This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
I believe I have seen somewhere that only the first letter of a website is to be capitalized in the references section "for example, "LATimes.com" should appear as "Latimes.com". If this is in fact the case, I believe it would be wise to note this in the Template:cite web. Zepppep (talk) 16:38, 30 September 2012 (UTC)
Citation style 1 does not seem to have a rule about this. If a rule is established, I would think one would want to mention it in all the CS1 templates, because many books, journals, newspapers, etc., are cited from their websites rather than the paper version. There is some merit to just using copy and paste to put a URL into a citation, and thus reduce the possibility of a typographical error. Jc3s5h (talk) 16:59, 30 September 2012 (UTC)
I don't think that MOS guidelines - certain to have exceptions - should be enforced through citation templates. Do we need a rule here? bobrayner (talk) 17:05, 30 September 2012 (UTC)
I am not worried about books or newspapers—for those in fact should be capitalized appropriately. Websites, however, often times appear as a single word (some might employ some use of hyphens, such as Baseball-Reference.com) and also employ a lot of abbreviations.
I am almost certain I've read that only the first letter of a website should be capitalized—after searching on a few different occasions, I haven't been able to fully retrace my steps. I shall check the MoS again (if anyone knows, feel free to link!), or perhaps other nooks and crannies. If it is indeed a rule, then I would encourage mention of it in the template because it's easily one of the most abused reference offensives out there. If it wasn't an oft abused stylistics offense I wouldn't suggest specific mention. Zepppep (talk) 17:32, 30 September 2012 (UTC)
Yes, that's the name. How about http:///www.nytimes.com? A lot of editors will put "work= or newspaper=NYTimes.com", even though it is simply the newspaper in a different medium. If the article/blog is mentioned in the web format only, however, they may very well put "work=NYTimes.com" when I believe it should be "work=Nytimes.com". Zepppep (talk) 04:39, 1 October 2012 (UTC)
Exactly. There seems to be some confusion here between "name of a website" and the url that points to it. E.g., "http://www.latimes.come/" is a url. And though it has been years since I've seen the relevant RFC I suspect that the general practice (if not the rule) is still that urls are case insensitive, so capitalization is irrelevant. ~ J. Johnson (JJ) (talk) 21:11, 30 September 2012 (UTC)
The address portion of the URL could be the name of the website, in the case of certain websites that brand themselves using ".com" as part of their name. The name for the newspaper in Ann Arbor, Michigan is AnnArbor.com, and like the online editions of its sister papers, it can also be accessed through the localized pages at MLive.com, http://www.annarbor.com/ and http://www.mlive.com/ann-arbor/ respectively. Imzadi 1979→ 21:37, 30 September 2012 (UTC)
For sure, some sites think it's cool to use their url as a "name", and use capitalization to make it more readable. But making the name (or title) look like a url does not make it a url; it's still just a name.
As to urls, there is, as Redrose points out, a caveat: the first single slash delimits the first part of the url (which is the domain name, and capitalization is as I said above) from the rest, which is file-system specific. And the file server may (or not) observe capitalization. In that regard it would be flat out error for "style" to over-rule technical reality. "AnnArbor.com/Forum" will probably work the same as "annarbor.com/Forum", but "annarbor.com/fORum" might not. ~ J. Johnson (JJ) (talk) 22:57, 30 September 2012 (UTC)
Regardless, I often view the page source and grab the content of the <title>. Doesn't work for every site, but most reliable sources will use a proper title. ---— Gadget850 (Ed)talk 01:49, 1 October 2012 (UTC)
If a website chooses to use its domain name as its name in general writing, it is a proper name and perhaps a trademark. It should be capitalized as the name-holder prefers, barring typographical difficulties such as mirror-image letters. As for the original poster's contention that only the first letter should be capitalized, it was reasonable to bring up that contention in the hope someone else would be able to find a Wikipedia policy or guideline that calls for that, or a reliable source indicating it is common English usage. But perhaps we have reached the point that we shouldn't give that contention further contention until a suitable source is found. Jc3s5h (talk) 14:41, 1 October 2012 (UTC)
Is this really the place for a MOS discussion? LeadSongDogcome howl! 06:29, 2 October 2012 (UTC)
Yes, where someone proposes to have the template reflect some style. But I think in this instance it has been adequately shown that capitalization of urls (whether of the domain name part, or the file-system part) is not something a template should be messing with. Hopefully that is settled. ~ J. Johnson (JJ) (talk) 23:14, 2 October 2012 (UTC)
BTW, it's me: I'm one of those editors who adds work=NYTimes.com, primarily because that's how the paper names its own site, and matches the proper capitalization of an abbreviation of the paper's name with ".com" appended. And yes, I'm one of those editors (oh dear) who thinks that Wikipedia should call things by the names and capitalizations offered by their copyright and trademark holders, or creators where such usage does not violate PROMO or the principle of least offense. Pardon me, dear, what was that word we reserved for people like that? Oh, yes, that's right: monsters, horrible horrible monsters. --Lexein (talk) 00:10, 21 October 2012 (UTC)
What's wrong with |work=New York Times? nytimes.com (HOweVEr yOU waNT to CApiTAliZE it) is not its name, it is its web address. —David Eppstein (talk) 00:15, 21 October 2012 (UTC)
Exactly the point I made earlier. The publication title is shown in the masthead of the page; where that is not clear, it is usually also in the <title> tag.
Should we clarify this in the documentation? ---— Gadget850 (Ed)talk 01:39, 21 October 2012 (UTC)
WP:SAYWHEREYOUREADIT says Say where you read it. The paper publication might not have carried the piece we're citing, and if it did, we don't know the page or section, so the masthead name is ambiguous for our purpose, and misleading. That is why I feel it's more descriptive to list work=NYTimes.com, the domain, not the full URL, and not the masthead name. Also listing publisher=New York Times Company is, of course, awesome. --Lexein (talk) 12:55, 21 October 2012 (UTC)
Continue as you will; others will certainly fix it. ---— Gadget850 (Ed)talk 13:21, 21 October 2012 (UTC)
Thanks for terminating civil discussion with a threat to stalk, without bothering to respond to the points, or thesis, raised. --Lexein (talk) 23:06, 21 October 2012 (UTC)
As previously noted, capitalization of domain names probably does not matter. But in general, we should not be changing capitalization. Even if (say) "NYTimes.com" is better (clearer?) than "nytimes.com", that is for them to do, not us.
I think what Lexein really wants is a way of distinguishing the website of the newspaper from the newspaper itself. I think this is a legitimate problem (I recently encountered the same problem with Scientific American). Lexein attempts a resolution by using the domain name to imply a website (a dubious approach), then wants to dress up the domain name capitalizing it. The real issue is distinguishing websites, but that is not what we are discussing here; it should have a new thread. ~ J. Johnson (JJ) (talk) 20:04, 21 October 2012 (UTC)
An article that appears in The New York Times could appear as follows:
Reporter, Joe (January 1, 1901). "An Article Never Online". The New York Times. p. 1. ISSN0362-4331.
Reporter, Joe (January 1, 2001). "An Article Never in Print". The New York Times. Retrieved October 21, 2012.
Notice how the presence or absence of page numbers, ISSNs, hyperlinks, and access dates indicates if an article is available online, in print, or both. I've done this with articles from The Mining Journal in Marquette, Michigan, because not all of their articles are repeated online each day. Some of the articles I've used at County Road 595 (Marquette County, Michigan) have hyperlinks (and even pre-emptively archived links) and some do not, but all of them have the page numbers and ISSN from the print edition. The one article from The Morning Sun in Mount Pleasant, Michigan, was online-only, so it lacks an ID or page number. Imzadi 1979→ 22:33, 21 October 2012 (UTC)
J. Johnson: C'mon, don't talk about editors. The nytimes.com searchbar has this obvious text: "Search All NYTimes.com", so the capitalization is indeed theirs, not mine, ffs. There's no dubious approach to associating domain name and website for our purposes: Wikipedia is no stranger to protocol-stripping URLs (http://nytimes.com) for discussion of websites (nytimes.com) in articles to avoid bare urls. If you're arguing that work=The New York Times website is best, I'd shrug, and not revert it if I saw it. But I'd fight for work=AnnArbor.com - heh.
Imzadi: I don't think implication is enough. We naively do not know if a modern item was ever, or never, in print, seeing only its web instance. We may know later, from news archive services, and/or citation services. I still think work=NYTimes.com is defensible as identifying where I read it. --Lexein (talk) 23:06, 21 October 2012 (UTC)
Well, since the newspaper published thrice weekly in print for Ann Arbor, Michigan, is titled AnnArbor.com on its masthead, which matches the online edition, that would be correct. However, the name for the online edition of The New York Times is just that. Hopefully you don't revert people correcting citations to use the proper publication name, per the websites' mastheads or <title></title> attributes. BTW, let me clarify something. I consult both the online and print editions of The Mining Journal when I use the combined citation; had I not been able to consult the print edition, I'd only cite the online edition sans ISSN and page number reference. Imzadi 1979→ 23:28, 21 October 2012 (UTC)
To reiterate, the masthead searchbar has this obvious text: "Search All NYTimes.com", so the name, and capitalization is indeed theirs, not mine. So your argument really is not with me, it is with the Gray Lady herself, until the literal text "NYTimes.com" is removed from the masthead of the online edition. --Lexein (talk) 00:30, 22 October 2012 (UTC)
Searchbar? Huh? What if it showed Search this web site? The masthead shows The New York Times; I see "NYTimes.com Search" on the search page, but not on the main page. This is starting to feel like a greased pig. ---— Gadget850 (Ed)talk 02:21, 22 October 2012 (UTC)
Okay, "searchbox", then. See screengrab. "NYTimes.com" appears on all pages but the homepage, including "Today's Paper". Greased pig? The original post was about capitalizing domains, which I do. Interestingly, the Sydney Morning Herald declares the domain fully lowercase, so I may have been doing that one wrong. Since work=domain.tld is accepted, I can be persuaded to keep lowercase as in latimes.com (since they show so in <title> text), and NYTimes.com (since they specify on-page). (To be honest, I'll still want to capitalize LATimes.com). --Lexein (talk) 07:01, 22 October 2012 (UTC)
I suppose we should be thankful you don't insist on writing . —David Eppstein (talk) 07:08, 22 October 2012 (UTC)
I swear that looks like "Los Ungeles", or, The Ungulates, which includes camels, which inspired CamelCaps titles (used on early wikis), presented as a deprecated capitalization example above. --Lexein (talk) 07:53, 22 October 2012 (UTC)
Yes indeed. Lexein, I think you misunderstand what I've said. First, let's not confuse the name of a publication, or website, with its url. (Which can be difficult, given that some sites incorporate their domain name into their proper name.) Second, I am not saying that a domain name should or should not be capitalized, I'm saying that we defer to the site's preferences. That the NYT has a web designer that used a capitalized form of their domain name certainly confuses matters, even gives others some license to do likewise, but we should be careful to not "improve" a domain name (or any part of a url) used as a url. ~ J. Johnson (JJ) (talk) 18:33, 22 October 2012 (UTC)
Using "page" as an alias for "pages"
Is it technically feasible to have "page" become an alias for "pages" and have the "pages" parameter detect the presence of any dashes, endashes, or mdashes, or commas to toggle "pp." instead of "p."? The difference between these two parameters I think is subtle and most editors in a hurry will not learn the difference. Unless a book uses really bizarre page numbering with dashes, this seems like a possibility. Jason Quinn (talk) 16:12, 25 September 2012 (UTC)
Consider option pages only: It would be easier to auto-adjust only option "pages=" which seems the worst case, as wrong in perhaps 45,000 articles as 9% incorrect (91% ok) but could auto-adjust most (as 99% ok). That would avoid issues people have raised here. See 2 sub-threads: #Feasible and fast for "pages=n", and also "#Cite_quick detects "pages=n" as singular". -Wikid77 12:27/13:12, 3 October 2012 (UTC)
The issue is performance. Detection would have to be with string templates which are notorious for eating resources. I looked at something similar for detecting first names ending in a period so if the spearator was a period it would suppress it. ---— Gadget850 (Ed)talk 17:32, 25 September 2012 (UTC)
Another issue is that there are always boundary cases (e.g. page numbers with hyphens in them) that will be interpreted incorrectly by code trying to be smart, forcing us to define additional parameters to allow us to work around the problem. Sometimes it's better just to keep things simpler. —David Eppstein (talk) 17:48, 25 September 2012 (UTC)
I would counter that while the current solution using two parameters dependent on context is simple from an implementation perspective, it is not "simple" from a usability perspective. Jason Quinn (talk) 18:56, 25 September 2012 (UTC)
There are bots which try to be clever and amend e.g. |page=8, 10 to |pages=8, 10 or |pages=8 to |page=8, but they do sometimes make incorrect changes such as those mentioned above involving hyphens etc. A human check is always better than leaving the guesswork to automatic systems. --Redrose64 (talk) 19:09, 25 September 2012 (UTC)
I found a bot which makes such amendments, see here. No bad edits found yet. --Redrose64 (talk) 23:03, 12 October 2012 (UTC)
See this thread and follow either of the two diff links in the grey box. The bot assumed that a hyphenated identifier of a single page was a page range and amended accordingly. --Redrose64 (talk) 15:30, 25 October 2012 (UTC)
I fail to see that any usability difficulty has been shown. Currently matters are quite simple: "page" is singular, "pages" is plural. If I forget the "s" the result is obvious enough to find and fix. Without having to second guess whether the software is being devious. ~ J. Johnson (JJ) (talk) 23:14, 25 September 2012 (UTC)
And there are definitely cases where books use something like "18-12" to mean the 12th page in the 18th chapter. Ucucha (talk) 00:13, 26 September 2012 (UTC)
My experience suggests that editors misuse those two parameters a lot. I myself at one point thought that "pages" referred to total pages because I hadn't read the documentation. I use the Reftoolbar (2.0b), as I suspect many people do. The entry forms on this gadget need some improvement. There needs to be some re-arrangement of items for more logical grouping. There needs to be more tooltips with descriptions and important notes. There needs to be some addition/deletion of items based on commonality of usage. This might be a good project for mw:Micro Design Improvements. Jason Quinn (talk) 22:27, 26 September 2012 (UTC)
Those certainly seem like worthwhile improvements, but do not constitute a problem with p/pp. (Except for failure to read the documentation, but that is generic.) And the miniscule advantage of trying to surmise an editor's intent seems vastly outweighed by the performance cost. ~ 18:25, 27 September 2012 (UTC)
I disagree with you. 1) There is a problem with p/pp for the reasons stated above. 2) Software that is unintuitive is a problem and existence of documentation is not a reason to ignore it. 3) The advantage would not be minuscule. 4) Performance concerns are a valid concern but not too much (see Wikipedia:Don't worry about performance). Regarding the other improvements we agree upon, yesterday I made a post about them on the Reftoolbar talk page. Jason Quinn (talk) 21:08, 27 September 2012 (UTC)
Strange, I thought I was agreeing with you to the extent that Refbar can be improved, the documentation generally can be improved, and that failure to RTMF is a general problem. (No? Perhaps you want to be more specific?) But as I said, all of those, as "usability problems" are not relevant to this alleged p/pp problem. As to whether "p=" versus "pp=" is unintuitive, well, perhaps if you have never seen any kind of bibliography before, which doesn't look like a s/w problem.
And you should read Wikipedia:Don't worry about performance more carefully. Such as the bit that "you, as a user, should not worry about site performance" (bolding removed, italics added). Or (under Editors still have a role to play): "Particularly in the area of template design, optimising server performance is important ...." And I am inclined to believe Ed's statement (above) that there would be a performance issue. ~ J. Johnson (JJ) (talk) 22:42, 27 September 2012 (UTC)
We agree that Reftoolbar and the cite documentation can be improved. The disagreement in my last comment was only concerning "p/pp". I think I misinterpreted what you wrote about documentation. Maybe about your stance on p/pp too. Regarding documentation, I just want to reiterate that good design can lessen the need for the user to read the documentation and sometimes eliminate it altogether. The less the user has to read the documentation, the better. I say this because I want to assert that documentation that properly describes the functions of software is not an argument for the current design/implementation of those functions. I think we'd agree on those points too. Regarding "p/pp", it's not clear to me where you stand except for thinking that the suggestion is infeasible. It may be. I did argue why I thought there was an issue with two parameters doing essentially the same job. So when you wrote that my idea "did not constitute a problem with p/pp" it seems to me a rejection of my assertion that there is a problem with p/pp. From my point of view, it boils down to, "Yes, there is a usability issue with having to parameters to handle page numbers. Would having a single parameter that autoformat's the entry be worthwhile?" Clearly both you and Ed believe no and that the "fix" would be infeasible. That's fine. If it's infeasible, it's infeasible. I am not saying that my idea is feasible. Just proposing it. What's unclear to me is to what extent you agree with me that there is a problem with the two parameters in the first place. Your last message seemed like you took personal offense. None was intended so I hope you don't take a disagreement too emotionally, especially when they may arise simply out of misunderstandings. Jason Quinn (talk) 01:08, 28 September 2012 (UTC)
So your concern is solely with p/pp, and not those other issues? Okay, I'm glad that is sorted out. I have run now, but will try to get back to you tomorrow. ~ J. Johnson (JJ) (talk) 23:09, 28 September 2012 (UTC)
Even if a reliably accurate algorithm existed for determining when to use the abbreviation "p.", or "pp.", or neither, it is bound to give rise to additional processing time, so we should not attempt to introduce one to the Citation Style 1 or Citation Style 2 templates as they stand because users do worry about performance. There are users who actively seek to eliminate these templates precisely because they are slow. Look at discussions earlier on, also at WP:VPT, concerning the time it takes to preview Barack Obama, for example. Any attempt to introduce smart code for p./pp. will definitely slow the templates down even further, thus giving more ammunition to the "delete them all" group. --Redrose64 (talk) 08:44, 28 September 2012 (UTC)
in general though, the performance argument is often spurious. users do not worry about performance afaik. Some editors (a small user subset) do. Preview is an editor function; its performance should not determine the fate of the citation system, which unlike performance, is not only an application of actual policy (WP:V) but may be also useful to the entire, much larger universe of users. at some point it has to be decided whether speed is better than completeness in an online encyclopedia, with actual proof offered that the largest by far group of users (the readers) prefers one over the other. until then policy should take precedence. 70.19.122.39 (talk) 14:12, 28 September 2012 (UTC)
And it is editors, not users, who need to consider whether using |page=, |pages= or |at= is most appropriate. --Redrose64 (talk) 16:42, 28 September 2012 (UTC)
Jason: let's take this back to the top. You asked if it was "technically feasible" to have the software determine the proper use of "p." versus "pp." depending on the supplied values. To answer very narrowly, I believe the specific answer is "yes, it is technically feasible".
But as has been explained here, this in no way means it would be a good thing to do. Especially as there would undoubtably be performance issues. Which you said we should not worry about (per WP:PERF), which I say you have misunderstood. You might not agree, but at this point it appears consensus is against you.
Now where you are being annoying is, in having asked a simple question in a seeming simple desire to know the answer, you now argue the point. It seems that, after all, it was a loaded question, that you really have an agenda, and that you want an answer other than what you have gotten. Which agenda is based on a postulated "usability issue". Sorry, but 1) I (for one, and I believe the others here as well) do not see that you have demonstrated any usability issue, and 2) I really do not feel like arguing the point.
If the bottom line is (as you said above) that you are "unclear" as to the extent I may agree with you "that there is a problem", then I can resolve this very readily: No, I do not agree that there is a "usability" problem. Beyond that, I do not care to argue the point.
While you may be able to get the template to detect whether there is a hyphen or dash in the page field (with uncertain performance implications) what there appears to be no solution for is how to determine whether this is a page range or some other page numbering system - numbering schemes like 5-103 are not that uncommon and need to be accomodated. This kills this proposal dead in the water.Nigel Ish (talk) 19:54, 29 September 2012 (UTC)
Feasible and fast for "pages=n"
It would be easier to auto-adjust option "pages=" which seems the worst case (leave "page=" separate), as pages=n is wrong in perhaps 45,000 articles as 9% incorrect (91% ok), but an auto-adjust of "pages=" could fix most, as then 99% correct use of "p." singular. I have run a "feasibility study" and found it feasible and quick (1/1,000 second) to detect a typical page-range "3-7" or comma-set "3, 7" or ampersand-pair ("5 & 9"). However, to reduce confusion, I would limit use to "pages=" rather than "page=" which could still be reserved for hyphenated page numbers, such as "A-7". Otherwise, "pages=A-7" would be considered as plural range "pp. A-7" but "page=A-7" would still be "p. A-7". If every citation used "pages=" then the extra overhead would be less than 1/3 second for every 200 citations combined. For citations which used singular parameter "page=" there would be no extra overhead. We can get lightning-fast performance from smart templates if we are careful about where to add clever features, and I have seen many people using "pages=56" to show incorrect "pp. 56". Templates can perform amazing tasks, such as:
((convert/spell |17,000,000,000,001 |mi|ly|abbr=off)) → seventeen trillion and one miles (2.8918325153954 light-years)
A better quick algorithm to detect the plural "pp." would be:
((#iferror: ((#expr: (({pages|528-32))}00000 ))
| pp.
|((#ifexpr:(({pages|528-32))}00000 < 1 |pp.|p.))
))
Any alpha page id would be treated as plural, such as "pages=A7" but that allows "pages=iv - 2" to correctly show "pp."
Anyway, I think "pages=56" could be detected, very quickly, as being "p." singular. In fact, I think I have added that to Template:Cite_quick. -Wikid77 (talk) 21:25, 2 Oct., 12:27/15:19, 3 October 2012 (UTC)
Hmm, that is an interesting idea. If I understand you correctly, you are suggesting that |pages= should render as "page" where it finds only a single page value. (But not having |page= expand to "pages", which is the initial proposal.) I speculate that the argument for this is that many editors have templates with "pages", and don't take the "s" off when the have only a single page number. (And perhaps the counter-argument is that templates should not condone sloppiness or laziness, but should issue big, red warnings?) I see one possible problem, where editors have intended (or understood) |pages= to be a page count, which is properly "pages" with a single number. Perhaps that should be deemed a misuse of the template? ~ J. Johnson (JJ) (talk) 23:34, 2 October 2012 (UTC)
The example templates have "pages=" which fosters incorrect use: So, the widespread use of the plural form "pages=n" is likely caused by some blank example templates showing option "pages=" rather than "page". As for the count of total pages, some people use "pages=999 pages" which would show as "pp. 999 pages" either way. -Wikid77 12:27, 3 October 2012 (UTC)
Yes, that last bit is a definite error, but not, I think, one that the template should try to anticipate and correct. ~ J. Johnson (JJ) (talk) 17:49, 3 October 2012 (UTC)
Cite_quick detects "pages=n" as singular
I have upgraded fast Template:Cite_quick to detect when only one page number is placed in "pages=n" to show "p." rather than "pp." for a singular page number. Examples:
((cite web | title=My page | pages=56 )) → "My page". p. 56. ((cite web)): Missing or empty |url= (help)
Since the method seems to work, I think it could be used in {cite_web}, {cite_book} or {cite_journal}, etc. Assuming only half of citations use "pages=..." then the overhead would average 1/2000th second per citation, or total 1/10th second more for 200 cites (half using "page=" or none). -Wikid77 (talk) 13:12/13:42/15:19, 3 October 2012 (UTC)
Cite_web/sandbox4 detects "pages=n" as singular
I have also upgraded fast Template:Cite_web/sandbox4 to detect when only one page number is placed in "pages=n" to show "p." rather than "pp." for a singular page number. Examples:
((cite web | title = My page | pages=56 )) → "My page". p. 56. ((cite web)): Missing or empty |url= (help)
((cite web/sandbox4|title=My page|pages=56)) → "My page". p. 56.
((cite web/sandbox4|title=My page|pages=5-9)) → "My page". pp. 5-9.
((cite web/sandbox4|title=My page|pages=3, 4, 9)) → "My book". pp. 3, 4, 9.
((cite web/sandbox4|title=My page|pages=34.2)) → "My book". p. 34.2.
((cite web/sandbox4|title=My page|pages=34/38)) → "My book". pp. 34/38.
((cite web/sandbox4|title=My page|pages=3, 4-6, 9)) → "My book". pp. 3, 4-6, 9.
((cite web/sandbox4|title=My page|pages=5 & 93)) → "My book". pp. 5 & 93.
((cite web/sandbox4|title=My page|pages=34–37)) → "My book". pp. 34–37.
((cite web/sandbox4|title=My page|pages=890)) → "My page". p. 890.
Since the page format works for {cite_web}, then similar logic could be used in {cite_book}. -Wikid77 (talk) 08:03, 5 October 2012 (UTC)
(Digression into use of citation templates)
70.19.122.39, there is no policy requirement to use citation templates; they are only one of many acceptable ways to provide citations. Jc3s5h (talk) 16:48, 28 September 2012 (UTC)
nowhere did i suggest that a citation system is policy. what i said was that cs1 is a templated application of policy, and i would not force anyone to use it. conversely i am not going to agree when someone forces me not to. let's run down the obvious:
a citation system has no independent justification. it only exists to apply verification. it is only a component, an important one, of a larger work. therefore the presence, quality and readability of citations directly impacts the enclosing work. this is obvious
unlike other citation systems that mainly support original research, such as in sholarly journals, nonfiction lit, theses, etc. the wikipedia citation system exists to provide a modicum of relevance to the whole project. this is because this is an encyclopedia that "anyone can edit". unlike in other encyclopedias (most of which have no need for a formal citation system) editors and contributors are not vetted, and it cannot be presumed they are experts, knowledgeable or neutral in what they write. therefore, the wikipedia citation system has a unique role, and its relative importance is magnified. this is obvious.
unlike the target readership of other applications with citation systems, the readership of wikipedia cannot be presumed to be knowledgeable or experts. the citation system should therefore be more readable and complete than other systems whose intended audience is relatively familiar with the subject(s). this is obvious.
to help both nonexpert readers and nonexpert editors, the wikipedia citation system should have a measure of stability and consistency in presentation. this is obvious.
verification is not art or creative endeavor. it is painstaking drudgery. it follows logically that its associated citation system is helped along by standardisation and not by free-form idiosyncracy. this is obvious.
so now you make the call on whether a templated, properly optimized citation system is needed, and whether it should be subservient to nebulous (because they are not measured properly) claims of performance. i'm sure readers will appreciate getting potentially lower quality content, if only they can get it faster. in the meantime, anyone can ponder whether a built-in editor is needed in wikipedia. personally i cannot find another wikipedia tool that is more redundant. there's an absolute glut of good quality text+image editors for free out there. i think that in-place editing may actually represent a significant drag in system resources and performance. ofcourse i have no properly researched proof. but from what i've seen, this is par for the course here. in any case, i'm done with repeating all the obvious stuff. good luck everybody. 70.19.122.39 (talk) 13:52, 29 September 2012 (UTC)
I agree with Redrose64 on his last. And this is the talk page for CS1, which uses templates. When Lua is installed and stable, we can look at this again. ---— Gadget850 (Ed)talk 01:32, 29 September 2012 (UTC)
Publication type
Prior discussion at /Archive_1#Cite_journal seems not to have concluded. ((cite journal)) takes the |type=Systematic review as in
Bloggins J (2012 Oct 24). "Recent advances in stuffology" (Systematic review). ((cite journal)): Check date values in: |date= (help); Cite journal requires |journal= (help)
Somehow though, adding the |journal=Curr Stuff suppresses the display of |type=, as in
Bloggins J (2012 Oct 24). "Recent advances in stuffology". Curr Stuff (Systematic review). ((cite journal)): Check date values in: |date= (help)
Fix with Citation/core/sandbox: I have a fix in Template:Citation/core/sandbox. The problem was 2 ways to show the title of "Periodical" and I have added TitleType so that both will show the TitleType. The problem has also existed with {cite_news}, similar parameters:
((cite news |author=Bloggins J |title=Recent advances in stuffology |type=Systematic review|newspaper=Curr Stuff|date=2012 Oct 24)) Result: Bloggins J (2012 Oct 24). "Recent advances in stuffology". Curr Stuff (Systematic review). ((cite news)): Check date values in: |date= (help) Expected: Bloggins J (2012 Oct 24). "Recent advances in stuffology". Curr Stuff (Systematic review).
The extra "TitleType" in {Citation/core/sandbox} should fix the problem.
Perhaps use "format=(...)". There is also the option to instead use current parameter "format=Systematic review" as follows:
((cite journal |author=Bloggins J |title=Recent advances in stuffology |format=Systematic review|newspaper=Curr Stuff|date=2012 Oct 24)) Result: Bloggins J (2012 Oct 24). "Recent advances in stuffology". Curr Stuff. ((cite journal)): |format= requires |url= (help); Check date values in: |date= (help)
However, the parameter "type=" would be used to appear after the journal's title (not after the article title, where "(format)" appears). -Wikid77 (talk) 20:20/20:47, 24 October 2012 (UTC)
Please don't misuse format in that manner.
format: Format of the work referred to by url; examples: PDF, DOC, XLS...
If you were to use format in that manner for ((cite journal)), then you would have to use it for all citations in the article.
The better solution would be to enable type. ---— Gadget850 (Ed)talk 21:06, 24 October 2012 (UTC)
Thank you for addressing that. I agree with Gadget850, that would be an abuse of |format=. So far as reasonable, we should try to mirror standard elements of bibliographic metadata, such as the Dublin Core or x.39. What we name these elements on WP doesn't matter so much (we do after all have entrenched choices), but commingling their purposes does. LeadSongDogcome howl! 16:40, 25 October 2012 (UTC)
Avoiding double-dots
I have created a fast utility, Template:Hasdot to check when a parameter ends with "." or wikilinked dot "[[Washington, D.C.]]" for limited use in rare cases. It is preferred that editors omit the dots when possible, such as by using the form "Washington, DC" (with no dots). Earlier, we had discussed having the templates remove double-dots ("..") for parameters which end with dot "." but that would increase the expansion-depth usage of Template:Citation/core, which would impact infoboxes where {cite_web} entries are automatically generated, perhaps causing those articles with infoboxes to exceed the expansion-depth limit. Now, next year, the Lua version can be enhanced to omit the double-dot ".." when it occurs. However, for now, it would be too risky to use {hasdots} for all cases, but perhaps some of the 24-25 cite forks could be changed to remove the end-dot, before invoking Template:Citation/core, such as in the "agency=Time Inc." or similar. Remember that only the "big 4 cites" (web/book/news & journal) need to be limited for total resources, whereas the other forks could have some more-extravagant options added, this year. Using Template:Hasdot will increase a template's expansion depth by +6 levels, even though it can check 200 names per second for the "." ending. -Wikid77 (talk) 18:26, 24 October 2012 (UTC)
I proposed something like this a while back Template talk:Citation/core/Archive 13#Duplicated period, but it added too much overhead. See User:Gadget850/Remove trailing period. When including publishers and agencies, we advise against including the corporate designator. ---— Gadget850 (Ed)talk 23:25, 24 October 2012 (UTC)
Since when (where?) is the abbreviation for "District of Columbia" done with out periods? ~ J. Johnson (JJ) (talk) 20:49, 25 October 2012 (UTC)
Cite family templates parameter name styling
If you glance at the parameters listed at ((citation)) you can find at least four different parameter-name styles. Parameters can have names that are hyphenated (|editor-first=), underscored (|trans_title=), and two versions of run-on (|accessdate=, |YearNote=).
I recently had a need for |trans_title= which I don't often use. Because there is no consistency in the parameter naming conventions in the citation family of templates (nor, it seems, in many other templates) after a couple of failed attmpts to get the parameter right (|transtitle= and |trans-title=) I had to go look it up.
Can we pick one style for parameters, document that style, and deprecate the other styles? This change would need to be accomplished across all of the citation family. I can certainly help with the documentation.
—Trappist the monk (talk) 16:40, 27 October 2012 (UTC)
Consider having a "help" parameter suggest the option names; see further below, "Have a "help" parameter correct options". -Wikid77 05:36, 28 October 2012 (UTC)
The Citation template is not part of Citation Style 1. Jc3s5h (talk) 16:57, 27 October 2012 (UTC)
But it does use ((citation/core)), as does CS1 and the parameters are mostly the same, and they use mostly common documentation. As various editors have added parameters over the years, they have use various styles. The problem is that editors have been accustomed to these quirks, and there are many tools programmed around these parameters. YearNote is only used in core. ---— Gadget850 (Ed)talk 17:05, 27 October 2012 (UTC)
@Editor Jc3s5h: If this is not the right place for this discussion, tell me where to take it.
@Editor Gadget850: I'm not arguing for removal of the "old" parameters. They can stay – they have to because otherwise thousands of templates would be broken. What I am requesting is additional synonymous parameters that all adhere to a single uniform style so that future (and current) editors don't have to learn about the quirks of some long-past editor's naming preference. Am I making any sense?
—Trappist the monk (talk) 17:21, 27 October 2012 (UTC)
If this were specific to ((citation)) it should be on the talk page for that template or maybe in Help talk:Citation Style 2, but this is also relevant to the CS1 templates. In any case, the way the templates are currently implemented, every additional synonym for a parameter slows things down, and they are already slower than they should be. In addition, adding more synonyms does not seem like a good solution to the problem that we already have too many synonyms. —David Eppstein (talk) 18:44, 27 October 2012 (UTC)
If synonyms aren't a good solution, what is a good solution? Or, is the problem of non-standardized parameter name styles insoluble? not worth fixing? a symptom of other more complex problems? something else?
—Trappist the monk (talk) 19:08, 27 October 2012 (UTC)
Synonyms are a good solution provided they are sensible choices, and don't cause trouble elsewhere. But when a given parameter has several synonyms, this can slow down template processing. For example, ((Citation/core)) has a parameter |Surname1=, which has no synonyms - but it's coded so that under certain circumstances, if it's blank or absent, |EditorSurname1= is used instead. These two parameters are passed in from ((citation)) as follows:
These are used left to right, and as soon as a match is found, the rest are discarded. Thus, if a given use of ((citation)) has |last1=Smith|surname=Jones|authors=Brown, only Jones will be displayed, because (({surname comes earlier than (({last1 and (({authors.
So, in this case, |last= is the primary parameter, and |surname=|last1=|surname1=|author1=|author=|authors= are its synonyms, in descending order of precedence. |editor-last= is also a primary parameter, and also has six synonyms.
Since they are used left to right, it's sensible for the most-used forms to be at the left, and the least-used forms at the right, in order to improve processing time in the majority of pages. --Redrose64 (talk) 19:42, 27 October 2012 (UTC)
FYI: That order is for ((citation)). I recently optimized and standardized the the CS1 templates (it also made template comparison a lot easier). See Help:Citation Style 1/snippets. ---— Gadget850 (Ed)talk 01:52, 28 October 2012 (UTC)
I know that it's for ((citation)) that's why I mentioned that template name twice in my post; I chose that template specifically because it's the one named in the very first sentence of this thread. --Redrose64 (talk) 12:58, 28 October 2012 (UTC)
I just glanced through Help:Citation Style 1/snippets and noticed that the order of parameter evaluation in the periodical sections seems odd. First there is:
Why, in |At=, is there not |newspaper= and why does |work= come last?
—Trappist the monk (talk) 14:34, 28 October 2012 (UTC)
Good point. I updated the snippets and will update the templates when there is a need. ---— Gadget850 (Ed)talk 15:43, 28 October 2012 (UTC)
@Redrose64: I've learned something. Thanks. I was under the misconception that the template parameters were processed individually, left to right when clearly, that isn't true.
—Trappist the monk (talk) 14:34, 28 October 2012 (UTC)
Have a "help" parameter correct options: In general, there will always be suggestions for better parameter names, or questions why the names are not easier. The experimental Template:Fcite_web offers the user suggestions for parameter names:
The tactic is to incur the slow, tedious help processing, to check the user's parameters, only when parameter 1 is "help" (or some common parameter is misspelled) to trigger all parameters to be checked for coherent usage. For example {cite_web} could warn:
((cite_web |help |last=Doe|author=Doe Jones)) → Cite_web: Found "last=Doe" & "author=Doe Jones" - use one, not both.
I have run some template-timing tests to show how 50 parameters could be tested, only when "help" or emergency, with no processing delay when the help-section was skipped for correct citations. -Wikid77 05:36, 28 October 2012 (UTC)
I think that I'm of two minds with this. I like code that emits useful, non-cryptic error messages. But. Editors are selfish. We will generally do what is most convenient so I suspect that editors using |help will fairly often leave the parameter in the cite – Yep, the cite displays the correct stuff, next edit ... – so it will be processed and thus incur the time penalty of spell-checking the template parameters until some other editor notices it and removes |help. One simple way to help ensure that |help parameters get removed sooner rather than later might be to always render the citation in red when |help is included.
Since processing time is important, and I know that it is from reading your posts elsewhere on the topic, having a uniform standard for template parameter name style will ease the processing load because it will reduce the number of possible names to check. Editors can be trained. We can learn to use |editor-last= instead of |editor-surname= (in my limited time editing wp, I don't recall ever seeing any use of a |surname= type parameter).
Choose a standard parameter name style. Add aliases where necessary. Publish the documentation with the new name style. Announce that the old style names will be deprecated on some date certain. Run a robot to change existing cites to the new name style (do this over all name spaces). At the specified date, deprecate names in the old style. Any new parameters must adhere to the standard name style before they are added to the templates. Have I missed anything?
—Trappist the monk (talk) 14:34, 28 October 2012 (UTC)
Looking through the listed parameters in Editor Gadget850's Help:Citation Style 1/snippets, these are the parameters that, in a non-hyphenated and non-underscored world (which I think is the path of least resistance), should have aliases to provide the CS1 templates with a singular parameter name style:
publication-date – add alias publicationdate (or perhaps better, pubdate)
contribution-url – publicationurl (puburl)
editor-last – editorlast, editorlast1, ... (like authorlink1 is done)
These aliases already exist but the order of precedence in Help:Citation Style 1/snippets should be swapped so that:
chapter-url is secondary to chapterurl
author-mask is secondary to authormask
On a different note, why don't |ARXIV=, |ASIN-TLD=, |BIBCODE= and |ZBL= (in §identifiers) have upper case versions like the others in that section? Are both cases necessary? This section seems to be the only place where both cases are used.
—Trappist the monk (talk) 00:55, 29 October 2012 (UTC)
I meant to mention this before: we use the term alias instead of synonym, but we know what you mean.
On your last: Again, different editors make for different standards. I added ASIN-TLD to core and asin-tld to each template; since it was new, I chose to keep the template version lower case.
Every added alias makes the template a bit bigger and adds some overhead. Using an order where the most used alias is first reduces overhead. I reordered some of the aliases when I reorganized all of the templates last month. We have a discussion above about reducing the aliases. ---— Gadget850 (Ed)talk 10:02, 29 October 2012 (UTC)
Ok, thanks, I've changed synonym to alias in my last two posts.
You wrote "... different editors make for different standards." That is the problem. Template editors forget or ignore the fact that their templates are used by editors who work in article space. For those article-space editors, it is incumbent upon the template editors to make the templates as easy to use as possible. One sure way of making templates as easy to use as possible is to have one, uniform parameter name style so that article-space editors only need remember the parameter name not whether this particular parameter includes a hyphen or an underscore or doesn't include either – the case that I tried, apparently unsuccessfully, to describe in my opening remarks.
Yes, adding these aliases will make the templates larger, but it's for a good cause. And, if we do the things that I've suggested, the deprecated parameters can go away and the template size will shrink.
I presume that you are referring to this discussion in archive 1? If so, you and I appear to be in agreement on many points. Doesn't seem that much has come from that discussion except your work that you've noted in this discussion. How can I help?
—Trappist the monk (talk) 13:24, 29 October 2012 (UTC)
Edit request on 10 November 2012
This edit request has been answered. Set the |answered= or |ans= parameter to no to reactivate your request.
Done. Actually, this needed to be added to Template:Cite book/doc, which is unprotected, so you could have done it yourself. I have made the edit for you, however. Best — Mr. Stradivarius(have a chat) 05:07, 10 November 2012 (UTC)
OK, mahalo! I will check the doc pages going forward. Peaceray (talk) 19:59, 10 November 2012 (UTC)
but it does not work here (note missing "tz"). Can someone please assist? Thanks. 219.79.74.171 (talk) 02:15, 17 November 2012 (UTC)
The interlanguage link for de:tz at 2012 Summer Olympics lacked the leading ":", with the result that an (erroneous) link to it appeared in the column listing the article's versions in other languages. Now fixed. I assume the same faulty construct worked here because talk pages are prevented from having interlanguage links the same way articles, categories, templates can have them. -- Michael Bednarek (talk) 06:14, 17 November 2012 (UTC)
Thanks! Not sure I understand, but anyway I was following the instructions at the interlanguage links help page. Either they need to be fixed or it's a bug, no? 219.79.74.171 (talk) 12:43, 17 November 2012 (UTC)
The first part of that page covers how to get links into the list which is seen on most pages at the bottom of the left-hand margin. For links within the text, you need to follow instructions further down, at Inline interlanguage links. --Redrose64 (talk) 13:04, 17 November 2012 (UTC)
(edit conflict) It's not a bug. There are two kinds of interlanguage links: those who link to an article's equivalent in another language (e.g. [[de:Olympische Sommerspiele 2012]] on the page 2012 Summer Olympics), and inline interlanguage links (which are somewhat discouraged) as explained on the page you quote in the section Help:Interlanguage links#Method where the leading ":" (colon) is shown. -- Michael Bednarek (talk) 13:12, 17 November 2012 (UTC)
Oh I see. Thanks for the clarification! 219.79.74.171 (talk) 14:11, 17 November 2012 (UTC)
language=Swedish not categorizing
When external links are tagged with one of the ((Language icon)) series of inline tags (short form e.g. ((sv icon))), the article is placed in a maintenance category corresponding to the language specified, e.g. Category:Articles with Swedish language external links. However, when specifying |language=Swedish this action is not effected. This is a discrepancy. I realize that some of the references are not to online texts, thus some additional provision must be introduced to make that distinction. __meco (talk) 14:09, 14 November 2012 (UTC)
Similar issues will exist for ((cite doi)), ((cite pmid)), and kin. Unless they are subst'd after being bot populated, applied categories will only be applied to the subpage, e.g. template:cite pmid/123456 will be placed in the category, not the transcluding mainspace article. For some reason I can't find an analogous category:Articles with Spanish language citations, perhaps there isn't one? How about category:Articles with deadlinks? LeadSongDogcome howl! 18:19, 14 November 2012 (UTC)
The Citation Style 1 templates do not categorize by language and have never done so nor do they support use of the ISO 639 name. ---— Gadget850 (Ed)talk 20:47, 14 November 2012 (UTC)
My point rather was that afaict they don't categorize the transcluding article by any criterion. Deadlinks seem more relevant than language to me, but perhaps I'm misunderstanding. It's yet another way the template subpages appear to be a bad idea. LeadSongDogcome howl! 21:48, 14 November 2012 (UTC)
LeadSongDog: I don't understand the connection. Templates such as ((cite doi)) use subtemplates that use ((cite journal)). Since ((cite journal)) does not do any language categorization, then ((cite doi)) certainly will not.
Meco: What are you looking for here? ---— Gadget850 (Ed)talk 03:16, 15 November 2012 (UTC)
Articles such as 2012 Norwegian C-130 crash should appear in Category:Articles with Norwegian language external links, where they belong since they are in fact articles with Norwegian language external links. __meco (talk) 13:16, 16 November 2012 (UTC)
Adding this to ((citation/core)) should do the trick:
((#if: (({language|))}
|((#if: (({URL|))}
|((#ifeq: (({language))}|English|
|((main other|[[Category:Articles with (({language))} language external links]]))
))))))
Limitations:
No check to ensure that the language is correct
No check to ensure the category exists
Does not support ISO 639 name
I have seen any number of misspelled language names in citations, so I expect a number of red categories to show. I also would expect to find languages such as Egyptian or Mandarin, which do not have categories.
So, this is technically possible with some limitations. Frankly, I don't understand the need for these categories. ---— Gadget850 (Ed)talk 14:52, 16 November 2012 (UTC)
I'm fuzzy on their intent as well, so I linked this discussion from wp:WikiProject Categories#Current discussions. Certainly there can be no doubt that such maintenance categories do exist, so at one point somebody thought they were worth implementing. My best guess is that it helps editors to be systematic in citation verification efforts, steering bilingual editors to the articles most needing their skills. LeadSongDogcome howl! 19:03, 16 November 2012 (UTC)
The problems with red-link categories being created is a minor one. People would notice the redlinks and fix the issue. __meco (talk) 17:17, 19 November 2012 (UTC)
Use wikisearch for those cite-language lists: I cannot emphasize, enough, the importance to use search-engine style lookup of parameter combinations, such as wikisearch of "language Norwegian" with other cite parameters. In many cases, the results are close enough, ignoring near matches, to provide useful results. For example:
That wikisearch reports more than 8,040 matching pages, and that is plenty of pages to analyze. For rare searches, feel free to wikisearch with accessdate & "language Egyptian" which matches 9 pages, most with "language Egyptian Arabic" or such. Remember in these recent years, the current bug/feature which allows such wikisearches, for template parameter names, is the magic that allows us to analyze cite-template usage in hundreds of possible combinations of rare parameters. This is the dream of so-called "associative retrieval" which many of us computer scientists have worked years to achieve. Please use wikisearch more, and create unusual rare categories less. -Wikid77 (talk) 00:16, 22 November 2012 (UTC)
Discuss phase-in of Lua-based citations
Because the wp:CS1 citation templates are so complex, with over 230 parameters, the transition to Lua-based cite templates should probably be a gradual phase-in, over a period of several months, beginning in Spring 2013. The initial Lua script cite module, Module:Citation has hundreds of parameters to provide support for the 24-25 forks of {cite_web}, including {cite_document}, {cite_encyclopedia}, {cite_video}, etc.
However, while Module:Citation has been designed to handle all wp:CS1 fork templates, I recently learned that it was also designed to cater to Vancouver-style cites, and special formats, such as small-caps for author name ("John G. Doe"). At this point, I fear too much confusion and advise having a separate Lua module (perhaps named "Module:CS1_cite_core"), just to handle the wp:CS1 templates, with no Vancouver-style features (no removal of dots in author initials) nor small-caps options or such. I can appreciate the techniques learned when dropping dots from names, but I fear the complexity will become unmanageable when mixing features, within a single Lua module, to handle parameters of wp:CS1, Vancouver-style, and small-caps custom formatting. Instead, a separate Lua module, dedicated to wp:CS1 features, should be designated for well-controlled updates. Of course, any clever new features accepted for other styles, in time, could be copied into the wp:CS1 Lua module, if approved.
In general, the Lua technology is a vastly different language system, where the stress testing of Lua has not been done to ensure support for 1.6 million articles, being reformatted hundreds at a time when some shared navboxes trigger all articles in the navbox, to be scheduled in the WP job queue for immediate reformatting. For example, I have seen Lua modules appear to "timeout" or go to sleep, to quit showing dependable results and say only, "Script error" (no other message) when used to process template parameters, perhaps based on a busy-load for the servers. Until full-out stress tests can prove Lua will keep formatting cites, and not "get tired to take a nap", then it should be phased-in by limited-use templates, added within a few articles, each time, to allow for safer testing, little-by-little. When Lua reaches a level of solid reliability, then the Lua-based templates and original markup-based templates can be name-redirected to point to the same templates. Meanwhile, there are still plans to quicken the current markup-based wp:CS1 templates, to allow faster operation, but also support all the current parameters. Also, there is the problem of Lua-experienced personnel, where some bug-fixes or proposed features might be delayed waiting for Lua updates, whereas the markup-based templates could be updated sooner by more people knowing the template markup language.
Any other thoughts about the phase-in of Lua-based cite templates? Should this topic be a separate subpage of CS1 discussions? -Wikid77 (talk) 04:52, 28 October 2012 (UTC)
Hi, Wikid77. I wasn't really following the cite-related discussion for a while and missed a lot of this lua stuff. Are these backend changes only? If I understand correctly, this should be transparent to editors and the cite templates themselves will continue to work the same as now and tools like the RefToolbar will continue to function uninterrupted. Is that true? Forgive me if the question is dumb. Jason Quinn (talk) 21:13, 21 November 2012 (UTC)
Lua-based cites have same parameters, nearly identical format: The Lua-based cite templates, as already developed on test2.wiki in September/October 2012, look exactly like using the identical {cite_*} templates, with all the exact same parameters, except over 10x times faster. If an article used the markup-based templates to format 250 citations, in 20 seconds, then the Lua-based footnotes will look identical, except instantly appear within 2 seconds. There are some slight format differences, such as numerous authors can be listed (far beyond just 9 names); however, all the major parameters appear in identical placement for {cite_web}, {cite_news}, {cite_book}, {cite_journal}, {cite_document}, and {cite_encyclopedia}. Testing of the other {cite_*} variations (such as {cite_video} ) has only been partially completed, and their speed is not much of an issue, so they could continue to use the markup-based {Citation/core} for weeks or months after the others are using Lua script. The COinS metadata, embedded for DASHBot or others, is still inside the Lua-based templates, formatted within the same 2 seconds (or faster). Already article "Barack Obama" is running with over 401 Lua-based citations on test2.wiki (see "test2:Barack Obama"), while also using markup-based templates for the President's infobox and naxboxes, etc. Hence, the RefToolbar will continue to work as before, even though some cites might use Lua-based templates while others use markup-based templates, during the transition period in early 2013. -Wikid77 (talk) 23:43, 21 November 2012 (UTC)
That's clears it up for me. Thanks for the reply. Jason Quinn (talk) 01:12, 22 November 2012 (UTC)
This sounds like a big win. The sooner we can get it sufficiently thoroughly tested to deploy, the better. Re the COinS data, I'm curious: does it fix the problem of <math> in titles? For the current template system, the title field is passed into COinS only slightly disguised by percent-encoding, and some later stage of wikimedia sees through the percent-encoding, thinks it should do something to the math in preparation for displaying it as Mathjax, and makes a mess. Other kinds of markup such as wikilinks in titles (or author names that are directly wikilinked rather than being linked through authorlink parameters) will also be passed into COinS, which isn't what should be happening even though the problems it makes aren't as visible immediately. The correct fix is to strip out wiki and html markup before putting the data into COinS, but that would be difficult using templates. Maybe Lua makes it easier? —David Eppstein (talk) 01:21, 22 November 2012 (UTC)
The Lua version can remove wikilinks and italics from COinS data much faster than in markup-based templates. Thanks for re-explaining the solution to work with math tags. -Wikid77 (talk) 18:16, 22 November 2012 (UTC)
It would appear the |others= parameter was removed from ((cite web)). Please update the documentation or put it back. Thanks. Goodraise 10:14, 23 November 2012 (UTC)
I have updated ((Citation Style documentation)) to include styling and templates to standardize the usage and examples section of documentation. --— Gadget850 (Ed)talk 16:00, 28 November 2012 (UTC)
Example headings?
Would it make sense to convert the bold text under the ==Examples== section into ===Tertiary=== TOC headings? Although it might clutter the TOC a little, I think that it would make the documentation more navigable. I am specifically starting this conversation with the Template:Cite web in mind, but I think this would apply to other documentation pages for Citation Style 1 templates as well. Peaceray (talk) 04:29, 30 November 2012 (UTC)
See #Citation Style documentation. I have been cleaning up the usage and example sections across the series and have documented the style. I don't see headings as useful for the examples. --— Gadget850 (Ed)talk 04:50, 30 November 2012 (UTC)
The guidance at Help:Citation Style 1#Titles and chapters, the third sentence in title reads: Should consistently use title case or sentence case throughout the article, and should use title case unless...
Regarding "title case," for WP CS1 references, is that intended to mean "capitalize all words, regardless of part of speech"? If not, which "rule" should it follow, based on the many examples at Title case#Headings and publication titles. Thank You. 71.212.92.154 (talk) 21:16, 1 December 2012 (UTC)
It Means to Capitalize All Words that Contain an Accented Syllable. (Also the first word is capitalized even if, as in this example, it is not an accented word). —David Eppstein (talk) 22:00, 1 December 2012 (UTC)
See WP:NCCAPS, third paragraph (the one about English titles of books, films, and other works). It primarily concerns titles of articles, but the same principles are used when giving book titles in references. --Redrose64 (talk) 22:31, 1 December 2012 (UTC)
Thanks for the responses. Redrose64's reference @ WP:NCCAPS says for details see MOS:CT and there are some specific guidelines there which I found very helpful. 71.212.92.154 (talk) 05:35, 2 December 2012 (UTC)
adding |agency=
I am aware that one is supposed to use ((cite news)) while citing news articles. However, in view of the fact that many editors, manually and using reflinks (for example here), complete citations using ((cite web)) template, and that citations are often news source with multiple contributors as well as agency credit, I think it desirable to add |agency=, so that citations that cite an agency have it rendered in screen output. -- Ohconfucius ping / poke 19:05, 30 November 2012 (UTC)
I agree. I feel like I run into this error often. -- Khazar2 (talk) 20:12, 1 December 2012 (UTC)
Done Added as an alias to series and version. --— Gadget850 (Ed)talk 19:25, 4 December 2012 (UTC)
Cite web or WebCite?
I would be grateful if I could get some opinions on two URL referencing tools, Cite web and Webcite. If we just use WebCite, the advantages are that (1) its one stable link that the reader can click and it will always be there. Multiple links may confuse the reader. (2) WebCite will also take a reliable snapshot of a website which may change later at any time. (3) Reader has access to the original URL at WebCite if he needs it. Example of WebCite:
We can use Cite web, where we can give both the actual URL and the archived URL. They can be switched around using the 'deadurl' URL. An example of Cite Web:
If we use Cite Web, we have to detect dead links and then switch around the URL's. Also the user may be confused seeing two links instead of one. I know Webcite has had some outages but its been online for most of the time.
So why should Cite Web be used instead of just WebCite? I know none of the available solutions fully solve the difficult problem of link rot so we're just making attempts to improve the chances that the reader gets to a working URL. --ApplePie3 (talk) 04:43, 8 December 2012 (UTC)
((cite web)) is a template for citing webpages, and it allows for full bibliographic information (authors, publication dates, publisher, work/website name, access date, etc). It can be used to link to the archived copy of a webpage on WebCite, the Wayback Machine/archive.org and other services. The other template doesn't do full citations, and it's limited to just one service. It's important to note that a webpage archived with WebCite is not permanent; the publisher/rights holder could request that the copy be removed from the archives at any time.
As for link rot, I just pre-emptively archive online copies of news articles. See County Road 595 where every digital copy of an article from The Mining Journal is pre-emptively archived already. I do believe that there is a bot that checks periodically through articles looking for dead links, and it will remove |deadurl=no as needed from the citation templates to perform the flip flop with the links. Since those are the online copies of news articles, they're cited using ((cite news)), which also allows linkages to ISSN or OCLC numbers and other identifiers to allow readers to search for libraries that have the print editions in their holdings. Imzadi 1979→ 05:59, 8 December 2012 (UTC)
((WebCite)) is not a full citation template— it does not support key information such as author, publisher, date of original publication and identifiers such as doi. It does not match the style of Citation Style 1 templates, thus they should not be mixed within an article per WP:CITEVAR.
((WebCite)) does not include the original URL. If WebCite were to go dark, then it is harder to find another archive of the site.
((cite web |last=House of Lords |title=Science and Technology - Sixth Report |work=UK Parliment |date=21 November 2000 |archiveurl=http://www.webcitation.org/5EsdOV0Sd |archivedate=3 April 2006))
House of Lords (21 November 2000). "Science and Technology - Sixth Report". UK Parliment. ((cite web)): |archive-url= requires |url= (help); Missing or empty |url= (help)
Thanks to both of you for the replies and additional information! --ApplePie3 (talk) 14:21, 8 December 2012 (UTC)
((webcite)) accepts just five parameters - four named and one positional; but the single positional param is merely an alias for |url=. One of the named parameters - |dateformat= - is used to implement an obsolescent technique for formatting dates; so really it has only three useful parameters |date=|title=|url=. --Redrose64 (talk) 15:13, 8 December 2012 (UTC)
Translator parameter in Cite book
I could not find a parameter for the translator, nor one for the original language of the book where the cited version is a translation. Often an analysis of authority can turn on the credentials of a translator, so I thought that it should be included in the citation. --Bejnar (talk) 17:29, 22 July 2012 (UTC)
See the template documentation; "others: To record other contributors to the work, such as "Illustrated by Smith" or "Trans. Smith"." ---— Gadget850 (Ed)talk 17:47, 22 July 2012 (UTC)
What about the original language? --Bejnar (talk) 11:00, 14 August 2012 (UTC)
I really think we should include a |translated_by= parameter (or something like it), as it is very common in many articles on classical literature, etc. --Thorwald (talk) 06:29, 8 September 2012 (UTC)
I too think that the absence of a |translated_by= parameter is debilitating, as well as the absence of an |original_language= parameter. --Bejnar (talk) 23:50, 11 December 2012 (UTC)
Class
All of the CS1 templates include a class such as book or video. Where is this used? --— Gadget850 (Ed)talk 13:07, 13 December 2012 (UTC)
I don't think it is used per se; but it's possible for custom CSS to use those classes to distinguish sources, if that's what you want. Try putting this into Special:MyPage/common.css and then look at East Linton railway station#References. Other colours are available... --Redrose64 (talk) 15:22, 13 December 2012 (UTC)
I well understand how the appearance can be changed, hidden or any variety of things. --— Gadget850 (Ed)talk 16:10, 13 December 2012 (UTC)
Maybe ping Andy? Might be a microformat. --Izno (talk) 17:38, 13 December 2012 (UTC)
I am pretty sure this predates Andy's involvement with microformats. --— Gadget850 (Ed)talk 21:51, 13 December 2012 (UTC)
What does [u.a] mean?
When I add references to articles using the Wikipedia:Reftoolbar using the auto-fill by ISBN feature, it very often adds "[u.a.]" to the city name in the location parameter. I have no idea what this means. Even our little encyclopedia doesn't answer that question. Anybody know? Jason Quinn (talk) 18:04, 13 December 2012 (UTC)
Unavailable? I would suggest that if it's not helpful in describing a source, leave it out. --Redrose64 (talk) 18:18, 13 December 2012 (UTC)
Well, usually it added entries like "New York [u.a.]". It's a good guess but I don't think "unavailable" works. Jason Quinn (talk) 18:40, 13 December 2012 (UTC)
Is it perhaps unitary authority to distinguish the city from the state? NtheP (talk) 19:05, 13 December 2012 (UTC)
Another good idea. Perhaps. I don't know. Jason Quinn (talk) 19:40, 13 December 2012 (UTC)
If we can't identify it, best to just leave it out. Annoying to have to re-edit, I know. Can't imagine a need for such for a book citation. --Lexein (talk) 20:09, 13 December 2012 (UTC)
I'd say it's German and means "among others". Goodraise 20:37, 13 December 2012 (UTC)
Is the an example ISBN that does this? RefToolbar uses http://toolserver.org/~alexz/ref/lookup.php to do the ISBN lookup, but I can't see the source. I don't know what database it pulls from, but I be the [u.a.] is in the database field. --— Gadget850 (Ed)talk 22:03, 13 December 2012 (UTC)
Hilary Clinton's book about the White House does it (ISBN0-684-85799-5). That was the book that was the last straw for me and made me investigate it. In my experience, many books do it, perhaps as high as 30 to 50%. The German theory is the most convincing so far. Regardless of the actual origin, I find the "if we [the editors interested in citations] can't figure it out" argument to carry great weight. I think it's enough to conclude that it should not be used. Jason Quinn (talk) 22:10, 13 December 2012 (UTC)
Good work. It's a real shame that such important functions of the default UI are inaccessible. I wonder if I could get a toolserver account, if only to see and copy this script. It appears that User:Mr.Z-man (the "alexz" above) has mostly quit editing Wikipedia. That being the case, someone else needs to maintain things. Jason Quinn (talk) 05:05, 14 December 2012 (UTC)
When a second or subsequent place is omitted, the omission may be indicated by "etc." or its equivalent in another script, enclosed in square brackets (see also 4.2.4).
It could be either "und ähnliches" or "und anderes", both of which are effectively "et cetera". As Seinfeld would put it, "[yadda, yadda, yadda]". LeadSongDogcome howl! 20:34, 14 December 2012 (UTC)
missing space between language and title?
This seems to be missing a space between language and title. Example:
((cite journal|journal=Sintra Cultural Guide|date=October 2009|issue=46|language=Portuguese))
renders as
Sintra Cultural Guide (in Portuguese) (46). October 2009. ((cite journal)): Missing or empty |title= (help)
but should (?) look like:
Sintra Cultural Guide (in Portuguese) (46). October 2009. ((cite journal)): Missing or empty |title= (help)
Problem is definitely in ((citation/core)); but funny things happen when you use ((cite journal)) and omit certain parameters. For example, it's normally best to fill in |title=, using the title of the specific article within that guide. When I do that, it spaces correctly:
"Article title goes here". Sintra Cultural Guide (in Portuguese) (46). October 2009.
Understood. I would - as normally do - include the article title, but I don't know it, I simply reformatted a previous edit, it is not my source. No big deal, and hopefully someone, someday will get it right. Tks - Nabla (talk) 01:25, 16 December 2012 (UTC)
archive 2
has anyone got any objection to adding |archiveurl2= and |archive2date= to the list of parameter as i have noticed one problem with archive if the archiving site is down you cant check the archived version so if you have 2 archive the chance of failure should be much lowerAndrewcrawford (talk - contrib) 14:49, 20 December 2012 (UTC)
How would you want to implement that? It would probably have to go in ((citation/core)). --— Gadget850 (Ed)talk 19:00, 21 December 2012 (UTC)
well you have just made me realise something this template isnt the core one, i thought i jsut need to alter this template but i will look at the sandbox and see if i can alter it to add a scond archiveAndrewcrawford (talk - contrib) 21:53, 21 December 2012 (UTC)
Again, how would you implement it? Would you show both archives in the citation. I have seen alternative archives noted in an HTML comment after the citation. --— Gadget850 (Ed)talk 22:05, 21 December 2012 (UTC)
i will have to see how it looks with listing both get feedback then go vrom there truthfully i aint sure the best way to do itAndrewcrawford (talk - contrib) 22:13, 21 December 2012 (UTC)
See ((Cite additional archived pages)). --— Gadget850 (Ed)talk 15:22, 25 December 2012 (UTC)
excellent that will save me work, but be nice to incorprate it into the main cite core in teh futureAndrewcrawford (talk - contrib) 16:41, 25 December 2012 (UTC)
It isn't a well used or publicized template, and it needs a bit of work on the output. --— Gadget850 (Ed)talk 14:20, 26 December 2012 (UTC)
Chapters and editors don't mix
((cite book|author=Whiting, Thomas A. L.|title=Maya|year=1998|chapter=The Maya Codices|publisher=Rizzoli|location=New York City|editor=Peter Schmidt))
results in:
Whiting, Thomas A. L. (1998). "The Maya Codices". In Peter Schmidt. Maya. New York City: Rizzoli.
Obviously, the chapter should be in the book, not in the editor! Probably something like:
Whiting, Thomas A. L. (1998). "The Maya Codices". In Maya. Peter Schmidt (ed.). New York City: Rizzoli.
Our citation style is heavily influenced by the APA style, modified for our particular uses. In looking over this webpage giving APA style examples, it uses:
Labajo, J. (2003). Body and voice: The construction of gender in flamenco. In T. Magrini (Ed.), Music and gender: perspectives from the Mediterranean (pp. 67-86). Chicago, IL: University of Chicago Press.
as one of the examples under the heading, "Essays or chapters in edited books". I guess if that's ok for the APA, it's ok for us too. Imzadi 1979→ 01:14, 21 December 2012 (UTC)
That the arrangement seems depends on your pov. I believe it reflects the scholarly/scientific eponymous use of the editor's name to refer to the work. I.e., "in" that person's work, not their physical person. ~ J. Johnson (JJ) (talk) 20:09, 26 December 2012 (UTC)
Perhaps pointing out the obvious here, Editor Kaldari's suggested citation rendering has similarities to Editor Imzadi1979 APA example. To wit:
In Maya. Peter Schmidt (ed.).
and
In T. Magrini (Ed.), Music and gender:...
Both have some notation identifying the editor as such. Punctuation is different and, to my mind, more correct in the APA example. Were I king of the world, I would fix the punctuation to be like APA and include the parenthetical lowercase editor abbreviation.
—Trappist the monk (talk) 20:59, 26 December 2012 (UTC)
Title not italic
"title: Title of source. Can be wikilinked to an existing Wikipedia article or url may be used to add an external link, but not both. If work is defined, then title is displayed in quotes, else displays in italics"
I've used this template extensively, but it's been brought to my attention at FAC that the title is not in italics as it's expected to be. Any chance this can be resolved please, otherwise I'll have to stop using this template to get through the review? (Example here) Thanks in advance. Socrates2008 (Talk) 09:58, 30 December 2012 (UTC)
Perhaps I've misunderstood - the example above has italicised the newspaper name, not the article name (as would be the case with a journal)? Socrates2008 (Talk) 21:47, 30 December 2012 (UTC)
Short works are in quotes, long works in italics. "Disastrous Gale in England" is an article, thus it is in quotes. Watertown Daily Times is the newspaper, thus in italics. The original citation confuses work and publisher. See Help:Citation Style 1#Work and publisher. --— Gadget850 (Ed)talk 23:21, 30 December 2012 (UTC)
Thanks Socrates2008 (Talk) 03:23, 31 December 2012 (UTC)
That's my query. It is not "aimed" at specific stylistic details (such as use of italics, commas, etc.) but at finding a general form for handling web citations. Comments invited. ~ J. Johnson (JJ) (talk) 00:24, 31 December 2012 (UTC)
Cite video: rename
I have made a proposal to rename ((cite video)) to ((cite media)). --— Gadget850 (Ed)talk 21:52, 10 December 2012 (UTC)
This edit request has been answered. Set the |answered= or |ans= parameter to no to reactivate your request.
Could someone replace the ((Template for discussion/dated)) transcluded into ((cite video)) with an embedded version with the text modified to make it clear that the template it being propsed for renaming not deletion? Thanks. DH85868993 (talk) 23:36, 14 December 2012 (UTC)
Not done: Sorry for the delay in answering this. I can see how this might be a good idea, but I think it would be better to fix this problem for all similar TfD nominations, not just this one. Maybe you could suggest this at WT:TFD, and if there is support for the idea we could update ((Template for discussion)) and/or ((Template for discussion/dated)) to allow renames to be specified by the proper template. Let me know if you need any help with the coding. Best — Mr. Stradivarius(have a chat) 14:06, 20 December 2012 (UTC)
No big deal either way, but web has only had noinclude because it checks for title, thus it gives an error message. If you do one, ten do the rest. But it only showed a period. --— Gadget850 (Ed)talk 01:53, 2 January 2013 (UTC)
Edit request
Can you update the template under Usage so that it has an access date of 2 January 2013, or even make it update automatically each day so that it needn't manually be done? Goodsmudge (talk) 19:33, 2 January 2013 (UTC)
Template:Cite web/doc tries to show today's date in the "Usage" section, but it may lag by a few days if Wikipedia's servers give you a cached version. If you purge the template, it should redisplay with the current date. -- John of Reading (talk) 19:45, 2 January 2013 (UTC)
I added a note at the top of the Usage section. You may have to purge the documentation to make it appear. --— Gadget850 (Ed)talk 21:53, 2 January 2013 (UTC)
((Cite web)) bug?
If the “separator” parameter is present but blank, and the “postscript” parameter is absent, then it seems that an extra blank space gets inserted between the publisher and the closing period, as in this example: "How does a bill become a law?". Michigan legislature. ((cite web)): Cite has empty unknown parameter: |separator= (help) Bwrs (talk) 01:46, 3 January 2013 (UTC)
Bug: issue not showing up
In converting to template style, I've found that journal issue numbers are not showing up. For example, on Joseph McElroy#Anthologies of McElroy criticism are two ((cite journal|...|issue=##)) templates, and the resulting text does not have the issue number. Choor monster (talk) 18:43, 3 January 2013 (UTC)
If you check the documentation, you will see that issue is a child of work— that is, issue is dependent upon work.
Markup
Renders as
((cite journal |title=The Review of Contemporary Fiction | year=1990 | volume=X | issue=1 | url=http://www.dalkeyarchive.com/book/?GCOI=15647100089080)) (table of contents)
And the same for Golden Handcuffs Review. --— Gadget850 (Ed)talk 19:06, 3 January 2013 (UTC)
Using ((Cite news)) to cite web-archived physical newspaper
I'm trying to cite this article from The Miami Times, 1979. Obviously, there is no original URL, so I put the the url in the archiveurl parameter. However, the resulting reference has the following error message: Error: If you specify |archiveurl=, you must first specify |url=. Why? What do I do in this particular case, where it an archive of an actual newspaper? —[AlanM1(talk)]— 23:55, 26 December 2012 (UTC)
Just use |url=, not |archiveurl=. --— Gadget850 (Ed)talk 01:02, 27 December 2012 (UTC)
I guess my point is that this fails to distinguish the fact that the URL is not the original publication, which is what I thought the point of having the archiveurl param was in the first place. —[AlanM1(talk)]— 07:48, 2 January 2013 (UTC)
But if you're going to be picky about the semantics of archiveurl vs url, using archiveurl fails to distinguish the fact that the version you read when you added the information from that source to Wikipedia was probably the online copy, not the original print publication. —David Eppstein (talk) 08:03, 2 January 2013 (UTC)
I'd consider using: ((cite news |newspaper=The Miami Times |date=August 28, 1979 |page=4A |title=Bolshoi ballerina greeted with tears |agency=Associated Press )) ([http://news.google.com/newspapers?nid=2206&dat=19790828&id=0dQzAAAAIBAJ&sjid=BuoFAAAAIBAJ&pg=839,2953781 image available online]) which renders as "Bolshoi ballerina greeted with tears". The Miami Times. Associated Press. August 28, 1979. p. 4A. (image available online) LeadSongDogcome howl! 23:23, 3 January 2013 (UTC)
That is a convoluted manner of accomplishing something so simple. --— Gadget850 (Ed)talk 01:21, 4 January 2013 (UTC)
It's not as simple a case as it seems, because Googlenews doesn't state an archivedate. As a result archiveurl is useless. One could, alternatively, archive the article on Webcitation or elsewhere to establish an archivedate, but the one provided is undated. LeadSongDogcome howl! 06:49, 4 January 2013 (UTC)
In this context, an archive is an electronic snapshot of a source originally published online; as such, there could be multiple snapshots as the page is repeatedly archived. This particular source was published in print, and the Google News version is an image of the original, thus there is no snapshot. --— Gadget850 (Ed)talk 12:41, 4 January 2013 (UTC)
Original URL not rendering
I used ((cite web)) on page Anna Carin Zidek. I have an archiveURL specified which renders properly, but the original URL does not. The citation comes out as:
^ Reimers, Johanna (16 July 2011). "Anna Carin Zideks besked: Jag slutar" (in Swedish). Expressen. Archived from [www.expressen.se/sport/skidskytte/anna-carin-zideks-besked-jag-slutar/ the original] on 8 January 2013.
The "Anna Carin Zideks besked: Jag slutar" is properly linked to the archived page but the original page is rendered with the external link brackets instead of rendering a URL. -- Whpq (talk) 11:52, 8 January 2013 (UTC)
Fixed You forgot the URI schemehttp://. MediaWiki needs the URI to detect that a string of text is a link. --— Gadget850 (Ed)talk 12:01, 8 January 2013 (UTC)
Thanks! -- Whpq (talk) 12:05, 8 January 2013 (UTC)
If the website supports both the http:// and https:// schemes (as Wikipedia does), the part before the double slash may be omitted, as in |url=//www.expressen.se/sport/skidskytte/anna-carin-zideks-besked-jag-slutar/ but the double slash is mandatory. --Redrose64 (talk) 20:40, 8 January 2013 (UTC)
That works in CS1 templates because ((citation/make link)) is adding the [] brackets around the link. --— Gadget850 (Ed)talk 15:09, 12 January 2013 (UTC)
Reuse of a named book citation, with a different page number?
Is there any way to reuse (in the same article) a named citation (e.g., <ref name=smith1995/>) but add a book page parameter to it?
In other words, if an initial cite book citation includes all the appropriate metadata (name, author, publisher, no. of pages in the book, etc.) and is named, say, <ref name=smith1995> ((Cite book |last =xyz first=abc |title=BookTitle etc.)), is there any way to avoid repeating all the book metadata again, reuse the citation, but add a unique page number for each subsequent reused citation? Something like "<ref name=smith1995/ |page=271–273>" would be very much appreciated. Cheers. N2e (talk) 13:41, 12 January 2013 (UTC)
Maybe the template ((rp)) can help? -- Michael Bednarek (talk) 13:50, 12 January 2013 (UTC)
My preference for this would be to use something like <ref>((harvtxt|Smith|1995|pp=271–273))</ref> for the second reference (and make sure to use |ref=harv in the first one so that the wikilink created by the template works correctly). But any style that unambiguously identifies the source and is consistent with the other citations should be ok. —David Eppstein (talk) 19:35, 12 January 2013 (UTC)
Which is the same as using ((sfn)), but I added an example. --— Gadget850 (Ed)talk 19:51, 12 January 2013 (UTC)
Actually, it's the same as ((sfnp)). I have a slight preference for sfnp over sfn (because that format works better for in-text references), but I think it's more important to choose one of the two formats consistently rather than mixing them. And I think if you want to re-use the footnote you need to do it with ref-harvtxt rather than sfn, which again argues for using sfnp for consistency even with the ones you don't want to reuse. —David Eppstein (talk) 20:16, 12 January 2013 (UTC)
You can reuse ((sfn)). Identical uses are combined in the reference list. ((sfn)), ((harvtxt)), et. al. all use ((Harvard citation/core)). --— Gadget850 (Ed)talk 20:51, 12 January 2013 (UTC)
(edit conflict) ((sfn)) operates like <ref>((harvnb))</ref>: there are no parentheses on either. However, ((sfnp)) operates not quite the same as <ref>((harvtxt))</ref>: they both have parentheses, but the position of the closing parentheses differs. The ((sfn))/((sfnp)) templates automatically merge duplicates. Whichever of the four is used, the associated Citation Style 1 template needs a |ref=harv --Redrose64 (talk) 20:55, 12 January 2013 (UTC)
Oh my, no, not ((rp))!! The basic answer here is to use some form of a short citation, each of which has the specific page number (or such) as appropriate for the text, and all refer to a full citation (or full reference) with the full bibliographic details, usually in a separate list. The suggestions of ((sfn)) and ((harv)) above are just different ways of implementing this. ~ J. Johnson (JJ) (talk) 23:19, 12 January 2013 (UTC)
((rp)) is a form of a short citation. If one uses ((rp)) aloing with named references to combine repeated citations, none of the meta data is duplicated. I major draw back of ((Harvard citation/core)) is that one needs to maintain two parallel sets of data. Futhermore changing one citation formate to another in an article that already has an established citation style without first gaining consensus contradicts WP:CITEVAR. Boghog (talk) 00:06, 13 January 2013 (UTC)
((rp)) puts some of the bibliography data in the reference section and some of it (the page numbers) in the text of an article. So, using this style, it is not possible to see the whole citation in one place and it becomes difficult to determine which pages of a source are actually referred to. It also adds to the reference-clutter in the main texts of articles. —David Eppstein (talk) 00:22, 13 January 2013 (UTC)
Well, first off, I very much appreciate that their is some way to do it. Having said that, it appears to be a bit complicated, based on the variety of opinions in the discussion above. Nevertheless, I think it is a problem worth the effort for me to figure out how to do it, and do it well, so I will probably invest the time to figure out the way these things work, sometime. Thanks again! Cheers. N2e (talk) 02:39, 13 January 2013 (UTC)
—Okay, I did a bit of research on ((sfn)) and ((sfnp)) to learn how to use them and was able to eliminate a number of "<!-- p. nnn -->" hidden comments in an article I had previously reused a <ref name=smith1995/>-type references in. It is better.
But it would seem to me to be much better if this multiple-reuse-of-a-single-book-<ref name=smith1995/>-ref were integrated with the <ref name=smith1995/> reference. That way, if any future editor eliminated the first use of the book citation (<ref name=smith1995/>), then the bots would pick it up and fix it. Also, it would be possible for the bots to keep the first use of the <ref name=smith1995/> ref "on top", where it is higher up in the References list than the repeat instances of ((sfn)) and ((sfnp)). YMMV, but that is my take. Cheers. N2e (talk) 13:34, 13 January 2013 (UTC)
Not integrated; segregated! The problems you touch on are inherent in the use of "named refs" ("<ref> name=..."). That's why I put only the specific short cite in the endnote (the <ref> tags or sfn template), and put the "reused" part — the full citation/reference that is common to all the short cites — in a separate section (not in a note). Which is the way you have to do it if you want to order the reference list (alphabetically, chronologically, whatever). ~ J. Johnson (JJ) (talk) 19:59, 13 January 2013 (UTC)
Sure, that also works. Didn't mean to imply otherwise, and I am clarifying my comment. The main point is to not trap the full citation in a "named ref". ~ J. Johnson (JJ) (talk) 21:52, 13 January 2013 (UTC)
I couldn't believe the amount of reading this template makes you sift through to tell you how to use it. I would suggest someone go through and collapse some of the more "so what" parts of it, like most of the Examples section that seems to go on forever—old format, new format, old format, new format—is this really necessary? Would someone using this template for the first time actually care?
Then there's a Deprecated sub-section that means nothing to people like me who don't care about what you used to be able to do. In addition, the Deprecated sub-section forgets to mention the other deprecated parameters, which are still (for some reason) listed in the Usage section (where you copy/paste the code); these parameters are listed as deprecated in their respective areas, but that seems really disorganized to me. Again, they shouldn't even be given as an option to use, if they're not going to work in the end.
In the end, I basically ended up with the most simplest example: Title of the show, season, episode, boom, done. But I had to spend 20 minutes trying to figure all of that out. While I'm not "new" at Wikipedia, I found the information on this template to be massive and almost intimidating. – Kerαunoςcopia◁galaxies 00:01, 14 January 2013 (UTC)
Also, does the time parameter include commercials or not? That'd be nice to know. Thanks. – Kerαunoςcopia◁galaxies 00:03, 14 January 2013 (UTC)
No. When chapter is used, it is usually an article, essay, short story manual section or the like in an edited collection, thus it is the primary source. --— Gadget850 (Ed)talk 00:07, 16 January 2013 (UTC)
Accessdate vs. Publication Date
Is it necessary to include the accessdate date when a publication date is available? I only use accessdate when a given webpage does not have a visible publication date, and almost never use accessdate when it does have one, since I think it's redundant, and can conceivably clutter up the citation. Is there any guideline or consensus on this? Nightscream (talk) 22:13, 15 January 2013 (UTC)
14.7 ACCESS DATES An access date--that is, the self-reported date on which an author consulted a source--is of limited value: previous versions will often be unavailable to readers; authors typically consult a source any number of times over the course of days or months; and the accuracy of such dates, once recorded, cannot readily be verified by editors or publishers. Chicago does not therefore require access dates in its published citations of electronic sources unless no date of publication or revision can be determined from the source (see 14.8). For such undated sources--or for any source that seems likely to change without notice--authors are encouraged, as an additional safeguard, to archive dated copies, either as hard copy or in electronic form.
If you check the documentation for accessdate, you will find that you can hide it as desired. --— Gadget850 (Ed)talk 13:09, 16 January 2013 (UTC)
I tend to agree. Archive dates are useful if not essential in the case of archived links, but I find access dates next to useless. Most pages are there or are later removed; some pages are dynamic and updated on a daily basis. Accessdates will not help finding the 'consulted' versions of these in either case once they are gone, unless they have been pre-emptively archived, yet Reflinks insert them systematically. I then have to remove them manually, which is a drain on productivity. Maybe I don't properly understand their uses, but as far as I'm concerned they potentially take up Gigabytes of disk space, and perhaps someone could make a case for their whole-scale removal. -- Ohconfucius ping / poke 14:02, 16 January 2013 (UTC)
The value I see in access dates is that they can be used for an editor to find an archived page, when an archive-url and date is not provided, the special case of which is when the original page url no longer works and it is desired to link to an archive to ensure verifiability. Without an access date in the special case, we need to guess at when the page may have been archived. --Izno (talk) 14:47, 16 January 2013 (UTC)
I have never understood Wikipedia's obsession with access dates. They rarely serve any purpose. Sadly, many editors apparently think that, if they put in an access date, they don't have to bother with the publication date. News references in particular should always show the exact publication date, which is much more important than the accessdate. I wish this could be spelled out more clearly. -- Alarics (talk) 18:10, 16 January 2013 (UTC)
Access dates serve a valuable purpose when citing undated web pages that might be subject to future changes. For newspaper or scientific journal references, though, I agree that the pub date is what we should be using. —David Eppstein (talk) 18:14, 16 January 2013 (UTC)
As said above, for sites or pages that are frequently modified (such as home pages that are cited), the only solution is pre-emptive archiving, because you can never be quite sure which is the version is cited otherwise. Merely having the access date could be helpful, but in practice it's pot luck. When a page has an archive link, the access date is redundant. -- Ohconfucius ping / poke 02:28, 17 January 2013 (UTC)
Seasons & Episodes
To whomever made the recent change to the template so that the words "season" and "episode" both appear in the saved info: THANK YOU! It looks MUCH more clear now!
Just one suggested tweak, though: Could we flip them so that the season is given first, and then the episode? For one thing, hierarchically-arranged information tends to be displayed from the general to the specific. For another, the widespread convention is for seasons to be given first. For example, among both sources and informal vernacular, "4.17" often used to denote Season 4, Episode 14 of a TV series. Lastly, putting the season first would emphasize, at a glance, that the episode given is not the overall episode number of the series, but only of that season.
I don't have a lot of experience in changing templates, so how do we go about implementing this? Do we need a consensus discussion? Nightscream (talk) 02:36, 13 December 2012 (UTC)
I presume you are referring to ((cite episode)). It was updated in April, and I used the same order as before the update. Season/episode does make sense to me, but lets see if there is any discussion. --— Gadget850 (Ed)talk 10:47, 13 December 2012 (UTC)
I can get on board with the switch. It just follows what is customary practice to specify the larger unit before the smaller one in citations. Imzadi 1979→ 23:54, 17 December 2012 (UTC)
So how do we go about doing it? Nightscream (talk) 13:48, 11 January 2013 (UTC)
It is on my todo list. --— Gadget850 (Ed)talk 13:51, 11 January 2013 (UTC)
Now in sandbox. Please review and comment. --— Gadget850 (Ed)talk 16:01, 15 January 2013 (UTC)
Current
((cite episode |title=Billy Crystal, 2nd Visit |series=Inside the Actors Studio |date=October 8, 2007 |url=http://www.bravotv.com/Inside_the_Actors_Studio/guest/Billy_Crystal_-_2nd_Visit |network=Bravo |season=13 |number=1307 |last=Lipton |first=James (host)))
Lipton, James (host) (October 8, 2007). "Billy Crystal, 2nd Visit". Inside the Actors Studio. Season 13. Episode 1307. Bravo.
Sandbox
((cite episode/sandbox |title=Billy Crystal, 2nd Visit |series=Inside the Actors Studio |date=October 8, 2007 |url=http://www.bravotv.com/Inside_the_Actors_Studio/guest/Billy_Crystal_-_2nd_Visit |network=Bravo |season=13 |number=1307 |last=Lipton |first=James (host)))
At Template:Cite book, the period should be after the quotation mark where it says: "Translated by John Smith." --82.170.113.123 (talk) 13:13, 18 January 2013 (UTC)
Fixed to use proper emphasis. --— Gadget850 (Ed)talk 13:24, 18 January 2013 (UTC)
Please change embedded explanation of "Work" in cite web template to match its description in documentation
If you use the dropdown to insert a "cite web" template, the vaguely titled "Work" field provides a "?" over which you can hover, which then says "What larger work this is part of?". Every time I see that, I ask myself "What the hell do they mean?". Then, buried in documentation, I see that it means "Title of website". Which explanation is short, sweet, and readily understandable. Can we please make the hint match the documentation? --Hobbes Goodyear (talk) 02:24, 2 February 2013 (UTC)
Hey, thanks a lot. I've placed a request on the talk page there. --Hobbes Goodyear (talk) 13:15, 2 February 2013 (UTC)
Cite additional archived pages
I updated ((cite additional archived pages)). It now supports multiple archives with different dates as well as an archived series of pages with the same archive date. --— Gadget850 (Ed)talk 22:30, 3 February 2013 (UTC)
Cite book from an on-line source
Can anyone say if they think I've cited correctly in this edit. My concern is that I accessed the book via the URL and not the physical book, and that perhaps this is not made clear enough from the cite. Either way, I think some wording should be added to the intro of Template:Cite book confirming how on-line book cites should be dealt with. Eldumpo (talk) 12:27, 3 February 2013 (UTC)
I fixed the |isbn= parameter. --Redrose64 (talk) 14:22, 3 February 2013 (UTC)
Thanks. I assume the format was OK and that there's no requirement to note that you haven't viewed the physical book, although I think the template intro should clarify matters. Eldumpo (talk) 20:05, 3 February 2013 (UTC)
I don't think there is one single answer that is always best. If the book is readily available and there is a fair risk that the website may disappear, it's probably better to do as you did. If the book is rare and the website is by a stable organization that seems prepared to make the website available for a long time, it might be better to treat the organization that runs the website as a republisher, the same way one might cite Dover Publications if one had read one of their paper editions if that is the edition you read. Jc3s5h (talk) 20:17, 3 February 2013 (UTC)
the existence of the hyperlinked title and retrieval date sufficiently indicate (and imply preference, imo) that the citation is verifiable online, irrespective of the source medium you consulted. 70.19.122.39 (talk) 14:48, 4 February 2013 (UTC)
Citing newspapers by contemporary title
In the previous item, note that I cited the paper as the
''[[Toronto Star|Toronto Daily Star]]''
This makes sense to me because I believe it is correct practice to refer to a periodical by its title contemporary with the cited article, with a Wikipedia link provided if possible
so people can connect that title to the current newspaper. If there isn't a Wikipedia article, I believe it would be correct to write
''Toronto Daily Star'' (now the ''Toronto Star'')
The ((cite news)) template does not support either of these styles. I think it should, using an added field called something like moderntitle=.
For publications no longer in print but better known under a later title, it could be latertitle=, which would expand to "(later the...)".
In the case when there are multiple references in an article to the same former title (again see the previous item), it would be nice if all the cites could be provided by the contributor
in the same style and any repetitiousness could be suppressed automatically; but that'd just be an extra bonus feature, not a requirement.
you can add info about the work (newspaper) in the |postscript= field (displays last).
escape the vertical bar in the wikilink by using the ((!)) template:
also, postscript does not appear in the mislabeled "full parameter list" (it's not the full list, there are other params not listed). you have to go through the detailed doc. 65.88.88.127 (talk) 20:40, 4 February 2013 (UTC)
to indicate that this is a renamed publication maybe series could be used, if it was available (it's not). as in |work=Toronto Star|series=Toronto Daily Star. 65.88.88.127 (talk) 21:14, 4 February 2013 (UTC)
Give the name of the publication as it was at the time. If there is potential for confusion, wikilink it to the article on the present name. --Redrose64 (talk) 21:54, 4 February 2013 (UTC)
I agree with Redrose on this point. The citation should list the source with its contemporary information. Imzadi 1979→ 15:31, 6 February 2013 (UTC)
Copyright tag
There is nothing to let you indicate the copyright holder. I'm sure Disney would not be amused, but actually I'm looking at Egyptian Grammar by Alan Gardiner where the Griffith Institute holds the copyright. Aarghdvaark (talk) 10:42, 28 January 2013 (UTC)
That might be, but the purpose of a citation is to allow someone to locate a copy of the source, specifically the same edition consulted where necessary. I'm not sure that indicating the copyright holder, or lack of one, would assist someone in locating a book in a library. Imzadi 1979→ 10:46, 28 January 2013 (UTC)
Agree. Citations are not intended to include every snippet of information about a source. --— Gadget850 (Ed)talk 11:14, 28 January 2013 (UTC)
The copyright holder is often the same as the author or publisher. In the rare cases that it's not, it isn't relevant. That said, some websites don't name a publisher but do name a copyright holder, so I put that information into |publisher= --Redrose64 (talk) 15:33, 28 January 2013 (UTC)
But that defeats the object of this tagged citation system! Aarghdvaark (talk) 15:45, 28 January 2013 (UTC)
Aarghdvaark, how does indicating the copyright holder benefit a reader looking to locate a copy of, or evaluate the reliability of, a source used in an article? Imzadi 1979→ 15:53, 28 January 2013 (UTC)
It is more information to track it down. For example the Gardiner book (1957) is listed on the Griffith Institute web site (as a 1996 reprint) [1]. Some places in Wikipedia say the Griffith Institute (part of Oxford University) are the publishers. I'm looking at my copy (mine's a 2007 reprint) and it says the Griffith Institute are the (c) holders, but UP, Cambridge are the printers. Confused? So am I. Are printers also publishers? Griffith Institute is on the binding - does that make them the publishers? So regardless of the actual publisher of Gardiner (although it does need clearing up), take it as an example where someone (i.e. me) looked inside at the place where publisher normally goes (although in small print it actually says 'printed by', not 'published by') and saw UP, Cambridge - a reputable publisher. And possibly that is wrong. So the reader then does not have the real publisher to help track down the book - or assess the reputation. Aarghdvaark (talk) 18:43, 28 January 2013 (UTC)
It sounds like you want a field for "published on behalf of". That is very different from the copyright holder. (For instance, most fiction and poetry has its copyright held by the author despite being published commercially.) —David Eppstein (talk) 18:45, 28 January 2013 (UTC)
Printers are not necessarily publishers; they are contractors engaged by the publisher. Some publishers carry out most of their own printing, but may also engage a contract printer if the demand for a particular title cannot be met in-house. One printer may handle work for more than one publisher; conversely, one publisher might use more than one printer.
If a book is libellous, it is the publisher that you sue, not the printer. --Redrose64 (talk) 20:02, 28 January 2013 (UTC)
In this particular case, the publisher of the book is not obvious from looking inside the cover, but the (c) holder is clearly marked. Hence a knowing the (c) holder could help a reader track this book down (although this is a famous book, so it wouldn't actually be hard to find). I could put the (c) holder instead of the publisher, as RedRose64 recommends, but these tags are supposed to be accurate. Aarghdvaark (talk) 00:11, 29 January 2013 (UTC)
Decided that where the publisher is not stated, but the printer and the (c) holder are, then the (c) holder should go into the publisher tag and the actual printer should be ignored, as Redrose64 suggested. But yes, as David Eppstein said above, what I was really looking for was a "published on behalf of" tag. Interestingly, I found out later the book I was trying to reference has its own wiki article: Egyptian Grammar: Being an Introduction to the Study of Hieroglyphs. Aarghdvaark (talk) 03:07, 8 February 2013 (UTC)
Citing newspaper ads and editorials
In a recent edit to Toronto subway and RT, I wanted to cite an advertisement and an editorial in old issues of a newspaper whose title has changed since they were published. I wanted to use ((cite news)), but this does not allow for citing things in a newspaper that are identified as not being news articles. To produce what I considered to be the correct appearance:
7. The Toronto Subway Referendum" (editorial), Toronto Daily Star, December 1, 1945, p. 6
8. "Rapid Transit for Toronto" (TTC advertisement), Toronto Daily Star, December 12, 1945, p. 26
I had to do it "by hand":
<ref name="editorial">"The Toronto Subway Referendum" (editorial), ''[[Toronto Star|Toronto Daily Star]]'', December 1, 1945, p. 6</ref><ref name="referendum-ad">"Rapid Transit for Toronto" (TTC advertisement), Toronto Daily Star, December 12, 1945, p. 26</ref>
I suggest that ((cite news)) should take an additional parameter or parameters to allow these types of items to be properly cited. The minimal change would be a "type=" parameter, consistent with ((cite)), so that "type=foo" would expand as "(foo)" after the title. Alternatively, something like "type=advertisement|advertiser=TTC" might be preferred.
Or if there is already a way to do this with ((cite news)), perhaps it needs to be explained somewhere. I tried doing it with ((cite)) but this produced a different format, with the title in italics instead of quotes.
--69.158.92.109 (talk) 19:48, 4 February 2013 (UTC)
as a[n] (unsatisfactory) stop-gap, use the undocumented parameter |department= in ((cite news)). it displays after the title item. 65.88.88.127 (talk) 20:11, 4 February 2013 (UTC)
To enable type in ((cite news)), a fix will have to be made to ((citation/core)), as type is disabled when one of the periodical parameters is defined.
department in ((cite news)) is documented under the periodical parameters. --— Gadget850 (Ed)talk 20:18, 4 February 2013 (UTC)
you are right, it is(/was?) undocumented in ((cite journal)). the doc is wrong, it does not display after work, but after title, which i think is the correct format. 65.88.88.127 (talk) 20:31, 4 February 2013 (UTC)
Fixed: department displays after title. --— Gadget850 (Ed)talk 23:44, 4 February 2013 (UTC)
Sheesh. Who knew that the none of the lists of parameters at the top of the documentation, including the one titled "Full parameter set", actually include all the valid parameters? Thanks for the suggestion, but "department" is documented as being for another purpose and I'm reluctant to use it in this way unless the documentation is changed. --69.158.92.109 (talk) 09:48, 6 February 2013 (UTC)
the doc is non-intuitive because a single doc template is used to document citation templates used for widely different types of sources, media, and purposes. so in any given citation template doc you have a lot of stuff that are irrelevant to the medium/source type/purpose of the citation, while more pertinent stuff, parameters, and parameter explanations are ommitted. personally, i ignore anything that 1. does not help readers understand and verify the citation 2. (distant second) does not clarify the citation, including my treatment of it, for other editors. i am mainly concerned with imparting proper info to readers, and complying with template usage is a non-issue when it conflicts. i use templates, it's not the other way around. but that's just me. 70.19.122.39 (talk) 14:37, 6 February 2013 (UTC)
I have done a lot of work on documentation, but really have not touched those full document sets. You truly can not include each and every parameter in those sets since they are intended to be copied and pasted. For example, you cannot use author and last in the same citation template, as only one can work. The full parameter list for ((cite book)) was updated fairly recently, with parameters like last[n] that will not work if copy/pasted, so it needs to be changed or a note added to explain it.
"the doc is non-intuitive..." Documentation was all over the place, which lead to my creation of ((Citation Style documentation)). If anything should have the hobgoblin of consistency, it is citation templates and their documentation. I have been considering a mini-doc section where the most-used parameters are explained. I am always open to suggestions, so if there are documentation issues, please start a new discussion with pertinent suggestions. --— Gadget850 (Ed)talk 16:58, 6 February 2013 (UTC)
how many times is the obvious going to be pointed out? and what for? to bulk up discussion archives while nothing is done?
i think that this constant talk and all the bureaucratic procedures are there to actually avoid doing something. after all, what is "controversial" or "disputed" or vague about the following:
wikipedia is a general-purpose encyclopedia whose vast majority of readers are nonexperts/nonspecialists
it is an encyclopedia that "anyone can edit". how can it be taken seriously? that's like the "novel that anyone can write" or the "bridge that anyone can build" or the "medicine that anyone can practice".
the only way for it to be taken seriously is to make obvious that the information is correct + pertinent. this is done by making content easy to verify.
so a citation system to aid verification can be desirable and vital.
the wikipedia citation system: throw out everything you know about previous citation systems, because this is supposed to be a citation system for, and edited by, nonexperts (see point 1). this is the first time (afaik) any such citation system has been attempted. all previous citation systems were targeted to specialists, experts, researchers, or committed students of their fields.
it follows the main design principles should revolve around ease of understanding+verification (for readers) and ease of use+existence of expansive options (for editors). these are not the design principles other citation systems use.
once these design principles are firmly established, you can use any helpful element/procedure from other citation systems to speed implementation
assuming you see the validity of the above, you'll just design the proper system for it. imo, asking for the clarification of minutiae or procedure-talk such as "bring it to the right forum", is proof that the fundamental errors (imo again) of the current system are not obvious to you. i am not taking anything away from the work you've done so far and your willingness to do it, which is to be commended. however "fixing" or "enhancing" a system that is fundamentally wrong, is making it more so.
fine, i'll adjust. but when i point out what i think are flaws to the occasional bewildered user, i'd rather NOT have another "discussion" about it. 65.88.88.127 (talk) 16:32, 7 February 2013 (UTC)
Is chapterurl broken?
Can anyone figure out why this chapterurl isn't linking?
Siwek, Magdalena; Henseler, Christina; Broich, Karl; Papazoglou, Anna; Weiergräber, Marco (2012). "Voltage-Gated Ca2+ Channel Mediated Ca2+ Influx in Epileptogenesis". In Islam, Md. Shahidul (ed.). Calcium Signaling. Advances in Experimental Medicine and Biology. Vol. 740. pp. 1219–47. doi:10.1007/978-94-007-2888-2_55. ISBN978-94-007-2887-5. PMID22453990. ((cite book)): Unknown parameter |chapterurl= ignored (|chapter-url= suggested) (help)
oh!!! doh!. thanks. i should have realized that —Chris CapocciaT⁄C 21:59, 13 February 2013 (UTC)
Position of "location" versus "agency" fields
As of this writing, the Cite news template has the location field follow the agency field, rather than the newspaper/work field (see the "No author but sourced to a news agency" example in the documentation), which is confusing, since per the documentation the location field refers to the location of the newspaper ("[g]eographical place of publication"), not that of the news agency.
I agree, this has needed fixing for a long time. The present situation is obviously illogical and wrong. (One of a number of such things often mentioned but never fixed.) -- Alarics (talk) 13:19, 15 February 2013 (UTC)
Please provide examples of current format and how you think it should be. --— Gadget850 (Ed)talk 15:31, 15 February 2013 (UTC)
See Singapore, footnote 62, which says "AsiaOne. Agence France-Presse (Singapore)." when it obviously should be "AsiaOne (Singapore). Agence France-Presse." -- Alarics (talk) 17:52, 15 February 2013 (UTC)
Another: The No author but sourced to a news agency example I referenced above, taken from the template's documentation:
I am not sure if this has been discussed before, but why doesn't this template provide a 'notes=' parameter? For an example: ((cite book |last=Sykes |first=Bryan |year=2001 |title=[[The Seven Daughters of Eve]] |publisher=W. W. Norton |isbn=0-393-02018-5 |pages=291-92 |notes=Sykes discusses the difficulty in genealogically tracing a maternal lineage, due to the lack of matrilineal surnames (or matrinames))). Or, is there another way to add notes to a reference? Note that this "note" is not the same as "quotes", where the quotes would be quoting something from within the reference and "notes" would be noting something about the reference. --Thorwald (talk) 01:37, 13 February 2013 (UTC)
I don't recall ever seeing notes in a citation. Why do you need notes to identify the source? --— Gadget850 (Ed)talk 01:48, 13 February 2013 (UTC)
I don't think you understood my question. Look at my example above. It has nothing to do with "identify[ing] the source"; it is about noting something about the reference/source. --Thorwald (talk) 01:52, 13 February 2013 (UTC)
I understand the question, but question your assumption that citations should be annotated. The citation identifies the source material ins such a manner that the source can be located for verification. As best I see it, the note in your example should be in the content with the citation to support it. --— Gadget850 (Ed)talk 02:36, 13 February 2013 (UTC)
If the note is about the text and the source supports it, then I place the note within the footnote in front of the Citation, if the note is about the edition, then I place it in parentheses within the footnote, after the Citation.
Examples:
<ref>2011 Census Village code = 621105, ((Cite web|title=Reports of National Panchayat Directory: List of Census Villages mapped for: Hoskera Gram Panchayat, Shahapur, Yadgir, Karnataka|publisher=Ministry of Panchayati Raj, Government of India|url=http://panchayatdirectory.gov.in/adminreps/viewGPmapcvills.asp?gpcode=221153&rlbtype=V))</ref> yields 1. 2011 Census Village code = 621105, "Reports of National Panchayat Directory: List of Census Villages mapped for: Hoskera Gram Panchayat, Shahapur, Yadgir, Karnataka". Ministry of Panchayati Raj, Government of India.
<ref>((Cite book|author=Kamath, Suryanath U.|year=1980|title=Concise history of Karnataka from pre-historic times to the present|location=Bangalore|publisher=Archana Prakashana|page=106|oclc=7796041)) (revised English version of his (1973) ''Karnatakada sankshipta itihasa'')</ref><ref>((Cite book|author=Kamath, Suryanath U.|year=1980|title=Concise history of Karnataka from pre-historic times to the present|location=Bangalore|publisher=Archana Prakashana|page=106|oclc=7796041)) (revised English version of his (1973) ''Karnatakada sankshipta itihasa'')</ref> yields 2. Kamath, Suryanath U. (1980). Concise history of Karnataka from pre-historic times to the present. Bangalore: Archana Prakashana. p. 106. OCLC7796041. (revised English version of his (1973) Karnatakada sankshipta itihasa)
for the 2nd example, assuming the 1973 version is the original, and guessing the language to be Hindi, i would do this: Kamath, Suryanath U. (1980) [originally published 1973 in Hindi as Karnatakada sankshipta itihasa]. Concise history of Karnataka from pre-historic times to the present (revised English-language ed.). Bangalore: Archana Prakashana. p. 106. OCLC7796041. 70.19.122.39 (talk) 13:57, 19 February 2013 (UTC)
Lua coming for more testing
On 18 February 2013 (Monday), the Scribunto interface for Lua script modules is planned to be installed on English Wikipedia (announced in wp:PUMPTECH). No cite templates will be affected during the first week. This first week is only for installation testing, and editors have been asked not to change any live templates, in case Lua must be removed for adjustments. However, next month, we need to consider changing some minor cite templates (not yet {cite web} or {cite book} ) to use Module:Citation (coming soon), which supports almost all current parameters, to format just a few cites, for initial use. As confidence grows, then other cite templates can be changed to use Lua script. Meanwhile, it appears, with the planned Lua design, that a future change to any Lua-based cite will require reformatting of all 1.7 million articles which have been using Template:Citation/core. Hence, we need to think about splitting the Lua-based templates, into test groups, such as:
Module:Citation - originally planned to handle all cite templates
Module:Citation/sandbox - to test Lua changes which might derail cite templates
Template:Cite_web/markup - the current markup-based template, for comparison
Template:Cite_book/markup - the current markup-based template, for comparison
Template:Cite_journal/markup - the current markup-based template
As Lua-based templates are installed, the current template names will be reused: the original Template:Cite_encyclopedia would be rewritten to simply #invoke the Lua Module:Citation, and an old Template:Cite_encyclopedia/markup could contain the prior markup-based template, to compare if new Lua features failed to support the same parameters. Another issue: the Lua-based cite templates have been implemented to again include the COinS metadata, but the overall size and speed of the Lua-based templates will be much smaller and faster than when COinS data was added by markup-based templates. As format differences are fixed, then more cite templates can be switched to use the Lua-based variations. -Wikid77 (talk) 01:03, 16 February, 11:29, 19 February 2013 (UTC)
Working on Lua to show same cite format: There have been some differences in the format of the Lua-based cite templates. Currently, {Cite_web/Lua} shows the equivalent for {Cite_web}:
Cite_journal: Doe, John (5 May 2009). "My Page". II (3). London: Acme: 6. Retrieved 8 Febrary 2013. ((cite journal)): Check date values in: |accessdate= (help); Cite journal requires |journal= (help)
Cite_web now: Doe, John (5 May 2009). "My Page". London: Acme. p. 6. Retrieved 8 Febrary 2013. ((cite web)): Check date values in: |accessdate= (help)
Cite_web/lua : Doe, John (5 May 2009). "My Page". London: Acme. p. 6. Retrieved 8 Febrary 2013. ((cite web)): Check date values in: |accessdate= (help)
To get parameters to match the same format, then lines of Lua script must be moved around into the same order. -Wikid77 (talk) 11:29, 19 February 2013 (UTC)
There is a similar issue with handling of the "origyear" parameter as shown by:
Thanks, I have switched us to Lua Module:Citation/CS1 to adjust for CS1 format, because other editors perhaps want the initial Module:Citation to show another style for origyear and such. Formerly, page numbers listed near the end, so I have shifted them back near the end, to match CS1. Values for volume=II and issue=3 are shown now, although {cite_web} had ignored those parameters. The results in Lua are very close now. -Wikid77 (talk) 14:54/17:56, 19 February 2013 (UTC)
Bug/typo in all cite templates
Hi, I noticed that seemingly all the "cite" templates have two instances of (({seperator))} when (({separator))} is clearly intended. I haven't tested what effect fixing the typo would have, so I won't implement it – I'll leave it to someone who's more familiar with the templates.
Seperately Separately, it might be nice to wrap all the templates in <includeonly>; right now it looks like the documentation on most template pages is preceded by a stray period (except template:cite web, which does have an includeonly wrapper: [2]).--Father Goose (talk) 21:48, 24 February 2013 (UTC)
What you are seeing is (({separator|(({seperator))}, which means if there is a parameter named separator, then use it else if there is a parameter named seperator, then use it. The latter is a common mispelling; it was added before I got into these templates, but I am sure there was a reason, and it doesn't really hurt anything. These duplicated parameters are known as aliases, and are heavily used in the citation templates for a variety of reasons. Most are documented, but this one is for a spelling error and I did not see any reason to tell editors they can use it.
I hate wrapping a template in <noinclude> when there isn't a real reason. If someone makes a mistake, it is not at all obvious when the template doesn't show. ((Cite web)) has it because there is a check for title and it shows a red error message. --— Gadget850 (Ed)talk 22:44, 24 February 2013 (UTC)
I see. Thank you.--Father Goose (talk) 06:11, 25 February 2013 (UTC)
Slightly smaller typeface for the "retrieved" date?
Citations using cite web may have a date or an accessdate, or both, specified within each entry. The date may appear near the beginning of the citation if author details are present, or near the end of the citation if no author details are shown. In many cases there may be a mix of date styles such as February 27, 2013 and 27 February 2013 within a list of citations.
When citation lists are presented as two or more columns, and the text of each citation therefore wraps to multiple lines, I don't find it easy to look through the list. I am sometimes looking for an entry with a particular date, or looking for entries before, or after, a particular date. I often find I am looking at the "retrieved" date rather than the article date, or vice versa. The accessdate is dependant on the editor adding the entry, and is not a part of the data created by the author or publisher of the original article.
I notice that several non-English Wikis have already overcome this difficulty in a very simple manner. They present the "retrieved" date (and the preceding word) in a slightly smaller typeface than the rest of the citation text (but still quite a bit larger than the "subscription required" or "registration required" text that appears on some entries).
Could this simple change be considered here in the English Wiki?-- 86.136.109.181 (talk) 20:55, 27 February 2013 (UTC)
We do wrap the accessdate text in a span with a CSS class attached, which you could use your browser (or a login plus your personal CSS) to even hide if you wanted. I don't think for IPs that this will change otherwise, however. (No comment on whether it should or should not.) --Izno (talk) 22:07, 27 February 2013 (UTC)
Also {cite book/lua} but timeout problems: I have also added ((Cite book/lua)); however, there have been some serious problems with the Lua-based cites hitting the 10-second Lua timeout error (and showing cites as "Script error"), so they are not yet reliable for use in large articles. Others have reported Lua problems with many templates per page. We need to re-focus on faster markup-based templates, in case Lua cites cannot be used in large articles. -Wikid77 (talk) 22:23, 2 March 2013 (UTC)
Cite web - full parameter
Hi! At Template:Cite web I noticed that the "Full parameter set" (either horizontal or vertical format) does not include parameters regarding the editor(s), but it does discuss the issue in detail (Template:Cite_web#Editors). I tried to use editor-first and last - and they actually did work. Shouldn't they be included in the full set? Sorry, if I asked something stupid here! Zoli79 (talk) 16:03, 2 March 2013 (UTC)
I have updated a lot of the CS1 documentation, but I have done only minor work on the copy/paste parameters sets. Please update as desired. --— Gadget850 (Ed)talk 16:45, 2 March 2013 (UTC)
In the ((cite episode)) template, there are several deprecated parameters, such as "serieslink" and "episodelink" which are listed under "Parameters" as deprecated; however, they are still used in the copy & paste example at the top of the /doc page, under "Usage: Full parameter set in horizontal format". They really shouldn't be listed there, if they are deprecated. This needs to be updated. --Funandtrvl (talk) 17:18, 2 March 2013 (UTC)
See previous discussion. --— Gadget850 (Ed)talk 18:40, 2 March 2013 (UTC)
When a journal article has two years of publication, one electronical and one print year
It is very common these days that articles are published online before the printed version, often in the year before the printed version is published. In such cases, an article may be cited using only the year and the digital object identifier, e.g.
Doe J (2010). "Article title". Journal name. DOI: 00000000000
Then the printed version may be published in the following year (2011), and Vol., No. and pp. become available to be used in citations. Also after the printed version has been published, it's common to continue citing the article using the original (first) year of publication.
How can this problem be solved when using this template? In such cases, the citation should ideally include both original year of publication (ahead of print) and the citation of the printed article. Bjerrebæk (talk) 19:58, 19 February 2013 (UTC)
origyear: Original publication year; displays after the date or year. --— Gadget850 (Ed)talk 20:25, 19 February 2013 (UTC)
You should give information pertaining to the edition which you actually consulted. There may be differences between the editions. --Redrose64 (talk) 20:41, 19 February 2013 (UTC)
There are seldom any differences between a version published online before print and the printed version of the same article. It is sometimes useful to have the actual full, correct citation, e.g. in a list of publications. Bjerrebæk (talk) 15:29, 4 March 2013 (UTC)
I have always taken "original year of publication" to be that of something that was subsequently republished. The cases referred to here are actually prepublication. And it seems to me that (in the sciences at least) the "real" date of publication is that of the printed item. I think we ought to give this some consideration. ~ J. Johnson (JJ) (talk) 22:16, 4 March 2013 (UTC)
Comparison of markup and Lua-based cites
The plan is to switch {cite_encyclopedia}, soon, to use the Lua-based template, as a first phase, to ensure proper handling of parameters. The output has been tested for the Lua version to match the basic functionality:
Parameters: ((...|
|last=Doe |first=J.B. |authorlink=John Doe|editor-first=Edie |editor-last=Tor
|editor-link=Editor |encyclopedia=My Encyclopaedia |title=Mein Artikel
|trans_title=My Article |url=http://www.aeiou.com
|accessdate=5 March 2013 |language=German |edition=2012 |date=1 December 2012
|year= |month= |publisher=AEIOU Staff |volume=IV |location=Vienna
|id= |isbn=1234567890 |oclc=Oclc 45 |doi=10.DOI_number
|pages=77 |quote=This is a direct quote |ref=harv))
Cite encyclopedia/lua:Doe, J.B. (1 December 2012). "Mein Artikel". In Tor, Edie (ed.). My Encyclopaedia (in German). Vol. IV (2012 ed.). Vienna: AEIOU Staff. p. 77. doi:10.DOI_number. ISBN1234567890. OCLCOclc 45. Retrieved 5 March 2013. This is a direct quote((cite encyclopedia)): Check |doi= value (help); Check |isbn= value: checksum (help); Check |oclc= value (help); Cite has empty unknown parameter: |month= (help); Invalid |ref=harv (help); Unknown parameter |trans_title= ignored (|trans-title= suggested) (help)
Among the slight differences, the option "trans_title" seems more applicable to follow the article title in an encyclopedia, but either way, a user can enter the title of the encyclopedia with an appended translation in square brackets "[__]". The parentheses around the publisher name should probably be removed, as too many items in the curved brackets. -Wikid77 (talk) 18:06, 5 March 2013 (UTC)
format PDF in the Template:Cite journal/doc examples
I am deleting format=PDF from the examples (four instances) in Template:Cite journal/doc. This specification is rarely used and should not be encouraged, as it clutters up the endnote with a "(PDF)" that is wholly unnecessary; the symbol provides the relevant information. Besides being a distraction for the reader, this usage often leads to inconsistent citation style, since other .pdf references in the same article seldom use this parameter. Peter Brown (talk) 17:35, 7 March 2013 (UTC)
The pdf icon isn't displayed when the url does not include the .pdf extension and not all urls do even though they address a pdf document.
—Trappist the monk (talk) 17:44, 7 March 2013 (UTC)
I did not know that. Nevertheless, the use of format=PDF should be discouraged when the .pdf extension is present, as in the Template:Cite journal/doc examples:
Perhaps Trappist the monk or someone else can update the documentation with an example without the .pdf, making clear that format=PDF is only appropriate in such cases. Peter Brown (talk) 18:16, 7 March 2013 (UTC)
Format=PDF is also useful for installations that don't load images and for WP:ACCESSIBILITY reasons. I think Format=PDF should be compulsory for all pdf files. -- Magioladitis (talk) 18:31, 7 March 2013 (UTC)
External link icons are surely not what WP:ACCESSIBILITY#Images and other policies and guidelines have in mind. Wikipedia:Image use policy, for example, requires that images have an image description; this is not applicable to these icons. Anyhow, if and were images in Wikipedia's sense, then, even if format= were used, their use would violate WP:ACCESSIBILITY since they don't have alt text. Peter Brown (talk) 21:34, 7 March 2013 (UTC)
I have for a long time been routinely deleting format=PDF from references when the URL ends in .pdf, but not when it doesn't. Although, I must say I don't really see why people need to be warned anyway that a document is in that format. Surely all browsers nowadays have a PDF reader built in? -- Alarics (talk) 21:44, 7 March 2013 (UTC)
Only Firefox and Chrome have PDF readers built in. Not every computer or tablet can read PDFs out of the box. Not everybody needs to have a PDF reader installed. Also, not all PDFs can be "read" via a screen reader. While many PDFs will work fine, only PDFs that are PDF/UA complaint will be guaranteed to work with screen readers. Only a few non-free products can create PDF/UA and not by default. Anything scanned and turned into a PDF will not work with screen readers. Bgwhite (talk) 23:42, 7 March 2013 (UTC)
Whether or not a tablet can read PDFs, (PDF) and are redundant. If the user doesn't have the PDF software, either indicates that the source is not directly available. Deleting format=PDF when the URL ends in .pdf still makes sense.
Screen readers are another matter, and many of us are not familiar with the technology. Bgwhite, would you explain further? Can screen-reader users follow external links at all? If not, and (PDF) are equally useless, so including format=PDF is of no help. Suppose, though, that they can follow links. Do the icons register the presence of an external link that might be followed? If they don't, some text indication should be provided in connection with as well as with. Peter Brown (talk) 00:31, 8 March 2013 (UTC)
I'm certain that screen readers can follow external links; but I don't know what happens when the target is a PDF file. I shall invite Graham87 (talk·contribs) here, who probably knows more than most people about the capabilities and limitations of screen reader software. --Redrose64 (talk) 09:08, 8 March 2013 (UTC)
Thanks for letting me know about this discussion. Yes, screen readers work with many, but not all PDF documents, as Bgwhite said above. I've always been a stickler for adding format=pdf or similar to links to any PDF files, because I do appreciate a warning that a link is in PDF format (the Adobe PDF plugins are slower than browsers with my screen reader JAWS, and the layout is obviously different). Until I read this conversation, I had no idea that a PDF icon came up automatically when a PDF file is directly linked. Alt text should be added to these icons, which would benefit not just screen reader users but also people who don't use images. Once the alt text is added, the format parameter can be depricated for all direct links to file formats with automatically generated icons, and a bot should go through and clean up all the redundant format parameters. Graham87 09:50, 8 March 2013 (UTC)
T47891 - External link icons should have alt text. --— Gadget850 (Ed)talk 10:33, 8 March 2013 (UTC)
Even if the link icons gained alt text and that rendered the parameter largely redundant, we'd still have the visual inconsistency that some links have an icon and other links will have the text. I'm sorry, but I'll stick to consistently adding |format=PDF to all citations to PDFs (or any other non-HTML-based source). Imzadi 1979→ 10:38, 8 March 2013 (UTC)
That suggests a larger problem with pdfs. We'd serve users better by distinguishing the purely scanned-image pdfs from those that have accompanying text. This applies not just to users of screen readers, but also to bandwidth-limited and mobile users. A scanned-image pdf of a book can be tens of megabytes.LeadSongDogcome howl! 14:25, 8 March 2013 (UTC)
Imzadi1979, your response is puzzling. There is currently an inconsistency in that format= is sometimes used for but never for. Graham87's proposal, providing alt text for both and using a bot to eliminate format=PDF where it is redundant, would get rid of the inconsistency. Are you opposed? Peter Brown (talk) 15:54, 8 March 2013 (UTC)
No, it shouldn't be. |format=PDF is telling the reader that the link is to a PDF, just as |format=MrSid tells a reader that a map is in the MrSid format. Both are potential warnings that additional software may be required to access the link. The fact that the MediaWiki software includes an icon for PDFs is irrelevant to that notice. My point is that if half of the links in an article's references have the PDF icon without "(PDF)" appearing, and the other half lack the icon but contain "(PDF)", there is still an inconsistency in visual display, and I'll err on the side of always including the format notation regardless if the link generates a graphic or not. Imzadi 1979→ 00:45, 9 March 2013 (UTC)
Well, if that view is accepted we should surely abolish the PDF icon altogether since there is no need for both, and it just becomes more irrelevant clutter. -- Alarics (talk) 07:46, 9 March 2013 (UTC)
abolish ALL such icons, not just the PDF. the |format= and its output provide more exacting and understandable information. in the doc, suggest to editors, that the "format" field's data should be linked, at least in first use. i do not understand the argument about "citation style inconsistencies". surely this is an article-style problem that should be fixed at the article, not at the citation system level. 70.19.122.39 (talk) 15:22, 9 March 2013 (UTC)
The use of accessdate when a page is archived or dead
What is the appropriate use of accessdate when a page is archived or dead? My initial feeling was that the accessdate applies to the archival page (if it exists) and otherwise to the "main" page which has been archived or is dead. Is there guidance somewhere? --Izno (talk) 00:34, 19 February 2013 (UTC)
Access date is simply the date the link was accessed, usually the date it was added. There is no requirement to include it, and it can be hidden by registered users. It is mainly useful for a web page that changes frequently, or when there is no discernible publication date. --— Gadget850 (Ed)talk 02:06, 19 February 2013 (UTC)
If the page is dead, and there is an access date, the latter apparently can serve to locate an archive version that most closely matches the date the citation was placed, but such utility is IMHO debatable. If there is an online archive of the article, I think it makes little sense to leave an access date that serves a purpose no longer. -- Ohconfucius ping / poke 02:22, 19 February 2013 (UTC)
I believe I'm the one who argued that, in fact, and applied just that logic today in cleaning up some citations. :) I might agree with the latter sentence, but I'd like thoughts on the below question before I do so. --Izno (talk) 04:54, 19 February 2013 (UTC)
Yes, but for which link? The archive link or the archived link? I suppose it seems a little incredulous to expect the archive link to go down... --Izno (talk) 04:53, 19 February 2013 (UTC)
however, it does happen (at least temporarily). i always archive online sources (when allowed), preferably at WebCite, and have encountered service disruptions several times. 70.19.122.39 (talk) 14:06, 19 February 2013 (UTC)
In my opinion: when the original link is dead, then there is no need for an access date, as the archive has an archive date that will normally not change. Where the original link is still live then the access date is useful if there is no publication date, regardless of whether it is preemptively archive. --— Gadget850 (Ed)talk 10:57, 19 February 2013 (UTC)
If you are adding an |accessdate=, it pertains to the original |url=. The |archivedate= (although being the date that the archiving service grabbed the page and not the date when you viewed it) does a similar job for |archiveurl= - we know that the archived page isn't going to change further. --Redrose64 (talk) 13:30, 19 February 2013 (UTC)
Yes, but our ability to access the archived page might [change]. That's the only Pandora's box I'm opening here. :) If the lesser consensus here disagrees with me, that's fine, and I'll remove (or at least, not add) accessdates when the cited pages are archived. To that end, I'm not sure the below proposal captures the point I'm making. --Izno (talk) 18:09, 19 February 2013 (UTC)
Documentation
This has come up before, so we should tweak the documentation. Current:
accessdate: Full date when URL was accessed; use the same format as other access and archive dates in the citations; do not wikilink. Can be hidden by registered editors.
Proposed:
accessdate: Full date when original URL was accessed; use the same format as other access and archive dates in the citations;[1] do not wikilink. Not required for web pages that do not change; mainly of use for web pages that change frequently or have no publication date. Can be hidden by registered editors.
the reflink [1] doesn't link to a ref here, do you mean the note at the template page? other than that i agree. 70.19.122.39 (talk) 15:13, 20 February 2013 (UTC)
I meant to leave that out for this. It links to an explanatory note on date formats. --— Gadget850 (Ed)talk 15:18, 20 February 2013 (UTC)
Could we maybe add a bit stronger guidance, something along the lines of "use of accessdate is [weakly] discouraged where the page has an archive and associated archive date, but is allowed where desired", per the above discussion? --Izno (talk) 22:11, 27 February 2013 (UTC)
^ abCite error: The named reference date was invoked but never defined (see the help page).
As far as I know, ((cite web)) is the only citation template that requires title= not be blank, e.g.
((cite web | title= | url=http://www.foo.com/ | first=James | Last = Ford )) = Ford, James. http://www.foo.com/. ((cite web)): Missing or empty |title= (help)
((cite news | title= | url=http://www.foo.com/ | first=James | Last = Ford )) = Ford, James. http://www.foo.com/. ((cite news)): Missing or empty |title= (help)
while at the same time cite web, for some reason, allows the URL to be omitted:
((cite web | title=Bob | url= | first=James | Last = Ford )) = Ford, James. "Bob". ((cite web)): Missing or empty |url= (help)
((cite news | title=Bob | url= | first=James | Last = Ford )) = Ford, James. "Bob".
This situation makes no particular sense to me. Shouldn't something called "cite web" require a URL? Also, isn't the lack of a title generally a problematic omission in all cases? For example, using ((cite book)) without naming the book seems silly.
In working on the Lua migration of the citation templates, I'm wondering what to do about the error conditions. We can, of course, enforce a rule that ((cite web)) must have a title but not the other templates, if that is what's needed. But that seems rather arbitrary. Do people here have suggestions for what error checks you would like to see? Dragons flight (talk) 04:57, 17 March 2013 (UTC)
I find it sensible that when a URL is produced, there should be a forced title for all of the templates. cite web might be the special case where the title is expected to be an online one (not sure that's relevant).
Forcing a URL though? Not sure about that. Even if it is cite web. --Izno (talk) 13:39, 17 March 2013 (UTC)
((Cite Nuttall)), which relies on this template, is broken (editorlink param). Bob Burkhardt (talk) 22:08, 17 March 2013 (UTC)
I fixed it, but it was never a Lua issue. That template was broken due to an erroneous edit last June. Dragons flight (talk) 22:26, 17 March 2013 (UTC)
the replacement of the modular ((cite encyclopedia)) application/((citation/core)) backend with the present iteration of the monolithic citation/cs citation/cs1 script is a mixed bag.
i didn't test enough to see if there are coding bugs (this should be done by the developers prior to deployment. syntax flaws are UNACCEPTABLE). several glaring logic bugs (the nonsensical dependencies among parameters) have been happily removed. some design bugs, namely the equally nonsensical unavailability of the full parameter set to the citation class have also been fixed. unfortunately other, equally glaring and potentially show-stopping design bugs remain. the most serious is the fact that the porting retains the parameter ambiguity of the previous system when it comes to args.title. this may mean either (a) title a work fragment (eg periodicals) or (b) title of the whole work (eg book). this also unacceptable ambiguity leads to uneccessary and inefficient workarounds of the args.booktitle kind.
as with all bugs it can lead to unexpected results such as below:
Author (2013). "Article". In Editor (ed.). Encyclopedia (Type). ((cite book)): |last= has generic name (help); Invalid |ref=harv (help)
as you see, the positioning and punctuation of type in the encyclopedia citation is wrong. this is because type is coded to come after (work) title. but in the encyclopedia/periodical case it instead appears after (fragment) title. the software has no routine to distinguish between the two. such a routine shouldn't be written; instead the parameters should be renamed and the relevant doc be rewritten so there is no semantic ambiguity.
my current evaluation of the lua-based citation implementation is neutral. 70.19.122.39 (talk) 12:43, 18 March 2013 (UTC)
Re title: I think you are referring to the title issue described in Module_talk:Citation/CS1#Weird_link_formatting_for_encyclopedia_class this discussion, where the old version used title in an odd manner that caused odd output. This has been fixed. I have reported the issue with type (which was not supported in the old version). I will work on the documentation. --— Gadget850 (Ed)talk 14:16, 18 March 2013 (UTC)
i'm referring to the fact that title may mean either a work title (args.BookTitle=args.Title) or a fragment title (see section "-- Account for the oddity that is ((cite conference)), before generation of COinS data."). this messes type up, as evident in the local tcommon routines of Module:Citation/CS1. 70.19.122.39 (talk) 14:36, 18 March 2013 (UTC)
Yes, "title" has an ambiguous meaning that usually translates as "article title", but sometimes translates as "encyclopedia title" or similar. We can't standardize that without breaking the existing base of citations, so it won't be fixed. Regarding "type", it has never been on the list of parameters expected by "cite encyclopedia", so it isn't something we test against (same with "series" and "issue" and "agency", and other options that aren't even supposed to be used for cite encyclopedias). We don't currently check whether the user submitted an invalid parameter set with extra options, though we might in the future. Dragons flight (talk) 15:08, 18 March 2013 (UTC)
P.S. If people want to add additional options, like type=, to classes that don't currently use them then we can work on making sure that they will look right; however, it would be up to users to define what parameters they want to support. Dragons flight (talk) 15:12, 18 March 2013 (UTC)
nope, this is wrong. it is much more important to fix meaning ambiguous to humans than to guarantee compatibility for software. first, find a way to correct the ambiguity even if it takes all your time and resources. THEN you can proceed with deployment of the new system without the old bugs. this is a unique opportunity to make the citation system understandable and logical. your way consigns a vitally important function of wikipedia to continuing confusion.
what set of parameters had been "expected" by ((cite encyclopedia)) is irrelevant. expected by whom? and when? if adding the full set of params to citation classes will not unduly break performance, then they should be there, period. and the code should properly present any and all parameters to readers. let the editors decide what to use, don't dictate to them. 70.19.122.39 (talk) 15:42, 18 March 2013 (UTC)
"Title" is accepted as an ambiguous word in the printed style manuals that inspired the cite templates. For example, Chicago Manual of Style 16th ed. p. 661 states "titles of larger works (e.g., books and journals) are italicized; and titles of smaller works (e.g., chapters, articles) are presented in roman and enclosed in quotation marks." If you want to change the parameter names now, it is up to you to think of a way to automatically change all the existing instances of the templates in the encyclopedia, and change all the software that helps people create citation templates. If you can't do that, or you can't find a way to live with the ambiguity, you will have to abandon your project. Jc3s5h (talk) 16:29, 18 March 2013 (UTC)
The old method is just plain wrong. I discussed this before with little response and let it go. I should have fixed it then and there and we would not be here. I think we should treat title as the quote marked included title and fix uses as they come up. I also think we should do the same with the old template, which may be reused on other wikis. --— Gadget850 (Ed)talk 16:52, 18 March 2013 (UTC)
Gadget850, are you suggesting that within the "cite encyclopedia" template, the parameter "title" means the title of an article, but within other types of cite template, such as "cite book", it could mean something else? Because it seemed to me that 70.19.122.39 wanted "title" to always mean the same thing for all citation style 1 templates.
But it already doesn't always mean the same thing, and I think it's too late to change. E.g. in ((cite book)) it's the title of a book while in ((cite conference)) it's the title of an article within a book (and the book title is given by the |booktitle= parameter). —David Eppstein (talk) 17:30, 18 March 2013 (UTC)
(edit conflict)Yes, title varies among templates. The problem is that these templates were developed separately and then updated to use core. When I updated cite conference, I had to use the parameters that were already in use. For encyclopedia, it should mean the article within the encyclopedia and nothing else. --— Gadget850 (Ed)talk 18:11, 18 March 2013 (UTC)
the limitations of the old setup are known. just see the logical + linguistic hoops in the posts above. if in ((cite conference))title means "paper title" why not rename the param paper? it is unambiguous and exactly describes the cited item. similarly in ((cite news)) etc. why not just use article instead of title? an article is what is being cited, and i challenge anybody to prove that editors will understand article to mean something other that "article title". this is/was a unique opportunity to design a citation system that is logical and accessible from the ground up. as well as establish an ancillary manual of style (the doc) for editors and readers that is simple and readily understandable by non-experts, so let's forget the expert-based systems for the moment. a parallel citation system based on lua or whatever could be designed correctly to coexist with the present, and editors should be given guidance to use it for new citations, while (with some automation) the older style is gradually replaced. or, and to continue with title in a very rough example: define eg book as an additional alias for work/title in ((citation/core)) and ((cite book)) (<- the main offender). replace all instances of title in the ((cite book))code with book. automate the replacement of |title= with |book= in article space ((cite book)) citations. then at least you will be left with title unambiguously meaning title of fragment-of-work. and even that should be phased out because we don't cite "titles". we cite works that may be articles, books, films, signs, etc. 70.19.122.39 (talk) 13:38, 19 March 2013 (UTC)
If we were starting from scratch, I would write a specification to work from. But, we have 23 base templates that were independently created. They have been kludged to fit the CS1 style and work pretty well considering.
3Es: engineering, education, enforcement. ((Cite conference)) has been mentioned. It uses title for the included work and booktitle for the main work. It has just under 5000 uses. We could update the parameters, make booktitle throw an error message and use a bot to fix all the uses. But trying to educate all the editors in the new use is going to involve pain and heartache. Or; we could just keep it as is and use the Lua technology to make it work transparently. --— Gadget850 (Ed)talk 14:26, 19 March 2013 (UTC)
i too, enjoy a good joke, but this is not very funny. in this very thread i show how handling of at least one error by the present lua implementation is anything but transparent. now whether the "klutzes" that make up the markup-based CS1 performed well is surely a matter of opinion. discussions in several talk pages about problem after problem (real or perceived) seem to suggest otherwise. the point is that the lua implementation is starting from scratch. it completely does away with the previous setup, using (at this point) templates only as wrappers for calling modules and passing them parameters+data. it could (should) have been designed intelligently to do away with all the bugs plaguing the previous system whether these are bugs in code, logic, process, or design. but right now the implementation, and comments on this thread, show that this is not the case, and it may be hopeless. so much for "lua technology" (?) fixing the citation system, solving the eurozone crisis, and brewing everybody a nice cup of tea. as a sidenote, it is interesting how the first reaction of yourself and others is finding an explanation to deny change requests by editors, instead of finding a way to actually implement them. and most of the editor requests imo are legitimate: that is why many are submitted again and again by different editors through weeks, months, and years. the explanations on the other hand, seem trivial: (a) well, that's how it's been done... (b) some irrelevant manual says so (as if we're publishing for a university press, specialist house, or the new york times) (c) can't do it because some editor might disagree about some vaguely defined item and then what are we going to do... (d) let's have an rfc! no, make that a hundred (e) there have been lots and lots of discussions that had no conclusion, so there (this actually is a nice example of recursion or self-reference, as most of the negative comments in these discussions are by those who then bring this up as an argument) (f) i have a toothache (ok the last is probably original research). fun in la-la land. 70.19.122.39 (talk) 00:18, 21 March 2013 (UTC)
You seem to misunderstand. The Lua implementation of citations is not starting from scratch. One of the design requirements is to fully support the existing parameter set so that existing templates can be migrated to Lua without losing the content of the preexisting citations throughout Wikipedia. That means working around any number of stupid historical choices for parameter naming and function. We can still fix many formatting and logic bugs (and we are working on that; it is still early), but we need to support all of the historical use cases which limits the set of changes that are possible. It was never the plan to make Lua citations completely independent of the currently existing template scheme. While there could be some value in reengineering everything from scratch, and rethinking parameter specifications and such, that is not what we are doing here. Dragons flight (talk) 00:36, 21 March 2013 (UTC)
This is one of the reasons I prefer ((citation)): one set of params to remember instead of 23 subtly-different sets. —David Eppstein (talk) 02:38, 20 March 2013 (UTC)
Page enumeration conflicts
According to the CS1 standard, the options the page=, pages=, and at= can all be used to provide page information, but no more than one of these is ever displayed, with page= overriding pages= and both overriding at=. As part of the Lua migration, we have added the hidden category Category:References with multiple page specifications to any page where a reference tries to simultaneously use more than one of these options. This will allow these conflicting parameters sets to be identified and cleaned up. Dragons flight (talk) 17:49, 19 March 2013 (UTC)
I think it would be better to throw a red error message in such cases, rather than just using a category, that way, the editor adding the superfluous detail is alerted to it. Otherwise, a subsequent editor nay not know which of the values given is correct. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:13, 20 March 2013 (UTC)
I had been considering that as well. I think that both page and pages are added when the editor is including the number of pages in the work, which we don't need. --— Gadget850 (Ed)talk 15:19, 20 March 2013 (UTC)
In the future I would agree with you, but not during the Lua migration. To simply turn on an error message that is likely to hit some few thousand pages (once all citation modes are migrated) seems unnecessarily disruptive. If after the migration the backlog gets cleared, then I think it would make sense to convert this to a visible error. Dragons flight (talk) 15:25, 20 March 2013 (UTC)
I concur, though not for the reasons Editor Andy Mabbett suggests. I've been plodding through the articles listed at Category:Articles with incorrect citation syntax. Finding the broken citations is relatively easy when there is a red error message. You'd think that editors would see the red error text and fix their citations. Apparently not.
—Trappist the monk (talk) 15:33, 20 March 2013 (UTC)
|publication-date= does not pass data to ((harv)) when verbose
pertinent examples are emphasized. i suspect whitespace is the prob.
Markup
Renders as
((harvnb|Author|2000))
((harvnb|Author|2001))
((strong|((harvnb|Author|2002))))
((strong|((harvnb|Author|2003))))
((harvnb|Author|2004))
(Bibliography)
((refbegin))
* ((cite book|ref=harv|last=Author|date=1 January 2000|title=Title))
* ((cite book|ref=harv|last=Author|title=Title|publication-date=2001))
* ((strong|((cite book|ref=harv|last=Author|title=Title|publication-date=January 2002))))
* ((strong|((cite book|ref=harv|last=Author|title=Title|publication-date=1 January 2003))))
* ((cite book|ref=((harvid|Author|2004))|last=Author|title=Title|publication-date=1 January 2004))
((refend))
The five ((harvnb)) are generating links to, in order, href="#CITEREFAuthor2000" href="#CITEREFAuthor2001" href="#CITEREFAuthor2002" href="#CITEREFAuthor2003" href="#CITEREFAuthor2004" whereas the generated anchors are id="CITEREFAuthor2000" id="CITEREFAuthor2001" id="CITEREFAuthorJanuary_2002" id="CITEREFAuthor1_January_2003" id="CITEREFAuthor2004" But the |harv=ref code only uses |publication-date= as a fallback, when both |year= and |date= are absent:
Markup
Renders as
((harvnb|Author|2010))
((harvnb|Author|2011))
((strong|((harvnb|Author|2012))))
((strong|((harvnb|Author|2013))))
((harvnb|Author|2014))
(Bibliography)
((refbegin))
* ((cite book|ref=harv|last=Author|date=1 January 2010|title=Title))
* ((cite book|ref=harv|last=Author|title=Title|date=2011-01-01))
* ((strong|((cite book|ref=harv|last=Author|title=Title|date=January 2012))))
* ((strong|((cite book|ref=harv|last=Author|title=Title|date=1 January 2013))))
* ((cite book|ref=((harvid|Author|2014))|last=Author|title=Title|date=1 January 2014))
((refend))
Author (1 January 2010). Title. ((cite book)): |last= has generic name (help); Invalid |ref=harv (help)
Author (2011). Title. ((cite book)): |last= has generic name (help); Invalid |ref=harv (help)
Author (January 2012). Title. ((cite book)): |last= has generic name (help); Invalid |ref=harv (help)
Author (1 January 2013). Title. ((cite book)): |last= has generic name (help); Invalid |ref=harv (help)
Author (1 January 2014). Title. ((cite book)): |last= has generic name (help)CS1 maint: ref duplicates default (link)
|publication-date= is intended for use in addition to, not instead of, |date=, so when you only have one, it's best to use |date=. --Redrose64 (talk) 15:30, 13 March 2013 (UTC)
Correct: It is the date of publication when different from the date the work was written. A real example:
White, T.H. (1941). The Book of Merlyn. University of Texas Press (published 1977). ((cite book)): Invalid |ref=harv (help)
|publication-date= passes data to hrefs when (a) only the year is given (b) the href is input with date formated exactly as the data in the "publication-date" field, or (c) ((harvid)) is used
Author2 (Jan 1, 2012). Title. ((cite book)): |last= has generic name (help); Invalid |ref=harv (help)CS1 maint: numeric names: authors list (link)
this is a bug. pls fix.
the suggestion to use |date= when only the publication date is known presents several problems.
what happens in online sources (eg ((cite web))) when neither work date nor pub. date are known? do we input the access date in |date=?
if not, there's an inconsistency. [when no other date is known, and for verification purposes (the reason citations exist) the access date is the defacto "publication date", which should then subst into |date= per your argument].
on the other hand, if the access date should become the variable data for |date= then the dependency between |accessdate= and |url= should be undone to avoid errors.
and what about archival dates when |deadurl=yes? another clarification should be provided.
secondly, |publication-date= has obvious semantic significance for editors/citation providers. substitution with |date= should not be encouraged. instead |publication-date= should be promoted to the same status as |date= within the template when |date= is absent. similarly for online sources |accessdate= should be promoted when |date= and |publication-date= is absent. and so on for |archivedate=.
or, just do the lazy thing and remove |publication-date= altogether from CS1.
Yes, it does not happen with |publication-date= but it is not a bug; it is a side-effect of the non-use of other parameters. Quite frankly I don't know why we attempt to extract a year from |publication-date= in the first place, since |publication-date= (see documentation) is intended to be supplementary to either |date= or |year= (see documentation).
Access dates must only be placed in the |accessdate= parameter (see documentation), no other parameter is intended for this.
The |ref= parameter (see documentation) is pretty much free-form. It is provided so that a manually-constructed anchor may be attached to the ((cite book)) when the special value |ref=harv does not give a suitable anchor. Use of the ((harvid)) template is not obligatory: it is a tool for constructing an anchor which is consistent with the links generated by templates like ((harv)), ((harvnb)) and ((sfn)). --Redrose64 (talk) 10:58, 14 March 2013 (UTC)
I need to fix the doc for this parameter. Somehow I thought the field did not display if year or date were not defined. --— Gadget850 (Ed)talk 12:46, 14 March 2013 (UTC)
No, that is how it works. If publication-date but not date, then publication-date is used as date and the anchor is formed:
Markup
Renders as
((cite book |last=White |first=T.H. |authorlink=T. H. White |title=The Book of Merlyn |date=1941 |publication-date=1977 |publisher= University of Texas Press |ref=harv))
White, T.H. (1941). The Book of Merlyn. University of Texas Press (published 1977). ((cite book)): Invalid |ref=harv (help)
'"`UNIQ--templatestyles-000000D6-QINU`"'<cite id="CITEREFWhite1941" class="citation book cs1">[[T. H. White|White, T.H.]] (1941). ''The Book of Merlyn''. University of Texas Press (published 1977).</cite><span title="ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=book&rft.btitle=The+Book+of+Merlyn&rft.pub=University+of+Texas+Press&rft.date=1941&rft.aulast=White&rft.aufirst=T.H.&rfr_id=info%3Asid%2Fen.wikipedia.org%3AHelp+talk%3ACitation+Style+1%2FArchive+2" class="Z3988"></span> <span class="cs1-visible-error citation-comment"><code class="cs1-code">(([[Template:cite book|cite book]]))</code>: </span><span class="cs1-visible-error citation-comment">Invalid <code class="cs1-code">|ref=harv</code> ([[Help:CS1 errors#invalid_param_val|help]])</span>
((cite book |last=White |first=T.H. |authorlink=T. H. White |title=The Book of Merlyn |publication-date=1977 |publisher= University of Texas Press |ref=harv))
White, T.H. (1977). The Book of Merlyn. University of Texas Press. ((cite book)): Invalid |ref=harv (help)
'"`UNIQ--templatestyles-000000DA-QINU`"'<cite id="CITEREFWhite1977" class="citation book cs1">[[T. H. White|White, T.H.]] (1977). ''The Book of Merlyn''. University of Texas Press.</cite><span title="ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=book&rft.btitle=The+Book+of+Merlyn&rft.pub=University+of+Texas+Press&rft.date=1977&rft.aulast=White&rft.aufirst=T.H.&rfr_id=info%3Asid%2Fen.wikipedia.org%3AHelp+talk%3ACitation+Style+1%2FArchive+2" class="Z3988"></span> <span class="cs1-visible-error citation-comment"><code class="cs1-code">(([[Template:cite book|cite book]]))</code>: </span><span class="cs1-visible-error citation-comment">Invalid <code class="cs1-code">|ref=harv</code> ([[Help:CS1 errors#invalid_param_val|help]])</span>
((cite book |last=White |first=T.H. |authorlink=T. H. White |title=The Book of Merlyn |publication-date=1 May 1977 |publisher= University of Texas Press |ref=harv))
White, T.H. (1 May 1977). The Book of Merlyn. University of Texas Press. ((cite book)): Invalid |ref=harv (help)
'"`UNIQ--templatestyles-000000DE-QINU`"'<cite id="CITEREFWhite1977" class="citation book cs1">[[T. H. White|White, T.H.]] (1 May 1977). ''The Book of Merlyn''. University of Texas Press.</cite><span title="ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=book&rft.btitle=The+Book+of+Merlyn&rft.pub=University+of+Texas+Press&rft.date=1977-05-01&rft.aulast=White&rft.aufirst=T.H.&rfr_id=info%3Asid%2Fen.wikipedia.org%3AHelp+talk%3ACitation+Style+1%2FArchive+2" class="Z3988"></span> <span class="cs1-visible-error citation-comment"><code class="cs1-code">(([[Template:cite book|cite book]]))</code>: </span><span class="cs1-visible-error citation-comment">Invalid <code class="cs1-code">|ref=harv</code> ([[Help:CS1 errors#invalid_param_val|help]])</span>
This should work. Let me look at the original examples again. --— Gadget850 (Ed)talk 13:23, 14 March 2013 (UTC)
Here is what is happening: ((harvnb)) is including the year only in the link and ((cite book)) is including the full date in the anchor, causing a mismatch:
Author (Jan 1, 2014). Title. ((cite book)): |last= has generic name (help); Invalid |ref=harv (help)
[[#CITEREFAuthorJan_1,_2014|Author & Jan 1, 2014]]<span class="error harv-error" style="display: none; font-size:100%"> harvnb error: no target: CITEREFAuthorJan_1,_2014 ([[:Category:Harv and Sfn template errors|help]])</span> '"`UNIQ--templatestyles-000000EA-QINU`"'<cite id="CITEREFAuthor2014" class="citation book cs1">Author (Jan 1, 2014). ''Title''.</cite><span title="ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=book&rft.btitle=Title&rft.date=2014-01-01&rft.au=Author&rfr_id=info%3Asid%2Fen.wikipedia.org%3AHelp+talk%3ACitation+Style+1%2FArchive+2" class="Z3988"></span> <span class="cs1-visible-error citation-comment"><code class="cs1-code">(([[Template:cite book|cite book]]))</code>: </span><span class="cs1-visible-error citation-comment"><code class="cs1-code">|last=</code> has generic name ([[Help:CS1 errors#generic_name|help]])</span>; <span class="cs1-visible-error citation-comment">Invalid <code class="cs1-code">|ref=harv</code> ([[Help:CS1 errors#invalid_param_val|help]])</span>
At this point, we can debate whether to include the full date in ((harvnb)) or to change the cite templates to include only the year in the anchor. --— Gadget850 (Ed)talk 13:43, 14 March 2013 (UTC)
It would be a massive change to put a full date into ((harvnb)) and related templates; people are very much used to using the year alone. As things stand, the cite templates already include only the year in the anchor if you use |date=. It is only |publication-date= where there is inconsistency. See my post of 15:30, 13 March 2013
If we should build the harv anchor from |publication-date=, we should do so in the same manner as |date=. That is, instead of:
|Year=(({year|(( <!-- attempt to derive year from date, if possible -->
#if: (({date|))}
|((
#iferror:((#time:Y|(({date|))} ))
|((#iferror:((#time:Y|(({publication-date|einval))} ))||((#time:Y|(({publication-date|))} ))))
|((#time:Y|(({date|))} ))
))
|(({publication-date|))} <!-- last resort -->
))
))}
we should put
|Year=(({year|(( <!-- attempt to derive year from date, if possible -->
#if: (({date|))}
|((
#iferror:((#time:Y|(({date|))} ))
|((#iferror:((#time:Y|(({publication-date|einval))} ))||((#time:Y|(({publication-date|))} ))))
|((#time:Y|(({date|))} ))
))
|((#iferror:((#time:Y|(({publication-date|))} ))||((#time:Y|(({publication-date|))} )))) <!-- last resort -->
))
))}
I agree. I just got publication-date added to the Lua version, and it needs some more fixes including the anchor, so this is a good time for that. --— Gadget850 (Ed)talk 14:27, 14 March 2013 (UTC)
yes, this is a good solution imo, and goes some way into normalizing date output across CS1. i still have problems with the rationale imposing limits on the use of access date when sources are online (eventually ALL sources may be online). the doc reflects these limitations. imo both should be changed, but that's another issue. Redrose, i don't mean to nitpick, but when software does not perform as expected, eg forming anchors in unpredictable fashion, then it is buggy software. but that's also another matter. 70.19.122.39 (talk) 14:45, 14 March 2013 (UTC)
This would break any instances where someone used a full date in a harv template, but I would expect this to be nonexistent or rare. Do need to update the anchor documentation to include publication-date. --— Gadget850 (Ed)talk 15:03, 14 March 2013 (UTC)
It would only break the harv linking if:
the harv template had used a full date (the docs explicitly state that the year of publication should be given)
and had entirely omitted |year= (not just left it blank)
Pretty big set of ifs. --Redrose64 (talk) 15:50, 14 March 2013 (UTC)
I searched the archives and don't see this was every reported before, so I think it is pretty safe to go forward. --— Gadget850 (Ed)talk 16:21, 14 March 2013 (UTC)
|Year=(({year1|(({year|(( <!-- attempt to derive year from date, if possible -->
#if: (({date|))}
|((
#switch: ((#time:Y|(({date|))))}
|Error: invalid time = (({publication-date|))}
|((#time:Y|(({date|))))}
))
|(({publication-date|))} <!-- last resort -->
))
))))))
this should be fixed too. you'd think this routine would be in ((citation/core)) since apart from ((harvard citations/core)) and ((cite book)) it applies to ((cite web)), and any template that uses |publication date=. nah, that'd have made sense. 70.19.122.39 (talk) 00:49, 22 March 2013 (UTC)
How to Cite eBooks
I have issues citing eBooks, in particular when referencing a page, which may vary by varying the font size. In Kindle the exact location can be expressed with a number called Location.
So maybe some rules and/or parameters are necessary, Or not? Carlotm (talk) 11:44, 20 March 2013 (UTC)
Thanks Gadget850. So I suppose a citation like the following one would be fine.
Epstein, Catherine (2010). "Model Nazi: Arthur Greiser and the Occupation of Western Poland". New York: Oxford University Press. location 2772-2778 (Kindle Edition). — Preceding unsigned comment added by Carlotm (talk • contribs) 07:54, 21 March 2013 (UTC)
Using CS1:
Markup
Renders as
((cite book |last=Epstein |first=Catherine |year=2010 |title=Model Nazi: Arthur Greiser and the Occupation of Western Poland |location=New York |publisher=Oxford University Press |at=location 2772-2778 (Kindle)))
Epstein, Catherine (2010). Model Nazi: Arthur Greiser and the Occupation of Western Poland. New York: Oxford University Press. location 2772-2778 (Kindle).
edited to finish what I started; guess I must have inadvertantly clicked save
I just replaced this citation a ((cite pmid|23019641)) with a ((cite journal)) equivalent at R. Duncan Luce. This page was listed at the end of all of the Template: pages at Category:Articles with incorrect citation syntax. None of the usual error messages were displayed and none of the other citations seemed to have anything wrong with them so I took a chance and replaced the ((cite pmid)).
So I guess my questions are: how did ((cite pmid)) cause this error? Why wasn't there an error message? Why was the R. Duncan Luce page listed with the Templates?
—Trappist the monk (talk) 14:31, 21 March 2013 (UTC)
It redirects to ((Cite doi/10.1126.2Fscience.1229851)), but I don't see anything there. --— Gadget850 (Ed)talk 14:32, 21 March 2013 (UTC)
Multiple ISBNs
[The key point of this request has been bolded big throughout to -- yes -- shout in desperation how often it's been repeated to no avail.] --EEng (talk) 14:57, 20 March 2013 (UTC)
Is there some way in cite book etc. to indicate multiple or alternative ISBNs e.g. paperback vs hardcover? Thanks. EEng (talk) 21:05, 16 March 2013 (UTC)
This is asked quite often; full answers are in the archives, but in brief: give the ISBN for the single edition/version which you actually consulted. --Redrose64 (talk) 21:37, 16 March 2013 (UTC)
That seems like a cop-out. Part of the purpose of ISBN is to allow the reader to find a copy he can consult for himself, so assuming that e.g. paperback and hardcover have identical content, it's useful to include both. Maybe there's a "misc info" parameter I can just stick the other ISBN into as free text? EEng (talk) 22:59, 16 March 2013 (UTC)
As a source we only need the one version - That said let me help - I do it like the following for bibliographies - not sure its ok but been doing it for 7 years - you dont need ISBN's in the ((cite book)) template itself to work - <ref>ISBN 978-0-8020-5016-8</ref> = ISBN 978-0-8020-5016-8
So this coding that has the ISBN's outside the ((cite book)) but with in the <ref> </ref> parameter will work as a reference.
<ref>((cite book |last = Taylor |first = Martin Brook|coauthor= Owram, Doug|year =1994|title =Canadian History|volume=|publisher= University of Toronto Press)) Hard cover pp 12, ISBN 978-0-8020-5016-8 and Paperback pp 24, ISBN 978-0-8020-2801-3</ref>
Will render this...
Taylor, Martin Brook (1994). Canadian History. University of Toronto Press. ((cite book)): Unknown parameter |coauthor= ignored (|author= suggested) (help) Hard cover pp 12, ISBN 978-0-8020-5016-8 and Paperback pp 24, ISBN 978-0-8020-2801-3
That's what I was doing -- was hoping for something more elegant. But thanks. EEng (talk) 23:30, 16 March 2013 (UTC)
Except that "pp" is the abbreviation for "pages", and since only one page number is given, it should be Hard cover p. 12, ISBN 978-0-8020-5016-8 and Paperback p. 24, ISBN 978-0-8020-2801-3 --Redrose64 (talk) 23:50, 16 March 2013 (UTC)
And that stuff outside the template will not render metadata in the new Lua templates as we migrate. --— Gadget850 (Ed)talk 05:06, 17 March 2013 (UTC)
Millions and millions of entries not in templates - any plans to address this?Moxy (talk) 05:30, 17 March 2013 (UTC)
No. ISBN is not required by any major citation style, and is not completely reliable as a unique identifier. We could provide the capability to add 16 different ISBNs, but it would not directly identify the particular edition used in creating the citation. --— Gadget850 (Ed)talk 13:43, 17 March 2013 (UTC)
Pardon my ignorance, but:
In what way is ISBN not completely reliable as a unique identifier? If you're going to say that there are occasional glitches such as the same number issued twice, or multivolume works assigned a single number, then by that token we might as well not bother with any kind of identifier of anything.
What does it matter that no major citation style requires ISBN? Here at WP it's recognized as a highly useful thing to include, so it's worth giving attention to making its inclusion as easy as possible.
Well, an editor should be specifying the exact edition of a work being referenced. The pagination and typesetting could differ between editions printed in the UK and and the US, for instance. The pagination normally differs between hardcover and paperback editions of the same edition of a book. An editor could be consulting a revised or updated edition as well. Each of these editions will have differing ISBNs, and the principle is to cite the actual source consulted. Imzadi 1979→ 15:58, 17 March 2013 (UTC)
I understand all that, and certainly any provision for multiple ISBNs should include some way to single out the particular ISBN that was consulted, because the core imperative is that the reader be certain he's looking at precisely the work to which the citation refers. It's still useful to include additional ISBNs, with the caveat that different ISBNs may be subject to the sorts of variation you mention. On the other hand, it's often the case the a publisher will explicitly state that two ISBNs are the same page images in different bindings, and it would be useful to be also able to designate those when that's known. EEng (talk) 17:19, 17 March 2013 (UTC)
Good— you have identified the issues with ISBN. JSTOR, doi and others use well verified systems with a central point of authority.
The purpose of a citation is to help the reader identify the source used to make a point in the content. If you have a pressing need to identify versions other than the one used, then put it in a further reading section. Alternate sources that were not consulted do not belong in references. --— Gadget850 (Ed)talk 22:38, 17 March 2013 (UTC)
Oh, honestly! Pressing need? It's amazing the amount of verbiage being wasted to characterize as illconceived something that's obviously useful. We've even had one person interject a correction about pp vs p, as if somehow anyone here needed that pointed out.
Let me state the situation again: Frequently a work is issued in paperback and hardcover, utterly identical except for the binding. Sometimes I'll have both on my bookshelf, so it doesn't make sense to talk about "the one actually consulted." It would be stupid to have two separate entries, one for the paperback, one for the hardcover. I could arbitrarily pick one or the other to use in the cite, but that unnecessarily reduces the ease with which the reader can find a copy for his own consultation.
So if you can find it in your heart, please make some way accommodate the entirely sensible desire to have two or more ISBNs in the same entry.
The point is that the reader should not be mislead into thinking both books have been verified to support the assertion against which the ref is called out. Hence we would need citations to the intent of:
Author. Title Location:Publisher (1999) ISBN 978-0123456789 pp.3, 5, 7-9. Also available as ISBN 978-9876543210 (not consulted).
Of course the problem then would be how many versions, editions, and printings to enumerate. Instead, we keep it simple and cite just the one consulted. There are many tools, magically linked via wp:Booksources, for finding the others. LeadSongDogcome howl! 17:08, 19 March 2013 (UTC)
I don't see the point of insisting on one ISBN "not consulted" and the other as primary when the person adding the citation has checked that both have identical text. Your comment comes across as not having actually read the previous one. Anyway, can't this all be handled by using |id= instead of |isbn=? —David Eppstein (talk) 17:27, 19 March 2013 (UTC)
Well, of course I read it here on this talk page, but how many article readers, or even editors, are likely to do so? If we start down the path of including multiple ISBNs, we will very shortly have many editors showing several ISBNs on the untested premise that they are equivalent. Even in the rare cases where an editor does have the multiple editions at hand and takes the time to check that they are equivalent, it will still make it more difficult for other editors to locate and verify them all. Still, in those rare cases, it would still be possible to say both were consulted, using free-form citations. The template, however, should reflect the usual case as its default. Unless our standard for wp:V is now "one or more of this cited list of editions supports the in-text statement", it would be a move backwards to list multiple editions (even without considering the visual clutter). So yes, EEng, I hold that it is better to just "arbitrarily pick one or the other" than list both. LeadSongDogcome howl! 21:27, 19 March 2013 (UTC)
If the only difference between two printings of a book is hard or soft cover, and a statement (etc.) can be verified in both (and possibly the editor has verified in both), then it seems a disservice not to provide as much information as possible. But it seems sufficient that in such cases the additional isbns be put into an ((isbn)) template following the citation, as illustrated above. I am doubtful on having supplemental isbn parameters in the citation templates themselves lest some editors think they need be filled in. And similarly for using |id= for that purpose. If there are two identical (except for the cover) editions, equally consulted, then it shouldn't matter which one is the "master" isbn. ~ J. Johnson (JJ) (talk) 23:08, 19 March 2013 (UTC)
((ISBN)) doesn't do what you think it does. ISBNs can use magic linking by simply typing ISBN a space and the ISBN. --— Gadget850 (Ed)talk 00:36, 20 March 2013 (UTC)
LSD, I don't know which alternative state of the universe is the sadder one -- David's, in which you didn't bother to parse the use case offered; or your own, incorporating as it does your strained (or, perhaps, hallucinatory -- if you will pardon my making the inference) logic for the trivialization of the use case -- logic by which we also musn't allow |chapter= together with |chapter_url= (because maybe the editor didn't check that the linked text is true to the original) or |title= together with |trans_title= (because maybe the translated title isn't accurate) or |authorlink=ever (because maybe the linked article is not actually about the author, but rather someone else with the same name). If a new |isbn2= is documented approriately ("For use only where both ISBNs are known to contain identical content at the page-image level e.g. paperback and hardcover editions differing only in their bindings") would we really be opening the door to unprecedented special danger?
If it will help you sleep at night we can call it |isbn2_but_only_if_they_really_have_same_content=, but in truth you are "straining at gnats and swallowing horses", as my Irish great-grandmother used to say. There can be little doubt that many or most ISBNs are not contemporaneously transcribed from the title page of the "work actually consulted" held so sacred here, but rather looked up online, after the fact, once home from the library (or whatever -- you get the idea). Indeed, the scales having now fallen from my eyes about the many potential (if only incipient -- even imagined) instances of insidious inaccuracy indulged by indolent editors, I propose we require that each ISBN be accompanied by a webcam clip of the editor reading aloud the ISBN -- in fact, the entire citation -- straight out of the physical "work consulted", with close-up shot of title page required as well.
The word edition may have been thrown about loosely in this discussion so far, but as has been made abundantly clear we're not talking about distinct editions in the sense of revised content, merely different bindings or printings of identical page content. One frequently sees, especially in titles intended for both library and textbook sales, ISBN J-JJJ-JJJJJJ-J / K-KKK-KKKKKK-K (pbk.) on the title page, which means exactly what one expects it to mean. Of course, there are sometimes minor corrections in the interval between hardcover and softcover issuance, but this happens silently from printing to printing of even the same "edition" (with the same ISBN!) anyway -- if you know what the mysterious runs of numbers like 8 9 10 11 12 13 14 15 16 17 18 19 mean at the bottom of title pages, then you know what I'm talking about. Why aren't you concerned about that?
You say the template "should reflect the usual case as its default". Well, by default |isbn2= will, of course, be null, and that's the usual case.
Finally, your preference that editors "arbitrarily pick one or the other" is, in the context of everything you wrote to that point, contradictory nonsense. If, as you insist, it's hard to know that two ISBNs are interchangeable as the basis of citations, then only one of them is OK -- yet you're suggesting flipping a coin without caring whether you're getting the right one or the wrong one. On the other hand, if picking one or the other arbitrarily is OK, then that must mean you know both of them to be right (i.e. they're identical at the page level), so you may as well help to the reader by including them both.
As to "visual clutter" -- oh, why am I wasting my time? It really makes one think more than twice before bothering to suggest anything. I repeat my earlier comment: It's amazing the amount of verbiage being wasted to characterize as illconceived something that's obviously useful. What a lot of know-it-all bullshit.
((cite book |title=Some Book |ISBN=978-0-306-40615-7 (hardback) ISBN 978-0-306-40615-8 (paperback) ISBN 978-0-306-40615-9 (online) ISBN 978-0-306-40615-1 (sign) ISBN 978-0-306-40615-2 (Tijuana bible) ISBN 978-0-306-40615-3 (comic book) ISBN 978-0-306-40615-4 (graphic novel) ISBN 978-0-306-40615-5 (audiobook) ISBN 978-0-306-40615-6 (eBook)))
Your example above parodies the use case proposed, while the one below illustrates it realistically. And see my further post, as this same indent level, below. --EEng (talk)
Markup
Renders as
((cite book |title=The Rise of Statistical Thinking, 1820–1900 | author=Theodore M. Porter| year=1986|publisher= Princeton University Press |ISBN=0-691-08416-5 ISBN 0-691-02409-X (pbk)))
Not to seem ungrateful, but why couldn't that have just been the answer to my original query? Were the tortures above some kind of hazing process intended to bestow strength of character through adversity? EEng (talk) 04:53, 20 March 2013 (UTC)
Now can you please explain to me just how a reader is supposed to know which of that long list of ISBNs is the pertinent one, which he should check to verify an assertion, or conversely in your special use-case, how is assured that the editions/versions/printings are equivalent? It is hard enough to get people willing to source-check articles from one specific edition, without asking them to do so in several, let alone having to listen to an audiobook to find out if it is the same as the print version. Protest the runaround all you like, but what you propose represents a massive waste of time for other editors and a reduction of quality for readers.LeadSongDogcome howl! 05:30, 20 March 2013 (UTC)
To help you focus, below Gadget's parody example above I've substituted a realistic one. I've also bolded, in earlier posts, the point emphasized ad nauseum from the very beginning, which for good measure I'll repeat here. Get ready. Ready? Here it is:
As has been made abundantly clear we're not talking about distinct editions in the sense of revised content, merely different bindings or printings of identical page content.
Here it is again, this time in bold:
Not distinct editions in the sense of revised content, merely different bindings or printings of identical page content.
One more, time, this time in big type:
Different bindings or printings of identical page content.
I'd be the last to suggest, LeadSongDog, that your username is meant to connote not only possible hallucinations (as previously noted) but prodigious denseness as well -- but can you please get with the program?
There is nothing in policy or common sense requiring the reader to verify that the ISBNs are correct or match each other. Verifiability means that the reader can verify the content of the article. Whether the ISBNs match each other is not part of the content of the article.
The point of supplying the ISBN is to aid readers in finding copies of the book. Multiple ISBNs make this easier, not harder, because there is a greater likelihood that at least one of the supplied ISBNs will work to aid the reader to find the book.
In practice the reader seeking verification is going to find whatever copy of the book is easiest to find (perhaps using the ISBN to do so), and look in it for the content he or she wishes to verify (ignoring the ISBN or indeed whether it's a correct edition, except possibly if it's a really wrong edition and the desired content isn't there). So all your questions about "how is the reader going to do this?" are completely pointless; none of those things are things that the reader is going to want to do.
In some usual circumstance -- perhaps an obscure foreign book with a confusing title -- I could see supplying an ISBN labeled (?) or something like that, if it's felt that would help someone locate the work in a dusty smoke-filled Karachi bookstore. But in general, any supplied ISBN or ISBNs (whether one, or several) should be one thought to match the page numbers etc. given in the citation. EEng (talk) 14:57, 20 March 2013 (UTC)
I don't see the need for this, but you have an answer. And the standard is to use <s> for strking text, not <del>. And <big> is obsolete. Out. --— Gadget850 (Ed)talk 15:06, 20 March 2013 (UTC)
The question previously posed to you, and still not addressed, is why (whether or not you saw the need) the sample markup finally supplied, after exhausting begging, couldn't simply have been given in response to the original request, 30 posts ago? Your fretting about <s> vs. <del> epitomizes the insistence on technical trivia, in lieu of understanding the situation posed, which has characterized this long and wasteful discussion. I repeat: What a lot of know-it-all bullshit.EEng (talk) 16:52, 20 March 2013 (UTC)
Choosing to insult fellow editors rather than work with them is seldom productive. EEng's oft-repeated and now shouted premise that different printings have identical content can for all practical purposes be based only on original research. David evidently doesn't think this matters. While he may be correct, I don't think he is. In any case, that isn't something I need further abuse over. Sorry for trying to answer your question. I'm done. LeadSongDogcome howl! 18:12, 20 March 2013 (UTC)
I just picked up three Dover books from my shelf to check. The first one said it was "extensively corrected"; the second said it was unabridged but had a new introduction; the third said it was "unaltered". I don't see how you can construe that as original research, and anyway the Wikipedia original research policy is supposed to be about content, and you are vastly overreaching by trying to apply it to citation data. My feeling is that in the second and third cases it would be appropriate to list both the original and Dover editions. In this case, it shouldn't be done just by giving an ISBN, since the other publication data (e.g. year) is different, but in some cases (electronic vs print isbn for a single book, or library edition vs textbook edition) just listing the multiple isbns in a single citation does make sense. —David Eppstein (talk) 19:26, 20 March 2013 (UTC)
Verifying an article's citations, page by page, in one ISBN, and then doing that again in a second ISBN, is no more original research than is just doing it in the first one and then stopping.
Reading, from the title page of a work, the words "This edition published simultaneously as ISBN XXX (hbk) and ISBN YYY (pbk)", and interpreting that appropriately, is no more original research than is getting any other bibliographic information off the title page.
"Abuse" can take many forms, and one of them is to waste others' time and goodwill by stubbornly exhibiting a level of cluelessness so extreme that only the most blunt form of correction has a chance of piercing the armor of self-imposed obliviousness -- so blunt, perhaps, that in another context it might qualify as an insult. But in this context, no. There's no post you've made to this discussion that wouldn't be completely laughable were it not for its contribution to muddling the discussion. EEng (talk) 19:39, 20 March 2013 (UTC)
Abuse, he says. I'll show you abuse. This is abuse. I try to help, and that's the thanks I get? I don't know why I fucking bother. I could have said "this has been asked many times before, check the archives" and left it at that. In fact, if you do bother to look through the archives here, or at many of the talk pages for the Citation Style 1 templates, or those of WP:VPT, you can see if I don't go out of my way to help. And don't shout at us to get your point across. It won't work. --Redrose64 (talk) 22:51, 20 March 2013 (UTC)
Well, since careful reason, examples from experience, detailed explanations, and conscientious response to concerns wasn't working, I thought I'd give it a try. Desperate times call for desperate measures. Look, I'm sure you meant to be helpful, but -- please be candid with yourself -- do you really think a reminder that pp. means pages (plural) was what was needed at that point, especially immediately following my post restating that nothing so far had answered my question? EEng (talk) 02:22, 21 March 2013 (UTC)
An adequate answer was provided by Moxie right at the start; the rest has been unneedful bickering. (And shouting.) Time to put a lid on this discussion? ~ J. Johnson (JJ) (talk) 21:46, 23 March 2013 (UTC)
You refer to the answer here, which was immediately followed by the response
That's what I was doing -- was hoping for something more elegant. But thanks.
So apparently not adequate, though well-intended and appreciated. The actual answer here was provided only much later after much begging. Everything after that was continued counterfactual insistence that this perfectly sensible request was somehow bizarre and wrongheaded.
But I do agree it's time to stop. Those who don't get it yet, probably never will.
My two penn'orth - I came across a work today, an encyclopedia in three volumes, each with its own isbn. It was also possible to have all three as a boxed set, no isbn. This was for a general bibliography, not a pinpoint citation, but basically I had to cite all three volumes separately, although it would have been more elegant to have just one citation with three isbns. Must try that isbn-outside-the-cite-book trick.
I am not sure what has happened to ((Cite book)) but any URL not in the URL parameter is no longer working. I understand that having a URL not in the URL parameter is not the way to do things but never the less we have thousands and thousands of links set up out side the normal URL parameter. What has happened anyone working on this problem?Moxy (talk) 18:43, 25 March 2013 (UTC)
Thank you for that link - Looking at that talk looks like people are not aware that at our content guideline we tell editors to link like this see - WP:BOOKLINKS.Moxy (talk) 18:58, 25 March 2013 (UTC)
To be more specific, it is not any URL. It is a URL stuffed into the page=/pages= field or the volume= field, and then only if the URL contains a hyphen ("-"). I suspect the number of broken URLs is actually pretty small, nonetheless we will fix that part. However, having a URL in the pages= or volume= field is incompatible with the COinS metadata generated by the citation templates (and always has been), so we should not be encouraging people to do that. Dragons flight (talk) 19:04, 25 March 2013 (UTC)
Fixed I added a note to BOOKLINKS, but if the practice is already in use, it will continue. --— Gadget850 (Ed)talk 19:30, 25 March 2013 (UTC)
Cite web format and language parameter position in rendered citation
Markup
Renders as
((cite web
|url=http://www.bnp.gob.pe/snb/data/periodico_mural/2009/7/fechas_civicas/28dejulio.pdf
|title=Aniversario de la Proclamacion de la Independencia del Perú
|language=Spanish
|format=pdf
|accessdate=2013-03-21
))
The format text normally follows the pdf icon when |language= isn't used. It seems odd to insert the language between the icon and the file format so perhaps these two parameters should be swapped in ((citation/core)). Right now, the code looks like this:
—Trappist the monk (talk) 13:48, 21 March 2013 (UTC)
That is fixed in the Lua version. Not sure when web will be deployed:
Cite web comparison
Wikitext
((cite web|accessdate=2013-03-21|format=pdf|language=Spanish|title=Aniversario de la Proclamacion de la Independencia del Perú|url=http://www.bnp.gob.pe/snb/data/periodico_mural/2009/7/fechas_civicas/28dejulio.pdf))
Could this be fixed please? Peter coxhead (talk) 16:59, 24 March 2013 (UTC)
This will be fixed in the next release of Module:Citation/CS1. Personally though, I think that's a rather silly parameter. It would make more sense to either always end with an ampersand, or never do so. I don't really see the point of having it as a user configurable parameter (which is turned on for fewer than 1 in 100 citations). Dragons flight (talk) 17:33, 24 March 2013 (UTC)
Well, I personally prefer the ampersand, but I don't feel very strongly about it. Peter coxhead (talk) 21:25, 24 March 2013 (UTC)
no it's not, and it shouldn't. it is a parameter that gives further info depending on media type: in "cite book" it normally (though not exclusively) refers to binding. in "cite web" or digital media may refer to file format (eg an e-book could be in EPUB format, a downloadable journal may be in Kindle format etc. no urls may be provided in these cases.) 70.19.122.39 (talk) 13:28, 25 March 2013 (UTC)
You are confusing format and type. Per ((cite book)) and ((cite web)):
format: Format of the work referred to by url; examples: PDF, DOC, XLS
type: Provides additional information about the media type of the source
and thus ambiguity in parameter meaning is unhappily retained. i am not discussing "type". "format" in book citations means primarily binding format: paperback, cloth etc. "format" in url/digital media citations means primarily (electronic/virtual) binding format: mobipocket format, portable document format etc. see? simple. so format can and SHOULD mean the same across the cite system. let's make it so. 70.19.122.39 (talk) 14:01, 25 March 2013 (UTC)
The format parameter is consistent across all 23 templates. --— Gadget850 (Ed)talk 14:29, 25 March 2013 (UTC)
please stop, this is not making sense.
1. spot the "consistency" in the display bug below:
Markup
Renders as
((cite book|author=Author|date=Date|chapter=Chapter|work=Book|format=Format|type=Type))
vs.
((cite journal|author=Author|date=Date|title=Article|journal=Journal|format=Format|type=Type))
Author (Date). "Chapter". (Type). ((cite book)): |author= has generic name (help); |format= requires |url= (help); |work= ignored (help); Check date values in: |date= (help); Missing or empty |title= (help)
vs.
Author (Date). "Article". Journal (Type). ((cite journal)): |author= has generic name (help); |format= requires |url= (help); Check date values in: |date= (help)
2. the dependency of |format= on |url= is just wrong and should be undone. the doc should immediately reflect this.
in citations, "format" has a more far-ranging meaning than the very specific "url". consequently, it can justifiably be used by editors to convey info about the item cited. assuming the citation is properly structured the average reader can easily divine the meaning of "format" from the context:
Author (paperback) Book (print) - in this case the item (book) type/medium is print and is formatted as paperback
Performer (mono) MusicWork (CD) - in this case the item (audio) type/medium is a CD and the audio format is monophonic
Author (CD) Book (audio) - in this case the item (book) type/medium is audiobook and the format is CD
Author (EPUB) Book (digital) - in this case the item (book) type/medium is digital/e-book and the format is EPUB
Creator (headstone) Sign (marble) - in this case the item (sign/visual) type/medium is marble and the work's format is headstone
"format" is a variable that can assume many more values than "url" which only refers to an internet address. therefore it is a more significant variable.
in the context of designed dependencies a basic principle of sound design is to make less significant variables dependent on more significant ones, not the other way around.
it is truly amazing that i have to expend time and effort to explain these self-evident things. please, just fix it or at least stop making condescending excuses. 70.19.122.39 (talk) 12:55, 26 March 2013 (UTC)
Looks like you're having much the same problem I ran into recently. See #verbiageEEng (talk) 19:05, 27 March 2013 (UTC)
well what can i say. i appreciate the effort, but responsiveness needs improvement. 70.19.122.39 (talk) 01:11, 28 March 2013 (UTC)
It appears that your vision of format conflicts with the manner in which it has been documented since the original templates were implemented. If you use a field in a manner in which it was not intended, then we cannot anticipate that changes affect your use or misuse. Each field in each template is well documented and is fairly consistent across the series. format is implemented and documented exactly the same across all templates: it reflects the format of the web link, not the medium type of the source which is intended for type. --— Gadget850 (Ed)talk 15:52, 26 March 2013 (UTC)
i have no "vision" for |format=. i explained above how "format" may be used logically and simply to make a citation more accessible to readers. making a citation more accessible is making it more easy to verify, and that is the only reason a citation exists, or else might as well do away with sources altogether. the illogical format/url dependency is a relatively recent change. previously the documentation assigned |format= a more pertinent and correct place. when the doc is fixed (it is substandard) you may want to restore it. anyway, don't want to be discussing this forever, you either see it or you don't. 70.19.122.39 (talk) 13:55, 27 March 2013 (UTC)
At this point I am not going to convince you of anything. The documentation will continue to be substandard until you either boldly change it or become involved in discussing specific issues. I am quite tired of the neeping and nopping. --— Gadget850 (Ed)talk 19:16, 27 March 2013 (UTC)
i have gone into details several times before. i have also mentioned that you imo cannot treat doc for wildly different sources as a monolithic item. technically the citations share certain parameters and certain display conventions (and code). this is of no importance to users, and the technical doc is better suited to developers. editors need to know how best to cite a specific type of source, and readers need to understand what the citation describes. so it is that "vision" thing again. don't think about parameters, think about what any one parameter means to someone (reader/editor) who hasn't seen that particular type of citation before. i thought you welcomed reports of problems. i see this as a problem. 70.19.122.39 (talk) 01:11, 28 March 2013 (UTC)
btw, unless the tfd process becomes more rational, don't expect any involvement on my part. i'm not gonna do work that can be deleted so easily, and with minimal discussion. 70.19.122.39 (talk) 01:11, 28 March 2013 (UTC)
At this edit on 30 May 2011, ((citation)) got text describing the purpose of |format=. The new((cite book))documentation of 9 September 2006 has text describing the purpose of |format= from its inception. In both cases, |format= is grouped under |url= implicitly identifying it as pertaining to |url=.
I have found no uproar in the ((citation)) and ((cite book)) talk pages to suggest that the meaning of |format= is or should be other than to identify the format of the resource addressed by |url=. You wrote: previously the documentation assigned |format= a more pertinent and correct place. Can you point me to that previous documentation?
—Trappist the monk (talk) 20:05, 27 March 2013 (UTC)
i will try to find it. however it may have been in the doc for Wikipedia:Citing sources or a similar, equally high profile page, and i will check. for years, i kept seeing "format" in ((cite book)) citations with such info as "hardcover" etc. nobody complained about that either. 70.19.122.39 (talk) 01:11, 28 March 2013 (UTC)
well i checked all the pages i thought possibly would carry the info but i could not find it. all i can say is that i do recall "format" being explained in the manner i suggested, and a "cite book" example being given with "format" describing the (physical) binding. i still stand by the merit of my suggestion. in any case, if there is a url involved, for verification purposes, the editor is no longer are citing a (physical) book, but a digital medium, and "type" as well as "format" should reflect that. the editor may have consulted a physical book when citing, but that is irrelevant: you are telling readers to verify the cited information online. edit: i assume the online publisher if different, (should get equal billing in |publisher=) is reliable 70.19.122.39 (talk) 14:35, 29 March 2013 (UTC)
((cite journal|department=Bryan on Scouting|first=Bryan|journal=Scouting|last=Wendell|title=Calendar of New Merit Badges|url=http://blog.scoutingmagazine.org/merit-badge-calendar/))
Take a look at sharpening. There are only two citations, and they are almost identical. However the second has correctly folded the url inside itself, whereas the first one shows both the url and its title (in square brackets). What's the difference? I tried various things, such as remove format, remove pages, remove www. from the url. Nothing helped
As usual, the cause of this is a newline within the linked title. Perhaps the new Lua version of the templates could find and fix this when it happens? —David Eppstein (talk) 18:29, 27 March 2013 (UTC)
This is already fixed in Lua, though we haven't yet deployed a Lua version of ((cite web)) so that page couldn't yet benefit. Dragons flight (talk) 18:45, 27 March 2013 (UTC)
Fixed Lua deployed so this issue will be automatically fixed in the future. --— Gadget850 (Ed)talk 11:01, 29 March 2013 (UTC)
lua: |origyear= does not display when no author
it used to.
now:
Markup
Renders as
((cite book|author=Author|date=Date|origyear=OrigYear|work=Book))
((cite book|author=<!-- No author. Who knows who wrote this?? etc etc -->|date=Date|origyear=OrigYear|work=Book))
Author (Date) [OrigYear]. ((cite book)): |author= has generic name (help); |work= ignored (help); Check date values in: |date= (help); Missing or empty |title= (help)
. Date [OrigYear]. ((cite book)): |work= ignored (help); Check date values in: |date= (help); Missing or empty |title= (help)
please don't ask why should anyone cite like this. probably because they have to?? let's not have a philosophical discussion again. 70.19.122.39 (talk) 01:17, 28 March 2013 (UTC)
Oppose. This page is for general issues about the structure of CS1 citations, whereas the talk-page Module talk:Citation/CS1 is for issues about the Lua script module, just as Template_talk:Citation/core is for issues about the markup helper template, {Citation/core}. Each talk-page focuses on different areas of concern. -Wikid77 21:33, 27 March 2013 (UTC)
((Cite DVD-notes)): update to 'Cite media notes' (436 uses)
Thoughts before I take this to TfD? --Gadget850 (Ed)talk 01:25, 31 March 2013 (UTC)
Sensible to merge, but is the meaning of "Cite media notes" sufficiently clear? "Cite packaging", perhaps? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:24, 1 April 2013 (UTC)
agree with the merge, but also with Andy Mabbett, in that the proposed template name may not be as precise as it should. however his "cite packaging" is too broad. the media notes referred to here are liner notes, aren't they? these pertain to the included work, not primarily to the work's medium or packaging. 70.19.122.39 (talk) 14:11, 1 April 2013 (UTC)
Copied from Wikipedia talk:Help desk -- John of Reading (talk) 10:18, 1 April 2013 (UTC)
Hi. I have been using language templates in references. They were appearing really great some weeks ago. However, currently, they appear really in a bad shape. Formerly, language categories don't appear when I add the lang templates in reference titles (I think they were hidden categories). Now, they are included in the links.
Don't. We have been updating the templates. The language should have been included 'language' field without using a template.
language: The language the source is written in, if not English. Displays in parentheses with "in" before the language name. Use the full language name; do not use icons or templates.
I am sure we can fix it, but my recommendation will be to show it and put the page in an error category. The language icon templates really don't do anything but show the language in gray. Down the line, I want to use the 'language' field to properly indicate the title language. --Gadget850 (Ed)talk 11:05, 1 April 2013 (UTC)
Thanks for the solution! And thanks for John of Reading for moving it here!--AR E N Z O Y 1 6A•t a l k• 12:57, 1 April 2013 (UTC)
'producer' would work works, if it is followed by "producer" or "producers".
'writer' would work works if there are no author parameters. I can see this used if the author is a reviewer and |type=review. Then follow 'writer' with "writer" or "writers".
Thank you. These don't seem to be documented. I'm not sure what you mean by following |producer= with "producer". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:29, 1 April 2013 (UTC)
I refactored, as I see the confusion. You would have to identify each role; for example:
Drucker, Sam (writer). Bloggs, Joe (producer) (January 13, 1999). "André the Giant". Biography. A&E.((cite episode)): CS1 maint: multiple names: authors list (link)
Written by Drucker, Sam. Produced by Bloggs, Joe. (January 13, 1999). "André the Giant". Biography. A&E.((cite episode)): CS1 maint: multiple names: authors list (link)
'translator' has been requested multiple times. Now that we can more easily add it without a performance hit I am going to add it to the feature request at Module talk:Citation/CS1/Feature requests. --Gadget850 (Ed)talk 19:50, 1 April 2013 (UTC)
Hyphen to endash conversion has been added for the short volume case. The bolding is intentional and related to style consistency across cite book / cite journal / etc. Dragons flight (talk) 15:05, 25 March 2013 (UTC)
Oh, I see, you are complaining that "1–2" in particular wasn't rendered as bold. Dragons flight (talk) 15:25, 25 March 2013 (UTC)
if ( Volume ~= nil and Volume ~="" )
then
if ( mw.ustring.len(Volume) > 4 )
then Volume = sepc .." " .. Volume
else Volume = " <b>" .. hyphentodash(Volume) .. "</b>"
end
else Volume = "" end
so "volume data" are bolded when they are 4-characters-long or less
but "volume data" are not bolded otherwise.
congratulations! you introduced one more design bug (in the form of presentational ambiguity)
what on earth? why not leave "volume data" always bolded? or always in plaintext? is there a competition to cram as many "if" statements (process forks) as possible? if you want to have volume numbersbolded and volume titles in plaintext why not just introduce different parameters? but I can hardly think of a need for that.
am I missing some other unfathomable design purpose here? 70.19.122.39 (talk) 13:13, 26 March 2013 (UTC)
The "purpose" is that people apparently like the bolding unless the volume label was long. For example:
People seem to like bolding on short numbers:
((cite book | first = John | last = Doe | title=History of America | volume=1 | date=1978 | publisher=McMillan and Company | location=New York ))
Doe, John (1978). History of America. 1. New York: McMillan and Company.
But people also complain about:
((cite book | first = John | last = Doe | title=History of America | volume=American Revolution to Civil War | date = 1978 | publisher=McMillan and Company | location=New York ))
Doe, John (1978). History of America. American Revolution to Civil War. New York: McMillan and Company.
So, now we have:
Doe, John (1978). History of America. Vol. 1. New York: McMillan and Company.
Doe, John (1978). History of America. Vol. American Revolution to Civil War. New York: McMillan and Company.
If you can convince the Wikipedia editor community to settle on using one style or the other consistently, then all the better. However, if the community requests content-dependent formatting then that is what they are going to get. For the record, I'm just reporting the logic as it was communicated to me. The actual decision to implement this approach was made before I got involved, so I'm not sure where the original discussions of this may have taken place. Dragons flight (talk) 14:58, 26 March 2013 (UTC)
There have been multiple discussions about volume bold on the talk pages for Help:CS1, ((Citation)) and ((citation/core)). At that point, we did not really have the ability to selectively bold the volume, so those discussions went nowhere. --— Gadget850 (Ed)talk 15:03, 26 March 2013 (UTC)
i don't need to convince anybody: volume data was always bolded by default before, and i am unaware of any convincing being done then. it is also not true that we could not "selectively bold the volume". all you had to do was add ''' markup around the data to unbold it. i did it in 1 instance. the doc was mum on that workaround so i suppose whoever needed convincing was convinced, or else thankfully the powers that be were unaware of the workaround, and thus could not screw things further. unlike in this case.
i shouldn't have to convince you that readers (they are the most significant target, the editors are second) are not well served by more ambiguity. and that is exactly what the new presentation of |volume= introduces, being sometimes bold and sometimes not. it is imo much better for editors (and readers) if you introduce eg the independent parameters "volume" (for volume title) and "number" (for volume number) as i suggested above. and then you can preformat them as you see fit. 70.19.122.39 (talk) 13:58, 27 March 2013 (UTC)
also, consider situations where multiple vols. may be cited:
Editor (ed.). "Editor Notes". Vol. 1–3, 6, 18–22. ((cite encyclopedia)): |editor= has generic name (help); |work= ignored (help); Missing or empty |title= (help)
numbers, not bolded. i used ((cite encyclopedia)) because is the recommended template for multi-vol works. also, is it impossible to have a 3-letter volume title? i don't think so. 70.19.122.39 (talk) 14:20, 30 March 2013 (UTC)
i notice that trailing punctuation for |series= has been removed if volume data<5. although this is a good idea, it compounds the confusion, because the punctuation is retained otherwise. see also my related proposal here. 70.19.122.39 (talk) 13:09, 3 April 2013 (UTC)
in related consideration, should "volume title" even be in plain text? every other "title" or "wannabe-title" is preformatted. 70.19.122.39 (talk) 13:22, 3 April 2013 (UTC)
Parameter="origyear"
Example
Result
((cite web
|url = http://hebrewbooks.org/38753
|title = Sepher Yezirah
|origyear=Published 1877
|location=New York
|work=HebrewBooks.org
))
Why does the parameter not work? ------ 00:59, 4 April 2013 (UTC)
My two penn'orth is that it should be a year, i.e. 9999, only John of Cromer in Philippines (talk) mytime= Thu 09:19, wikitime= 01:19, 4 April 2013 (UTC)
'origyear' is a child of 'year'; that is, if 'year' is not defined, then 'origyear' does not show. See the ((cite web)) documentation. The proper citation is:
Markup
Renders as
((cite book |title=Sepher Yezirah |last=Kalisch |first=Isidor |location=New York |publisher=L. H. Frank |year=1877 |language=Hebrew |url=http://hebrewbooks.org/38753))
Kalisch, Isidor (1877). Sepher Yezirah (in Hebrew). New York: L. H. Frank.
Gadget850: Thanks for your quick and helpful response.
John of Cromer: Thanks for your even quicker response. But:
Next time, please post your response after the end of my signature & timestamp; not before.
According to Template:Citation Style documentation/date, it is preferable to "supply specifics; example: |origyear=First published 1859 or |origyear=Composed 1904." ------ 02:27, 4 April 2013 (UTC)
((cite web |url=http://www.autoblog.com/2010/03/29/hennessey-venom-gt-a-600k-mid-engine-cobra-for-the-21st-centur/ |title=Hennessey Venom GT: A $600k mid-engine Cobra for the 21st Century |accessdate=2010-03-29 |last=Lavrinc |first=Damon |date=2010-03-29 |work=[[Autoblog.com|Autoblog]] |publisher=[[Weblogs, Inc.]] ))
Gets displayed as:
Lavrinc, Damon (2010-03-29). "Hennessey Venom GT: A $600k mid-engine Cobra for the 21st Century". Autoblog. Weblogs, Inc.span class="reference-accessdate">. Retrieved 2010-03-29.
Note malformed span class. Oddly, if the publisher field ("Weblogs, Inc.") is changed to something else (e.g. "Weblogs, Inc", without the period, or "Weblogs, Inc.", without the square brackets), it works fine. GregorB (talk) 09:59, 4 April 2013 (UTC)
I suspect it is something in the new algorithm that removes extra periods. Reported. --Gadget850 (Ed)talk 10:39, 4 April 2013 (UTC)
In reviewing many examples of press release citations (Module talk:Citation/CS1/test/press), one of the things I noticed was that the organization issuing the press release was almost always listed a "publisher" rather than "author". For example:
(( cite press release | publisher=HTC Corporation | title=Scartel and HTC Launch World's First Integrated GSM/WiMAX Handset | date=12 November 2008 | url=http://www.htc.com/www/press.aspx?id=76204&lang=1033 | accessdate=1 March 2011 ))
This seems a little odd to me. Obviously, there are many press releases where no specific named individuals are listed, but still the corporation is responsible for the content of the message, and personally, it seems more logical to say that "author=Big Corporation" rather than "publisher=Big Corporation". One of the reasons this might come up is that ((cite press release)) doesn't list |author= in its standard prototype (using "first" and "last" instead), so people may be unaware that "author" is an option.
If people agree that it makes sense to list the issuing corporation as the "author", it might be good to update the documentation for ((cite press release)) accordingly. Dragons flight (talk) 20:25, 5 April 2013 (UTC)
I don't really agree. I think of "author" as meaning a person. "Publisher" seems to me more appropriate for an organisation. -- Alarics (talk) 05:49, 6 April 2013 (UTC)
The flip side of this is that there are also people using "publisher" to indicate the newswire or other third party that distributed the press release. That's the more traditional meaning of publisher. It also means that some press release citations currently have no immediate indication of who wrote them unless you follow through and read the release itself. If Mom's Motors wrote a press release and Reuters distributed it, then I would say it is natural to list "Mom's Motors" as the author and "Reuters" as the publisher for essentially the same reason that if John Jacobson wrote a news article distributed by Reuters then "John Jacobson" is the author and "Reuters" is the publisher. Anyway, that's my two cents. Dragons flight (talk) 07:20, 6 April 2013 (UTC)
"Work" Parameter for Website Citation
Please explain what is meant by "Work" for Website "cite web" citations. How is it different from Author and Publisher?
I would also suggest that an explanation be given for this parameter within the Website "cite web" template, which is the only template that uses "Work" as a parameter and yet has no examples or explanations for it.Wondering55 (talk) 15:09, 6 April 2013 (UTC)
From ((cite web)): "'work': Title of website; can be wikilinked to an existing Wikipedia article or url may be used to add an external link, but not both. Displays in italics." And many other Citation Style 1 templates use 'work' but with a slightly different meaning. --Gadget850 (Ed)talk 16:26, 6 April 2013 (UTC)
ORCID
ORCID identifiers disambiguate journal contributors with similar names; and unite records for authors who write under more than one name. Think of them as an ISSN for people. For example. my ORCID is 0000-0001-5882-6823 and the corresponding URI is http://orcid.org/0000-0001-5882-6823
I believe that we should add an option for an ORCID parameter for each author of a cited work, to CS1 templates. Whether or not to make that a clickable link can be debated. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:33, 8 March 2013 (UTC)
I believe ORCIDs are a good idea within the ((Authority control)) template on Wikipedia articles about individual researchers. I am less convinced that we should clutter our bibliographies with them, especially in cases where we have a wikilink to the author's article. —David Eppstein (talk) 22:17, 8 March 2013 (UTC)
I suspect (and would welcome evidence) that in the vast majority of cases, we have no article about the author, and thus no such link. ORCIDs would not be clutter, but valuable information. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:53, 9 March 2013 (UTC)
I support this. It should be a clickable link and the template should check the format of the ORCID for mangled ones. Stuartyeates (talk) 06:12, 9 March 2013 (UTC)
I agree with David, they should be in the authority control system and not in the citation templates themselves. I don't see how they help a reader locate a copy of the source to verify the information being cited. I do see how the ISSN would help a reader locate a library that maintains a copy of the journal/magazine, but I'm not sure the same can be said for ORCIDs. Imzadi 1979→ 06:37, 9 March 2013 (UTC)
You're quite right. But then no such claim has been made. An ORCID doesn't help find the publication; it helps find the - and unambiguously to identify - the author; and thus to find that author's other works, affiliation, reputation, etc. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:51, 9 March 2013 (UTC)
The purpose of a citation is to allow a reader to find the publication being cited for verification purposes. The other works of the author, etc would best be spelled out in an article on that author, which conveniently enough, |authorlink= would provide. From that article then, the ORCID and other authority control information would be linked. Anything else in the citation is clutter. That isn't to say that the ORCID isn't valuable, just that it isn't needed, in this context. Imzadi 1979→ 12:31, 9 March 2013 (UTC)
Please refer to my above comment, where I have already addressed the |authorlink= red herring. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:35, 9 March 2013 (UTC)
one solution would be to have "authorlink" output either a wikilink or an ORCID? add logic in the template so that if the field's data start with a number link to ORCID. i realize there could be situations where people (legally) use numbers as names. in which case my idea would not be optimal. 70.19.122.39 (talk) 15:37, 9 March 2013 (UTC)
If we introduce data like this (sounds useful for cross referencing, metadata), we have to be able to strike a balance between the additional data and a default display of the reference that is not too detailed for an average reader. Not sure how we would do this/how it would be configured though. Rjwilmsi 10:17, 12 March 2013 (UTC)
If we must, we could hide ORCID numbers behind an icon, or the text "ORCID"; but displaying them would be better. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:31, 12 March 2013 (UTC)
.39's solution would be technically easy, just "if authrolink exists, do not display ORCID". Numeric names would not be a problem. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:31, 12 March 2013 (UTC)
your amendment is more elegant and takes care of names that are numbers. the only potential problem i can see is that #ifexist is expensive, or more so than plain #if statements that don't check article namespace. conceivably an article could bump into limits when there are many citation templates? 70.19.122.39 (talk) 14:00, 12 March 2013 (UTC)
And it would mean that when printed or ported, the ORCID would be hidden. --— Gadget850 (Ed)talk 14:15, 12 March 2013 (UTC)
That's easily dealt with by CSS, as we do with external links in citations, for instance. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:22, 12 March 2013 (UTC)
Would someone like to knock up a test implementation in a sandbox, please? If you need a sample ORCID, mine is on my user page. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:26, 12 March 2013 (UTC)
I support this conceptually but am not sure how to participate or what is next to be done. Blue Rasberry (talk) 14:35, 27 March 2013 (UTC)
RESOLVED: Use current "postscript=" for any additional data: The current wp:CS1 format allows parameter "postscript=. ORCID [http://orcid.org/0000-0001-9977-9999 9977-9999]." to add the ORCID id, or any of millions of other extra data items into each citation. That option has been available for years, so feel free to discuss with other editors of each article when wanting to append more details. The postscript parameter can accept over 50,000 characters of data, including maps with driving instructions to the library, to help readers find each source document. -Wikid77 (talk) 20:49, 27 March 2013 (UTC)
NOT RESOLVED and please don't unilaterally declare your suggestion as a resolution. Using |postscript= would mean that an ORCID identity would not be identified as such; not automatically linked to the relevant URL; and that its display would not be suppressed in the presence of an |authorlink=, as discussed above. Your map comment is asinine. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:04, 27 March 2013 (UTC)
Cool your jets. Wikid77 wasn't saying that a resolution had been reached, but rather was following the quaint forensic tradition of prefacing a proposed decision with RESOLVED i.e. it means I move that it be RESOLVED that.... See [3]. It's a peculiar idiom indeed, but one that, to be frank, should be well within the experience of anyone dealing with the niceties of research and citation, so please have a care in future before flying off the handle. An apology seems in order. EEng (talk) 23:17, 27 March 2013 (UTC)
Um, no. That may well be what was in Wikid77's head, but this page uses the English language, not some "peculiar", "quaint forensic tradition". You don't see that usage elsewhere on Wikipedia, nor even on the rest of this page. It's not unreasonable to suppose that other readers will see the word "resolved", emboldened and in all-caps, and pass on to the next section (c.f. ((resolved))). Now, can anyone help with the ORCID parameter requested? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:01, 28 March 2013 (UTC)
There is much that is peculiar or quaint which may reasonably be expected to be within the experience of the culturally literate. Apparently you've never served on a board of directors, public commission, or student council; participated in any kind of formal discussion or debate; been present at a faculty or school board meeting, or a union vote; owned stock; or read about any of these things with any depth of comprehension. That's OK, of course; what's not OK is that, faced with an opportunity to expand your knowledge just a bit, you choose instead to insist that the rest of us restrict our discourse to your cramped radius of experience. It's kind of like the old joke about opposition to foreign-language instruction in the schools: "If English was good enough for Jesus, it's good enough for our kids!" EEng (talk) 21:48, 28 March 2013 (UTC)
Well, that's quite a collection of attacks and failures to AGF. No wonder you leap to so many false conclusions. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:14, 29 March 2013 (UTC)
'postscript' is for the terminating punctuation of the citation- please don't suggest misusing it. I am sure some editors are doing something odd and wonderful, but we have already broken a few misused parameters. An it won't work for ORCID as there will be one ORCID per author and it has to be attached to the author.
Putting the ORCID on the author page and linking to it with 'authorlink' will not work if the page is ported to another wiki but the author page is not.
The ((authority control)) template is for the subject of the article, not for authors of individual sources used as citations. ORCID would be good there for use when the article is about a specific author
The key question is: does ORCID help identify the source? As I see it, ORCID does help to uniquely identify the author, regardless of duplicate names or name changes.
But, ORCID is not the only author identifier. It is actually a subset of the International Standard Name Identifier which further allows for pseudonyms and publisher imprints. ORCID is for researchers and academic authors, whereas ISNI covers music, film and other media. I think we should use both. I would also like to see if they can both be included in one parameter.
As noted, ORCD/ISNI would be attached to individual author names.
We are in the process of updating CS1 templates to Lua. Lets add this to the queue as a new feature while I look at adding this to ((citation/core)). --— Gadget850 (Ed)talk 13:55, 28 March 2013 (UTC)
Naming the illustrator or photographer in Cite book
There doesn't seem to be an obvious way to add the name of an illustrator or photographer to the Cite book template. How should it be done so that it renders "Illustrations/Photographs by Jane Doe" in the citation? Roger (talk) 11:57, 7 April 2013 (UTC)
Isn't that what the |others= field is for?Nigel Ish (talk) 12:05, 7 April 2013 (UTC)
Should the original url= be required when using archiveurl=
People here may be interested in commenting on the issue described at
Consistency between work and publisher parameters in citation templates
Will there ever be consistency between the work and publisher parameters in citation templates? Every Featured Article I've seen has website names in the publisher parameter of a citation template. This keeps the website from being italicized. However, ((cite web)) states that the work parameter is specifically for websites, and when using the auto-template form (not sure what to call it—appears in the edit window when you click on "Cite > Templates > Web"), the little question mark above the work parameter also states that this is for websites, which means that most people will follow this advice, since more people see the auto-template form advice than anything demanded by FA criteria. This seems to be a conflicting way of citing websites, and I'm wondering if, or what, steps have been taken to try to standardize the work/publisher parameter definitions. – Kerαunoςcopia◁galaxies 18:46, 8 April 2013 (UTC)
The short answer is no. 'work' is not a good parameter name in this context, and I have a feature request to add 'website' as an alias; Module talk:Citation/CS1/Feature requests. The toolbar is WP:RefToolbar. But, editors will keep doing as they have done. And many put the domain name in as the website name, which is only occasionally correct. The definitions for 'work' and 'publisher' are clearly documented.
I'm not sure if website as an alias would solve anything. It seems to me that users at Featured Articles are using the publisher parameter as a way to avoid the auto-italicizing. I don't believe websites should be italicized at all, that would be like italicizing "Scholastic Books". I think so long as one parameter italicizes, and the other lets you decide whether you should italicize or not, as is shown in the Rogues gallery, then editors will choose which parameter they would like to use despite any actual parameter definitions given on the template documentation. What bothers me is that users are being told to put websites in the auto-italicizing parameter, which makes no sense to me, although someone made the big decision, so I'm sure at some point there have been discussions regarding this already. But if articles that are featured on the main page don't follow this guideline, then I'm not understanding why this isn't a bigger issue on the template talk pages—hence my bringing it up here. – Kerαunoςcopia◁galaxies 21:46, 8 April 2013 (UTC)
—Trappist the monk (talk) 13:08, 12 April 2013 (UTC)
Parameter |day of week= is broken
What's happened to this parameter? Showing error messages all over the place! Mjroots (talk) 13:39, 13 April 2013 (UTC)
Probably because there has never been a |day of week= parameter and the new error checking is showing this. Be specific: what articles are showing this? --Gadget850 (Ed)talk 13:45, 13 April 2013 (UTC)
—Trappist the monk (talk) 14:06, 13 April 2013 (UTC)
My SWAG is that someone propagated the |day of week= from ((cite newspaper The Times)) to ((cite news)). Regardless, remove the parameter and value (which never worked) and the problem will be resolved. --Gadget850 (Ed)talk 15:19, 13 April 2013 (UTC)
I beg to differ, it worked perfectly before. Surely the solution is to fix the template so that it works again. Mjroots (talk) 15:27, 13 April 2013 (UTC)
It never showed the day of week. The error check is new. Comparing the old version to the new:
Cite news comparison
Wikitext
((cite news|date=8 January 1803|day_of_week=Saturday|issue=Issue '''6587'''|newspaper=The Newcastle Courant etc.|title=MARINE INTELLIGENCE))
Live
"MARINE INTELLIGENCE". The Newcastle Courant etc. No. Issue 6587. 8 January 1803. ((cite news)): |issue= has extra text (help); Unknown parameter |day_of_week= ignored (help)
Sandbox
"MARINE INTELLIGENCE". The Newcastle Courant etc. No. Issue 6587. 8 January 1803. ((cite news)): |issue= has extra text (help); Unknown parameter |day_of_week= ignored (help)
The old version of the template is at ((cite news/old)). Feel free to test or to look for the parameter. Day of week has never been supported in any of the CS1 templates.--Gadget850 (Ed)talk 15:33, 13 April 2013 (UTC)
Looks like this is being cleaned up. --Gadget850 (Ed)talk 20:27, 13 April 2013 (UTC)
FN1 at My Funny Valentine is messed up due a programming bug, someone please fix or unprotect it and let me do it.--Launchballer 20:27, 13 April 2013 (UTC)
It's broken for two reasons: The |isbn= parameter contains text other than an ISBN (a Please check ISBN) and because the check digit doesn't match the calculated value. Deleting the Please check ISBN template and finding the correct ISBN will fix the problem.
—Trappist the monk (talk) 20:40, 13 April 2013 (UTC)
The ISBN can be fixed easily, however I'm actually going to recommend leaving it with the error because the error message is messed up. I'm fairly sure it shouldn't render as [[Special:BookSources/1-4144-0140-9|1-4144-0140-9 [[:Category:Articles with invalid ISBNs]]]], because links within links don't work.--Launchballer 21:31, 13 April 2013 (UTC)
In article namespace, Please check ISBN is replaced with this text Category:Articles with invalid ISBNs. Doesn't happen here because this is Help talk namespace.
Here is the original citation:
((cite book |last=Trager |first=James |year=2005 |title=The People's Chronology: A Year-by-Year Record of Human Events from Prehistory to the Present |edition=3 |location=Detroit |publisher=Gale |isbn= 1-4144-0140-9 ((PleasecheckISBN|reason=Check digit (9) does not correspond to calculated figure.))))
→((cite book |last=Trager |first=James |year=2005 |title=The People's Chronology: A Year-by-Year Record of Human Events from Prehistory to the Present |edition=3 |location=Detroit |publisher=Gale |isbn= 1-4144-0140-9 ((PleasecheckISBN|reason=Check digit (9) does not correspond to calculated figure.)) |no-tracking=true))
So, without the Please check ISBN text substitution, ((cite book)) shows that the ISBN isn't correct. Leaving the Please check ISBN template in place in article namespace is corrupting the citation's COinSmetadata and the Special:BookSources link.
I don't think that what we're seeing at My Funny Valentine is necessarily anything wrong with ((cite book)) or with Module:Citation/CS1. However, I will mention this discussion and see what comes from that.
The ISBN appears to be ISBN 978-1414401409, a claim that should be verified by someone else.
—Trappist the monk (talk) 22:29, 13 April 2013 (UTC)
User:Helpful Pixie Bot added the Please check ISBN after the ISBN. Problem is, that has never worked. Adding that template into the citation breaks the ISBN link to Special:BookSources and always has:
Cite book comparison
Wikitext
((cite book|edition=3|first=James|isbn=1-4144-0140-9 [[:Category:Articles with invalid ISBNs]]|last=Trager|location=Detroit|no-tracking=true|publisher=Gale|title=The People's Chronology: A Year-by-Year Record of Human Events from Prehistory to the Present|year=2005))
Live
Trager, James (2005). The People's Chronology: A Year-by-Year Record of Human Events from Prehistory to the Present (3 ed.). Detroit: Gale. ISBN [[Special:BookSources/1-4144-0140-9 Category:Articles with invalid ISBNs|1-4144-0140-9 [[:Category:Articles with invalid ISBNs]]]]. ((cite book)): Check |isbn= value: invalid character (help)
Sandbox
Trager, James (2005). The People's Chronology: A Year-by-Year Record of Human Events from Prehistory to the Present (3 ed.). Detroit: Gale. ISBN [[Special:BookSources/1-4144-0140-9 Category:Articles with invalid ISBNs|1-4144-0140-9 [[:Category:Articles with invalid ISBNs]]]]. ((cite book)): Check |isbn= value: invalid character (help)
The nature of the error is a little different now, but it is no more broken than it already was. Someone should find out if the bot is still doing this and probably tell it to stop as it is both unnecessary (now that the same function is handled by the template), and breaks the citation template. Dragons flight (talk) 00:21, 14 April 2013 (UTC)
Actually, I think I can see what the problem is, but I'm going to need the page unprotected so I can get into it.--Launchballer 00:32, 14 April 2013 (UTC)
The bot has not run since May 2012. --Gadget850 (Ed)talk 01:01, 14 April 2013 (UTC)
Helpful Pixie Bot (talk·contribs) is blocked, as is its operator. Whilst its operator may yet be unblocked, the ArbCom decision means that the block of the bot is permanent. --Redrose64 (talk) 10:04, 14 April 2013 (UTC)
The thing that needs fixing is the citation in My Funny Valentine, not ((cite book)). There should be nothing in the |isbn= parameter's value except one ISBN.
—Trappist the monk (talk) 11:17, 14 April 2013 (UTC)
Catscan shows 1909 pages in both categories. Frankly, its not making anything worse, as both indicate bad ISBNs. --Gadget850 (Ed)talk 11:22, 14 April 2013 (UTC)
In the documentation of this template, we can read the following indication:
work: Title of website; can be wikilinked to an existing Wikipedia article or url may be used to add an external link, but not both. Displays in italics.
But when I see the Lua module, the arg work seems to be the same arg than magasine or newspaper, etc. You can see it here :
local Periodical = args.journal or args.newspaper or args.magazine or args.work
or args.periodical or args.encyclopedia or args.encyclopaedia
On the other hand, we can read that encyclopedia and encyclopaedia are both used for Title and Periodical. Is it normal? Is that someones knows why?
We can see it here:
local Title = args.title or args.encyclopaedia or args.encyclopedia or args.dictionary
Thanks by advance for the answers. Automatik (talk) 20:57, 3 April 2013 (UTC)
The CS1 templates started as some 20+ separate template authored by multiple editors. Editor Gadget850 has diligently worked through all of them to make them as much the same as possible. ((Cite encyclopedia)) is the one that doesn't easily shoehorn into the others. Editor Dragons flight, author of the Lua code, has made noises about trying to fix the ((Cite encyclopedia)) mess though I suspect it's not high on the list.
Did I answer your question?
—Trappist the monk (talk) 13:07, 12 April 2013 (UTC)
Ah, this one. This was already in place and I should have pushed harder to get it fixed a year or so ago. I have never understood why this is like this and we should just fix it and see if someone complains. --Gadget850 (Ed)talk 18:48, 12 April 2013 (UTC)
@Trappist the monk: Thanks for your answer; I didn't understand the whole response, however. What does this argument work, anyway?
@ Gadget850: Otherwise, if fix what you want to fix would make things clearer, you have my full support.
Regards, Automatik (talk) 09:52, 14 April 2013 (UTC)
The documentation pages include full parameter sets. These have always bugged me, as they seem to lead to copy/paste leaving a bunch of blank fields. The full sets include rarely used parameters such as lay summary and the display set. And none are actually full sets, as they include only a few identifiers. And now we support hundreds of authors and editors for the updated templates. Thoughts? --Gadget850 (Ed)talk 15:20, 16 April 2013 (UTC)
There are at least two opposing viewpoints here. One, expressed at #Cite episode deprecated parameters above, is that the absence of a parameter from the full set implies deprecation. The other is that when users see the full set, they believe that they must fill in as many as possible - even when irrelevant. Some editors often comment about parameters (such as |publisher= or |issn=) being used when they are often superfluous. I must admit that I tend to be on the "as many as possible" side, although you won't find me trying to add DOIs or ARXIVs or laysummaries. --Redrose64 (talk) 15:27, 16 April 2013 (UTC)
This is the problem when you have any template that includes a lot of parameters. ((Infobox settlement)) is a rather bloated template with many parameters to cover geographical and administrative differences across the entire planet. The documentation includes separate sets of parameters for metric and non-metric values, but includes a set of all parameters as well. ((Infobox Australian place)) uses a different approach. Documentation is spread over several pages, with one page containing examples of the most common uses for the template, using the most commonly used parameters. Another page has multiple blank parameter sets, including a set of all parameters. When I created ((Infobox hut)) from ((Infobox Australian Hut)), ((Infobox mountain hut)) and ((Infobox Schutzhütte)) I had to include parameters specific to European Alpine huts. The blank parameter list includes all fields with the usage table broken into sections to cater for this, and examples for the various uses have been included. The approach is obviously going to be different for each template, but the problem identified by Redrose64 (parameters being used when they are superfluous) can be addressed in the documentation. Don't simply assume that editors will know what is meant, you often need to guide people with unambiguous direction, and that usually requires more than a brief, one line comment. For example, it was necessary at ((Infobox television season)) to list some series specific parameters separately and include a notice that "the following fields are only to be used in articles that previously used forks of this template that have been deleted or are otherwise no longer used" in the parameter description table. --AussieLegend (✉) 16:32, 16 April 2013 (UTC)
ISBN
I usually change any dash in isbn to ((nbhyph)), because Sod's law says the line will break there. However now although ((nbhyph)) gets transformed correctly to ‑ that is how it stays. So isbn gets shown as for instance 978‑1892214973
Yep. This hack also breaks magic linking ISBN 978‑0812695939 v. ISBN978-0812695939. The better solution would be to add markup in the template to keep this from wrapping. --— Gadget850 (Ed)talk 15:09, 26 March 2013 (UTC)
Please do not make any such change. Not only is it not undesirable for ISBNs to be allowed wrap, it's positively desirable, especially where cites are in 2- or 3-column format. In fact I was about to suggest that linebreaks be allowed between PMID and what follows, between doi: and what follows, and (within doi) at the slash. EEng (talk) 19:43, 26 March 2013 (UTC)
I think my reference to ((nbhyph)) is bit of a red herring. In fact it breaks if I enter the hard hyphen (】) directly. John of Cromer in Philippines (talk) mytime= Wed 04:46, wikitime= 20:46, 26 March 2013 (UTC)
Done. Could we encode the URL? I know we can't do the pipe. At least this will give an error when we make this one visible. --Gadget850 (Ed)talk 17:01, 18 April 2013 (UTC)
That help text now includes a first line that describes the URI scheme. Is Special characters the right place for that snippet of text?
—Trappist the monk (talk) 17:24, 18 April 2013 (UTC)
Not precisely, but I used a template to keep the text the same across several pages. The page needs some work, especially regarding the updates. I could add a conditional to the template. --Gadget850 (Ed)talk 17:39, 18 April 2013 (UTC)
Because of the way percent encoding is defined in the URI spec it is not universally safe to always encode any of the URI reserved characters !*'();:@&=+$,/?#[]. If you always encode (or never encode) those, then you are guaranteed that some population of URLs that are valid per spec will fail. We could try encoding characters that have Mediawiki significance such as <> and space (" ") but lack URI significance. I think that would work. Or at least it will work as long as people haven't been intentionally embedding things like " " or [] in their citation URLs with the intent of exploiting the Mediawiki meaning. As an aside, I'm not sure it is ever possible to use [] in a Mediawiki based URL in the way that the URI spec intends (IP address delimiters), though on the other hand, I've never seen a valid URL that actually used [] for their spec meaning, so it might be okay to encode those as well. Dragons flight (talk) 17:48, 18 April 2013 (UTC)
section parameter missing
The parameter "section" from ((cite news)) now shows as an error. Please fix it.--Aurictalk 12:26, 21 April 2013 (UTC)
There is no 'section' parameter, which is why the error is displayed. What article? --Gadget850 (Ed)talk 12:30, 21 April 2013 (UTC)
Previous advice (not necessarily on this page) has been to use the |at= parameter, for example |at=sec. B p. 28 col. 3 --Redrose64 (talk) 14:39, 21 April 2013 (UTC)
That results in an Extra |pages= or |at= error.--Aurictalk 21:23, 21 April 2013 (UTC)
Can you tell us where this citation is or put a copy of it here?—Trappist the monk (talk) 21:49, 21 April 2013 (UTC)
It doesn't result in an Extra |pages= or |at= error if you put the page number inside the |at=, as shown in my example. --Redrose64 (talk) 21:54, 21 April 2013 (UTC)
Replace the |section=Property - Commercial Property|page=18 with |at=Property - Commercial Property section, p. 18 --Redrose64 (talk) 22:00, 21 April 2013 (UTC)
That works. Thanks for the clarification.--Aurictalk 22:07, 21 April 2013 (UTC)
Or use |department=, if "Property - Commercial Property" is a regular department of the newspaper, and return to |page=:
((cite news |newspaper=[[The Sydney Morning Herald]] |title=Leasing Ladder |page=18 |department=Property - Commercial Property |date=17 July 2010 |accessdate=25 October 2010 |url=http://newsstore.fairfax.com.au/apps/viewDocument.ac?docID=SMH100717BGA2B40LRQ7))
—Trappist the monk (talk) 22:16, 21 April 2013 (UTC)
Very weirdly, I've happened across several examples of |section= being used in ((cite news)) and ((cite web)) in the past. I discovered another today.[4] --AussieLegend (✉) 14:03, 23 April 2013 (UTC)
Looks like that one could have used 'department', a parameter I added July 2012. --Gadget850talk 14:07, 23 April 2013 (UTC)
It works, but it doesn't seem a "natural" parameter. "Department" seems more related to departments in a department store or, for example, the service department in a car dealership. It doesn't really seem right in a newspaper. --AussieLegend (✉) 16:27, 23 April 2013 (UTC)
I initially proposed 'column' but it was thought that it might be confusing. We also discussed 'section' with the same conclusion. I then consulted Chicago 16 where 'department' was used.[5] --Gadget850talk 00:14, 24 April 2013 (UTC)
Cite book, at= should not be overridden by page=, simultaneous use should not be an error
The documentation for cite book (Template:Cite_book/doc#In-source_locations) shows that at: is used where page= is inappropriate or insufficient. If page= is insufficient (meaning necessary but not sufficient), then at: should not be overridden by page=, IMHO, nor should simultaneous use of the pair constitute an error. Example: page=234| at=Table 3.23 where six tables are present on a page. In my opinion, the sensible use of these params is a matter for editor discretion and discussion, not a hard and fast rule encoded in software. --Lexein (talk) 03:36, 22 April 2013 (UTC)
This has been covered several times before (not necessarily on this talk page). If the whole table is on page 234, use |at=p. 234, table 3.23 if the table spans more than one page and the pertinent information is only on page 234, use |at=Table 3.23, p. 234 In either case, omit |page= --Redrose64 (talk) 10:07, 22 April 2013 (UTC)
Seems not to be covered in the documentation, else I wouldn't have asked. That would be good. --Lexein (talk) 15:31, 22 April 2013 (UTC)
The "OR" before 'pages' and 'at' weren't sufficient? And the "Overridden by". --Gadget850 (Ed)talk 15:40, 22 April 2013 (UTC)
Whoosh. Did I touch a nerve? Are you a documentation minimalist or something? Good grief. Examples, examples, examples. Such as those good ones suggested by Redrose. Not difficult. --Lexein (talk) 15:58, 22 April 2013 (UTC)
What can we do to make the documentation better? Sorry. --Gadget850 (Ed)talk 18:20, 22 April 2013 (UTC)
I think a short table of examples is a good idea. Such as
Use case
Syntax
Display
Single page
|page=24
p. 24
Range of pages
|pages=24-38
pp. 24–38
List of pages
|pages=13-24, 28, 37
pp. 13–24, 28, 37
Unnumbered location
|at=Back Cover
Back Cover
Numbered location that is not part of a numbered page
Lets say we allow 'at' and 'page'. What do we put into the COinS rft.pages key? Do we ignore 'at' and just include 'pages'? Or do we combine them and guess at the intended order? --Gadget850 (Ed)talk 11:01, 22 April 2013 (UTC)
Oh no! An unhandled case! This comes from overconstraining, and worrying about creating code, where, as I've indicated elsewhere, discretion and discussion completely address issues attempting to be solved by code. Sorry. --Lexein (talk) 15:31, 22 April 2013 (UTC)
It is handled. Pick one of the three parameters. What is your specific alternative that allows us to properly support COinS? Sorry. --Gadget850 (Ed)talk 15:40, 22 April 2013 (UTC)
I agree that |at= should not be overridden by |page= (or vice versa). There are cases where it is very handy to have both page numbers and (e.g.) section numbers. But combining them in |at= is objectionable, as it breaks the metadata association. I don't know what COiNS wants, but if it asks for "pages" then probably that is what it should be given, not "at".~ J. Johnson (JJ) (talk) 23:20, 22 April 2013 (UTC)
It has always been the case that using |page= or |pages= will cause |at= to be ignored. That said, we could change it. What syntax would you suggest, e.g. "p. <PAGE>, <AT>"? The examples above aren't entirely consistent since one example goes essentially "<AT>, p. <PAGE>". One could also imagine a number of other variations using colons or parentheses as separators for example. Dragons flight (talk) 23:35, 22 April 2013 (UTC)
I propose we show 'at', 'page' and 'pages' and ignore rft.pages. --Gadget850 (Ed)talk 23:47, 22 April 2013 (UTC)
We don't presently simultaneously show both 'page' and 'pages', and surely that behavior can and should be kept. Right? As to rft.pages, I am not acquainted with where and how it displays; perhaps you would favor us with an explanation? ~ J. Johnson (JJ) (talk) 00:02, 23 April 2013 (UTC)
We also don't show 'at' if 'page' is defined. You want to define and show both 'page' and 'at'. There are a lot of folks who are defining both 'page' and 'pages', so we should accommodate them as well, thus show all three. rft.pages is a COinS key. It does not show, but someone with reference management software like Zotero can scan the page and copy all the references into a database. They can then reuse those references elsewhere. Currently, one of the three in-source location fields ('at', 'page', 'pages') is inserted into the rft.pages key. I think we should just ignore rft.pages and not use it. --Gadget850 (Ed)talk 00:10, 23 April 2013 (UTC)
In fact, this is over constrained. We should add parameters for 'act', 'back cover', 'book', 'canto', 'colophon', 'column', 'dust jacket', 'folio', 'hours', 'indicia', 'liner notes', 'minutes', 'paragraph', 'part', 'scene', 'seconds','section', 'stanza', 'track', 'verse'. Did I miss anything? --Gadget850 (Ed)talk 00:37, 23 April 2013 (UTC)
I think you forgot table, figure, code listing, subsection, theorem, lemma, corollary, definition, observation, and equation. Probably a few others besides... —David Eppstein (talk) 02:24, 23 April 2013 (UTC)
Two of the three parameters are merely format specifiers that aid editors in getting the format right. That leaves |at= which allows for free-form specification of any kind of locator – but, the editor has to take care to use the proper p. or pp. for the rendered citation.
I suspect that |page= and |pages= by far outnumber |at= in citations. That would suggest that |at= should rank lowest in the hierarchy.
If I understand the definition of rft.pages, it will take any kind of in-source locator description. That being the case then, keep it.
So, this editor, who is constantly arguing for change, really doesn't see much need for change here. Perhaps inclusion of the example table and a bit of rewording in the documentation (though, I don't find it ambiguous or hard to understand) is in order.
—Trappist the monk (talk) 01:17, 23 April 2013 (UTC)
(Are you constantly arguing for change? Ah. Didn't know that.) I'm not opposed to filling in |at= with page and other detailed location, but (with Johnson, above) I feel a tug at the loss of metadata association. By the way, all, my concerns are completely addressed by better examples, and I appreciate Dragons flight's table, whatever fine tuning might be considered needed there. --Lexein (talk) 03:12, 23 April 2013 (UTC)
Ed, you seem to be feeling unusually grumpy. No, we don't need all those other specific identifiers, as |at= is generic, and can be used for all those. We do have the specific page/pages parameters because that is by far the most common form of "at". What we are asking is why "at" is essentially not allowed when page/pages is specified.
As to having separate |page= and |pages=, I think that is justified for ensuring that "p." or "pp." results, as trying to determine that in software is problematical. That people use both is probably an error (which I see gets flagged as such), where they consider 'pages' to be the page range (say of an article in journal) and 'page' to be the specification (the particular page where the material cited is to be found). As far as that goes, I would have 'pages' override 'page' within cite and citation, the more inclusive term likely being the more appropriate for a full reference. (And reversed in ((harv)) for p/pp, as a short cite should be as specific as possible.) ~ J. Johnson (JJ) (talk) 23:34, 23 April 2013 (UTC)
Citation error query
This edit to Pi showed up on my watchlist, and I'm wondering if that was the correct resolution.
It seems to me that the citation templates could do something sexy and mark up the title in italics with the lang attribute set according to the definition in the template of the language of the document. Alternatively, we could add a new parameter for "title-language", to take an ISO code language letter code compliant, to do the same thing.
Just tossing some thoughts out there. Let me know if I'm way off. :) --Izno (talk) 20:01, 24 April 2013 (UTC)
Even with the |nocat=true, ((lang)) introduces content into the WP:COinS metadata that does not belong there. Specifically, with((lang)) the COinS title metatdata looks like this:
without((lang)) the COinS title metatdata looks like this:
&rft.btitle=Elementa+Trigonometrica
The correct resolution, for the time being, is to remove ((lang)) from |title=. As Editor Gadget850 has pointed out, and integrated solution to the language issue is needed.
—Trappist the monk (talk) 13:13, 27 April 2013 (UTC)
Category:Pages using citations with accessdate and no URL
Hi, I just noted this category on a page that I edited (Journal of Reinforced Plastics and Composites). I added two references there that seem to trigger this cat. The references are online, which is why I included an accessdate. However, I cannot include a URL. The reason is that this resource is behind a pay-wall and uses dynamic session-specific URLs. This means that any URL that I would copy would not work for anyone else and, once I have closed and restarted my browser, not even for myself. I don't really see a workaround and in this case, having an accessdate but no URL seems reasonable to me. I could, of course, include a URL to the homepage of the resource, but as that is not the page where I found the information sourced by this reference, that seems less correct to me (and also runs the risk that someone will add a "fails verification" template to the reference). I'd appreciate any advice on how to handle this situation. Thanks! --Randykitty (talk) 15:34, 27 March 2013 (UTC)
Just a note for people that may be unaware. The hidden tracking category was recently added as part of the Lua migration; however, the behavior hasn't changed. It has always been the case that accessdate= is ignored if no URL is specified. Dragons flight (talk) 16:41, 27 March 2013 (UTC)
Actually, it seems to me that in the case outlined above, accessdate should perhaps not be ignored and displayed... --Randykitty (talk) 16:56, 27 March 2013 (UTC)
Yes, there is a valid question about whether accessdate should be included even if no url is given. Historically, the answer has been no, but perhaps that should change. Personally, I don't have a strong opinion either way. I was just trying to make clear that this is a discussion of a possible problem with how the accessdate parameter has been used in citations, which is different from many of the other recent reports on this page which deal with bugs specific to the Lua migration. Dragons flight (talk) 17:08, 27 March 2013 (UTC)
Per the the documentation: "Not required for web pages or linked documents that do not change; mainly of use for web pages that change frequently or have no publication date." In this instance, the access date really does not help identify the source. --— Gadget850 (Ed)talk 18:45, 27 March 2013 (UTC)
That's an excellent point. The URL may be dynamic, the page content isn't, so the accessdate is indeed not needed. Problem solved, I think! --Randykitty (talk) 19:01, 27 March 2013 (UTC)
No, your suspicions of the hidden-accessdate problem are correct, when no "url=" data, because the id parameters (such as "doi=") generate an internal URL which needs the accessdate to assure the time when access was allowed. The recent shake-up at Billboard.com disabled 50,000(?) weblinks. Thanks for raising the issue so quickly. -Wikid77 23:20, 27 March 2013 (UTC)
Use parameter "postscript=" to force a date to appear: Previously, the Lua version displayed the "accessdate=" data, with no judgmental restrictions, but merely echoed all parameters as added by each user. However, with accessdate conditionally ignored, the postscript can be used to show that data:
postscript=. Retrieved 27 March 2013.
In general, ignoring parameters is typically extremely frustrating to new users who are likely to misspell parameter "url=" as perhaps capital-letter "Url=" which would be ignored, then causing the "accessdate=" value to also be hidden as well, and a new user is likely to go bonkers wondering why this "trashy" citation software does not show the URL address nor even the damned accessdate as ignoring "everything" they put in the cite. In general, there is a fine line between "smart" software and "smart-ass" software, and the automatically ignored parameters, triggered by arcane rules of citation hierarchies, will be considered by many users to be poor-quality, and non-user-friendly. Beware when the interface acts as "minefield" of traps, where one misspelled word triggers peculiar sinkholes where other parameters also disappear. In general "keep it simple" and mimic the concept of "what you see is what you get" (WYSIWYG), so if a user enters a parameter, then display that parameter without prejudiced, judgmental restrictions as to what will be deemed permissable, and instead: if they see it in the markup, they get it. However, with the speed of Lua, we will be able to process an assistance parameter (such as "help") which could warn the user, at each specific citation, about which unexpected parameters were being ignored, according to what elaborate rules for citation etiquette, or common misspellings. -Wikid77 (talk) 21:33, 27 March 2013 (UTC)
Then we should remove all restrictions and allow any and all combinations of parameters to show. We could add 'url' 'URL' 'website' 'site' 'link' 'internet' 'tubes' and for those who just can't remember 'thingy'. Then we can delete the documentation and allow editors to pick their favorite parameter names.
But really- postscript is the terminating punctuation. Yes, editors stuff all sorts of crap in it, but we aren't allowed to send them to citation reeducation camp. Unless you can come up with some sort of heuristically programmed algorithm that will sort through a free-form citation and clean it up. then we just have to live with rules. If editors can't take how the templates work then they have other options. --— Gadget850 (Ed)talk 22:13, 27 March 2013 (UTC)
Hey, I did not invent the way people think about computers, I just spent 5-7 years in universities, to learn what issues to beware. And "People are the way they are". Working with computer users is like a complex chess game, and to "win" their approval, there are a lot of complex issues to consider, such as trying to make complicated operations seem simple to them, without over-complicating the underlying simplicity or freedom of choices. -Wikid77 (talk) 23:20, 27 March 2013 (UTC)
Accessdate needed for doi/PMID/bibcode URL links
I found the need for the unrestricted "accessdate=" parameter, in the first article I checked ("1843 in science"), which ironically, links to an article about the "first computer program in history" (Ada Lovelace and Charles Babbage), but the cite links the webpage using a doi-parameter to generate a URL address, so the accessdate disappeared (when no "url="). Compare:
Cite journal comparison
Wikitext
((cite journal|accessdate=2010-10-01|doi=10.1109/MAHC.2003.1253887|first1=John|first2=Jo|issue=4|journal=[[IEEE Annals of the History of Computing]]|last1=Fuegi|last2=Francis|month=October–December|pages=16–26|sandbox=true|title=Lovelace & Babbage and the creation of the 1843 'notes'|volume=25|year=2003))
The easiest fix is to "keep it simple" and not have any restriction to hide the "accessdate=xx" parameter. That would be another benefit of using the Lua version of the {cite_*} cites. -Wikid77 23:20, 27 March 2013 (UTC)
What does the access date mean to you? --— Gadget850 (Ed)talk 23:26, 27 March 2013 (UTC)
accessdate is when I or anyone else last visited the url. I hardly come across the doi etc. parameters, but aren't they as fixed as isbn, even more so, so why would they need any associated date? I'm not sure how your examples differ - I think you should subst: them in so they're fixed. Incidentally, I always specify dates as ((date|date)) because I understand that then emits date according to user's local style sheet. May also translate to user-language, I'm not sure.
Apparently, people think the "accessdate=" is when they read the webpage, even if they did not know/want the URL address to include it. Perhaps that explains why many of those 40,576 people put "accessdate=" in their cites, with an empty "url=" because they are stating they accessed the page on that date. See: examples among the few 40,576 pages already in "Category:Pages using citations with accessdate and no URL". There might 90,000 people who also put the lone accessdate in {cite_web}. That is just the way people think. -Wikid77 07:29, 28 March 2013 (UTC)
I just noticed ((date)) being used in a citation. No it does not and cannot use the user date preference. With no other parameters it formats the date as DMY, thus it is rather useless here. Date formatting was dropped years ago since it relied on logged in user preferences and gave a mix of formats for casual readers. --— Gadget850 (Ed)talk 00:59, 28 March 2013 (UTC)
Perhaps leave the lone accessdate until {cite_web} counted: This is an opportune time to allow the lone accessdate to remain hidden (when no "url="), until after Template:Cite_web is transitioned to Lua, to category-count all the people/pages where "accessdate=" is used without the URL parameter, among those 1.3 million {cite_web} pages. Then, based on "overwhelming demand" we can report that a consensus of "95,000" editor teams wanted to show the lone accessdate, and then perhaps sample among those thousands as to how many use "doi=" and "accessdate=" together. Remember, once {cite_web} uses Lua, then all those 1.7 million articles can be reformatted, to show accessdate, 4x times faster than any prior change to {Citation/core}. Optionally, we could upgrade to show the requested lone accessdate, but still log those pages in the Category, and later re-change to hide some of those lone accessdates. Think of this as: 3 extra chances to add new features, to the Lua cites, before incurring the overhead of 1 old update to {Citation/core}. Because once {cite_web} uses Lua, then general editing will reformat thousands of pages for "free" with the 3x faster edit-saves. -Wikid77 07:29, 28 March 2013 (UTC)
Let's remember that the purpose of the citation is to identify the source, and each element should help in that goal. The access date is intended only for web pages that change frequently (e.g. Wikipedia main page) or have no discernible date, and has been documented that way from inception. The date someone added a citation does not help us in identifying the source. A Bibcode, doi, JSTOR or similar link does not need an access date as the content of those links by design do not change.
Simply because editors use a field in some odd manner does not mean that we should automatically assume that there is some sort of gestalt consensus to change the use. Most casual editors will not read the documentation and will simply follow what they see in an extant page, thus bad practices are propagated. We have already seen that both page and pages are included, where the editor believes that pages is for the total number of pages in a work, which is not needed to identify the source. --— Gadget850 (Ed)talk 09:33, 28 March 2013 (UTC)
Yes, yes, this. 95,000 examples of improper use out 1.605 million or 1.205 million (which ever number is correct?) does not constitute consensus. Where is the discussion?
I think that Gadget850 is inadvertently helping to make my case for visible error messages (Error trapping and checks and Format but no URL) because without visible cues to the contrary, editors assume that improper use is proper and so continue with and adopt new bad practices.
—Trappist the monk (talk) 13:31, 28 March 2013 (UTC)
Error checking is good, but we need to spec it out and link to help pages for each specific issue. --— Gadget850 (Ed)talk 14:06, 28 March 2013 (UTC)
I've been thinking exactly that and will probably open a discussion to talk about messaging, help pages, linking from the error, etc when I return from spring break.
—Trappist the monk (talk) 23:38, 28 March 2013 (UTC)
make sure to include error-checking for when software exceeds the bounds, with overly-restrictive recommendations and the related generated errors. 70.19.122.39 (talk) 15:43, 29 March 2013 (UTC)
The Chicago Manual of Style (16th ed., p. 657–8) discusses access dates for electronic sources in general. Such a date might include electronic sources not accessible through a URL. For example, help files from software that is constantly updated. While some sort of version number might be available, the editor might not know how to access such a number, or it might be so difficult that the editor would feel the reader would be unable to access it or understand it. I admit this would be a fairly unusual occurrence. Jc3s5h (talk) 09:54, 29 March 2013 (UTC)
a version number is an edition, and should probably carry some kind of date. online sources are a subset of electronic sources, and internet/intranet sources a further subset of online sources. for such sources accessdate is the de facto edition date, if there is no other info available. 70.19.122.39 (talk) 15:43, 29 March 2013 (UTC)
Online sources are always subject to computer-dependent revision at any time: Although stored documents might seem theoretically static, for the displayed contents, the broader reality in data processing is that all documents are subject to change when retrieved, or dynamically reformatted, from the document database(s). The problem of date-stamped materials has been so flexible that computer software, for years, has included a "build version" beyond just a computer program version number, where the contents of a "static" version might change because the underlying runtime library might have changed to alter operation of the upper-level software. For document storage, the underlying database might alter, or truncate, data in various ways, depending on unknown bugs in the current release, perhaps upgraded last week, of the database system used to display the stored document. For those reasons, an accessdate parameter helps to indicate which database or text-formatting product versions, at a point in time, were active in displaying the cited document, especially if national laws required censorship, of prior documents, for new content restrictions. -Wikid77 (talk) 19:07, 29 March 2013 (UTC)
If an online document was subtly changed and the visible date did not change, how would you or another editor know it? --— Gadget850 (Ed)talk 19:23, 29 March 2013 (UTC)
well this can get complicated. if one drills down, the effects of caching may also have to be accounted for. personally i tend to think that in situations where the document (a) has not changed and (b) it is not retrieved from a cache, every access could be functionally considered a "reprint". 70.19.122.39 (talk) 00:39, 30 March 2013 (UTC)
Discussion regarding the usage and cleaning of accessdate
As this discussion seems to have puttered out without really reaching a broad consensus, I've decided to try taking the issue to a somewhat broader forum, as well as discussing what to do next. Please see the community discussion at Wikipedia:Village pump (policy)#Use of accessdate in citation templates about the appropriate usage of the |accessdate= parameter and what steps (if any) should be taken to clean it up. I invite your comments at that page. Dragons flight (talk) 03:57, 29 April 2013 (UTC)
Discussion regarding the usage and cleaning of accessdate
[edit]Regarding ((cite book)) – Kerαunoςcopia◁galaxies 04:12, 11 April 2013 (UTC)
Is this new? That's very, very strange to me. What's wrong with using the "ed." abbreviation, even with an author? – Kerαunoςcopia◁galaxies 04:12, 11 April 2013 (UTC)
This is not new, per the template documentation:
If authors: Authors are first, followed by the included work, then "In" and the editors, then the main work.
If no authors: Editors appear before the included work; a single editor is followed by "ed."; multiple editors are followed by "eds."; more than three editors will be followed by "et al., eds."
Well, I saw that in the documentation, but I don't recall seeing "in" in any college papers. That just seems like a weird thing to say. Author, in another person. I've always seen "ed.", whether there's an author or not, so I'm wondering where this rule stems from. – Kerαunoςcopia◁galaxies 08:16, 11 April 2013 (UTC)
I can't answer for the template creators, but this is generally in line with APA Style.[6] --Gadget850 (Ed)talk 15:43, 12 April 2013 (UTC)
See, those examples are normal. (ed.) and (eds.) are used to signify editors. "In" is just bizarre (I wasn't clear, but I meant that I didn't know where the "In" rule was coming from); I have never seen that until now. I suppose the only option would be to write the reference out manually and not use the template. – Kerαunoςcopia◁galaxies 17:56, 12 April 2013 (UTC)
But, if the article uses Citation Style 1, then you need to continue to use it unless you gain consesnsus to change all of the the citations to a different style. See WP:CITEVAR. --Gadget850 (Ed)talk 18:11, 12 April 2013 (UTC)
Historically, the templates have always used "In" with books that had authors, editors, and named chapters:
Doe, John (1945). "My chapter". In Brown, Mary. My book.
If the book had authors and editors, but no chapter title, then "ed." was historically used:
Doe, John (1945). Brown, Mary. ed. My book.
However, this creates a problem with the lack of a clear demarcation between the end of an author list and the start of an editor list if the date is missing:
Doe, John. Brown, Mary. ed. My book.
So, in the new Lua templates, "In" is used whenever both authors and editors are specified, regardless of whether a chapter title is present:
Doe, John. Brown, Mary (ed.). My book.
Lastly, I would note that if a book has no authors, then editors are still marked with "ed.":
Brown, Mary, ed. (1945). My book.
The style guide linked above suggests using both "In" at the beginning and "Ed." at the end to demarcate editors in edited volumes that contain both authors and editors. Personally, I wouldn't be opposed to adding the "Ed." at the end as well. Dragons flight (talk) 18:23, 12 April 2013 (UTC)
((cite news)) uses Citation Sytle 1, and this produces awkward references for news articles that have both an author and an editor, for example:
Is this really correct for a newspaper citation? —RP88(talk) 12:44, 24 April 2013 (UTC)
The word "in" should only be used between the description of a component with a named author and a containing work with a named editor. It is not typical to name both an editor and an author for exactly the same work; indeed, RP88's example one would normally only cite the author, not the editor. If a case were discovered where it would be plagiarism to not site both the author and the editor, one would hand-write a citation because apparently citation style 1 does not support citing both an author and editor for exactly the same work; it only supports citing the author of a contained work and the editor of the containing work.
Of course all the stuff in WP:CITE about preserving the existing style of an article is no excuse to suppress information or commit plagiarism, so the preserving style stuff must be ignored when you need to cite a work that is not supported by a template. Jc3s5h (talk) 13:02, 24 April 2013 (UTC)
Yes, the "In" arises not from having an editor, but from a "work" being contained within a larger work. I seem to recall once seeing an article that had several authors, and two editors named as such. (Perhaps showing they were not responsible for the substantive content?) But most certainly that is not typical. ~ J. Johnson (JJ) (talk) 17:58, 24 April 2013 (UTC)
So, from my understanding, there's the case where there is an editor for the same work, as might be true for a newspaper articles or books, and then there is an editor for the "larger" (greater) work such as might be the case with an encyclopedia or a journal. Doesn't that suggest we should have two sets of parameters for it? I'm not sure I would know how to distinguish the two, but I have at least been using the parameter incorrectly. --Izno (talk) 19:42, 24 April 2013 (UTC)
Well, on second thought, what if it is the chief editor who is editing a newspaper article? I could easily claim that he certainly serves as the editor to the greater work. Just musing. --Izno (talk) 19:44, 24 April 2013 (UTC)
This is clearing things up for me. If I understand correctly, a book with an author and an editor shouldn't have the "In" line because only the author should be credited, or if the editor should be mentioned, it should be done manually; but a work by an author that is "contained within a larger work" would use "In", and this actually does make sense to me now. – Kerαunoςcopia◁galaxies 19:48, 24 April 2013 (UTC)
This is the first instance I have encountered where a newspaper article has an editor. I don't see the need for the editor, and I see no connection to plagiarism. --Gadget850talk 20:03, 24 April 2013 (UTC)
I am unable to think of an example where exactly the same work would have both an author and an editor, and the editor needs to be cited in order to avoid plagiarism. But I won't guarantee there isn't some such work lurking out there. Jc3s5h (talk) 20:45, 24 April 2013 (UTC)
It isn't at all about plagiarism, it is entirely about correct attribution. (Even an incorrect or an "anonymous" attribution technically avoids plagiarism.) The normal use-case for "in" would be something like an advanced reference deskbook or conference proceedings, where various chapters have different authors. Sometimes only the editors are show in the cataloguing data for the book, making it helpful in locating it among similarly-titled works in open stacks, yet the cited chapter in the work is due to a specific named author.
Historically (not so much for the Web) all published works had editors. They are not cited where they are essentially part of the publisher. They are cited (as editors) where they (and not the publisher) have assembled (and are responsible for) a work. ~ J. Johnson (JJ) (talk) 22:01, 29 April 2013 (UTC)
Authorformat parameter
A new parameter named authorformat was introduced during development of the LUA version.
This allows the display of author names to be somewhat decoupled from how the name is entered, and provides consistency: I particularly requested it because some planned instances of ((cite doi)) were in places where different styles were required (I didn't edit-war, honest ;-).
Allowable values at present are:
vanc which applies a "Vancouver-like" style: first parameters are reduced to a set of initials without full-stops; author parameters are left untouched to allow for names in different configurations
scap which applies a small caps effect to each author
There is a corresponding parameter named editorformat which performs as one might expect on the corresponding editor-related parameters.
These parameters require documentation. I think the above is roughly correct text-wise, but needs formatting.
If someone with more familiarity with how to navigate the tangled strands of the documentation templates can add it in the correct place, I would be obliged.
HTH HAND —Phil | Talk 15:41, 17 April 2013 (UTC)
I don't think it is that tangled, except for just adding a 'lua' parameter to several of the doc sets so we can start documenting the updated templates. Regardless, these are display options, thus they belong in ((Citation Style documentation/display)).
As new display features, what are the guidelines for their use? When would I use the Vancouver style, which already has a set of little-used templates as documented at Citation Style Vancouver? Ditto for small caps. --Gadget850 (Ed)talk 15:58, 17 April 2013 (UTC)
I have been using them solely to make the display conform to existing practice while allowing me to record names properly: I have encountered some inexplicable hostility in the past from people who do not like their articles disturbed for the sake of conforming to some style or other which applies in the world outside Wikipedia. My intent is to record full names where possible so that prolific authors can be detected and articles created for them, same as for frequently-cited journals. HTH HAND—Phil | Talk 17:19, 17 April 2013 (UTC)
Phil: My SWAG is this involves academic journals where the editors use first name initials and probably shorthand journal names? --Gadget850 (Ed)talk 17:37, 17 April 2013 (UTC)
Pretty much, yes, it involves those who think that simply copying and pasting the names into a single authors parameter is better than splitting them out into individual entries and investigating the possibility of linking them to an appropriate article, because it conforms to a style used by certain (but not all) publications which have no actual bearing on how we do anything here on Wikipedia. Sorry, but if I sound jaded and bitter, it's the number of years I have wasted on this kind of "discussion" when we could have been collecting and storing useful data. HTH HAND —Phil | Talk 20:36, 17 April 2013 (UTC)
What I would like to see is a way to specify the authors' first and last names using the first?= and last?= parameters, but still display them in the format typically used in the author= parameter: "firstname lastname" etc. Splitting the name into first and last names is a good idea for sorting and searching, however, this should be decoupled from a certain display style. One scenario, where it might be useful to force the template to display the name in "firstname lastname" order despite being given using first= and last= parameters, is when the majority of the other references in an article use this style. Also, in the long run, it might be useful to make this configurable in order to allow registered users to specify the desired display order in their personal preferences. It looks like this new authorformat= parameter would be a good place to add this kind of configurability. --Matthiaspaul (talk) 00:48, 21 April 2013 (UTC)
CS1 uses last, first. What would you call this style set? --Gadget850 (Ed)talk 11:34, 21 April 2013 (UTC)
Hm, perhaps something trivial like natural (assuming "firstname lastname" is the natural order to write names in Western locales). --Matthiaspaul (talk) 12:31, 21 April 2013 (UTC)
For Citation Style Natural, we will need a help page to define it and a list of the new templates. --Gadget850 (Ed)talk 12:36, 21 April 2013 (UTC)
To get closer to the "Vancouver-like" style, one also needs to add author-separator and author-name-separator parameters:
This syntax seems unnecessarily verbose. In my opinion, if "authorformat = vanc" is specified, then author-separator and author-name-separator should also be automatically set as above without having to explicitly include these two parameters. Thoughts? Boghog (talk) 06:39, 30 April 2013 (UTC)
'Cite web' used more than once.
How to use
Could you please help me with the situation that a 'cite web' is used for more than once? How to shorten the cite?
Thanks, New worl (talk) 08:55, 30 April 2013 (UTC)
Change the <ref>((cite web ... ))</ref> to <ref name="Vatican guesthouse">((cite web ... ))</ref> and then later on you can put <ref name="Vatican guesthouse" />
Markup
Renders as
Pope Francis does not live in the Papal apartments.<ref name="Vatican guesthouse">((cite web |url=http://ncronline.org/news/vatican/pope-francis-live-vatican-guesthouse-not-papal-apartments|title=Pope Francis to live in Vatican guesthouse, not papal apartments|first=Cindy |last=Wooden|date=26 March 2013 |accessdate=26 March 2013|work=[[National Catholic Reporter]]))</ref>
Something else about the Pope.
The Pope lives in the Vatican guesthouse.<ref name="Vatican guesthouse" />
Pope Francis does not live in the Papal apartments.[1]
The part of the name= inside the quotes can be almost anything you like, but must be unique within the article. --Redrose64 (talk) 09:48, 30 April 2013 (UTC)
Thanks Redrose64 very much. New worl (talk) 10:55, 30 April 2013 (UTC)
I like to use author date. In this case wooden2013. --Gadget850talk 12:16, 30 April 2013 (UTC)
Thanks Gadget850 for the nice tip. New worl (talk) 13:32, 30 April 2013 (UTC)
A shortcut to insert 'Cite web'
Could you please let me know if we have a shortcut to insert 'Cite web'? I'm new so any shortcut is very helpful. Many thanks, New worl (talk) 11:05, 30 April 2013 (UTC)
Just above the edit box you should have a strip of buttons and menus which looks like the example that I've added here. The right-most drop-down menu, "Cite", should have the function you require. I don't use it myself; but there is more at Wikipedia:RefToolbar/2.0. --Redrose64 (talk) 11:48, 30 April 2013 (UTC)
Oh, got it. Many thanks, Redrose64. New worl (talk) 11:59, 30 April 2013 (UTC)
I just heard of Wikipedia:VisualEditor. Do you know if it will replace "Cite web", < ref >, etc. at least for editors with limited coding experience? Many thanks, New worl (talk) 12:04, 30 April 2013 (UTC)
It's a new front-end, offering different means for entering everything from italics and links to references, but the underlying Wikicode will remain the same. I shan't be using it myself: I dislike WYSIWYG web editors, since they always produce code bloat (Microsoft FrontPage was a particularly horrendous example). --Redrose64 (talk) 13:05, 30 April 2013 (UTC)
Thanks Redrose64. I think you are right. The problem is that it is not easy to have a common tool for all because there are so many preferences, strengths, etc. amongst individual editors. New worl (talk) 13:23, 30 April 2013 (UTC)
Proposal for comment parameter
I would like to suggest the introduction of a comment= parameter to the various ((Cite)) templates. (The equivalent templates in the German WP have a similar kommentar= parameter already.)
This could be displayed at the end of the cite info similar to quote=, but perhaps in a smaller font or framed with "(NB. ...)", "(Note: ...)" or something similar.
The rationale for such a parameter is that it is sometimes necessary or helpful to comment on a provided source, for example, if some information in a source is used as a reference, but other information in the source is known to be false or misleading, or if there is a typo in the title, the author's name or other bibliographic data, or if something else needs to be explained about the reference, which does not belong into the article's text, but could confuse or mislead a reader, if not mentioned at all while using the reference.
While it is possible to provide this information inside the <ref> environment, but outside the ((cite)) template, having a dedicated parameter for it would help to bring such comments into a stringent format and improve maintainability in the future.
I think that it would be more reasonable to display the specified language just after the title but not after the "website" field. The reason is that we specify the language of the particular page, but currently a reader may think about it as the language of the whole web-site which may have a different language generally. --DixonD (talk) 18:38, 1 May 2013 (UTC)
I'm not sure that moving the language element is necessary. Doesn't the reader know that the leftmost element in the citation is the thing to which every element to the right refers? In the case of a ((cite book)) citation to a chapter, |title=, |publisher=, |date=, |page=, |language=, etc all refer to the value in |chapter=.
Says: "title: Title of web page." Maybe clarify that this is about the <title>, or - if it's not - the article title on the web page, or that the Wikipedia editor should pick the one that fits best if the <title> and article title on the web page are different. What if, for example, the <title> says "My Website" and the article title on the web page is "Ten ways to improve your memory" (or vice versa)? What if the <title> says "Best way to lose weight" and the article title on the web page says "Why Ireland has no snakes"? --82.170.113.123 (talk) 17:50, 28 April 2013 (UTC)
Changed to "Title of source page on website". Also moved the note about linking to 'title' where it belongs. --Gadget850talk 18:10, 28 April 2013 (UTC)
Thanks. A question, merely out of curiosity and this is purely hypothetical... if, for example, the title is "My Website" and the article title on the web page is "Ten ways to improve your memory", and I want to use the page as a reference/source for text somewhere on, say, the Memory page, I should literally add "|title=My Website" as part of the cite web? --82.170.113.123 (talk) 18:20, 28 April 2013 (UTC)
I'd use whichever will be most helpful to the person reading the Wikipedia article. -- John of Reading (talk) 19:19, 28 April 2013 (UTC)
There is a reason that CS1 has |title= and |website= (aka |work=). In IP Editor's example, the citation is:
((cite web |url=http://example.com |title=Ten ways to improve your memory |website=My Website |accessdate=2013-04-28))
Essentially, this is much like citing a chapter in a book:
((cite book |chapter=Ten ways to improve your memory |title=My Website))
→"Ten ways to improve your memory". My Website.
—Trappist the monk (talk) 19:32, 28 April 2013 (UTC)
That's just because of my example though. What about the other example I gave, where the title doesn't say "My Website" but another relevant text that isn't about the website in general. See, cite web is used a lot throughout Wikipedia, so maybe a broader discussion about this is desirable? Because I agree with what John of Reading writes about whichever will be most helpful, but because of my comments here the official text now says "Title of source page on website"... --82.170.113.123 (talk) 19:36, 28 April 2013 (UTC)
Example please? My intial thought is that any website that does not clearly identify itself may not be a good source. Do we need further instruction on how to identify the title of a website? --Gadget850talk 20:15, 28 April 2013 (UTC)
[7], [8], [9], [10] - examples where the displayed page title was preferable to the hidden HTML title. -- John of Reading (talk) 20:39, 28 April 2013 (UTC)
I didn't have any trouble finding page titles and website names for any of those citations except for Melling, and that citation (such as it is) hardly supports the claim in the article.
"J.D. Chakravarthy Profile". Telugu Movie Talkies. Retrieved 2012-09-15. – changed |publisher= to |website=;
"Minister for Economics – Daniels Pavļuts". The Cabinet of Ministers of the Republic of Latvia. 1976-05-14. Retrieved 2012-08-11. – deleted |publisher=, split old |title= into |website= and |title=, added Minister for Economics
—Trappist the monk (talk) 21:25, 28 April 2013 (UTC)
So from these examples can I infer that you happy for the citation not to copy the HTML "<title>" element from the cited web page, in cases where other text from the page makes the citation clearer for the reader? I think that was the OP's point. -- John of Reading (talk) 09:34, 29 April 2013 (UTC)
Yep. I am happy for an editor to find and use a human readable/understandable title. <title> may be a perfectly good title source for whatever bots go around seeking such things, but human editors know better than bots what is appropriate for |title=. As far as I can determine by looking back through the ((cite web/doc)) history, <title> has never been required nor even recommended as the value to be used with |title=.
However, editors should not be inventing titles which implies that <title> may have to be used regardless of how strange it may be.
—Trappist the monk (talk) 12:56, 29 April 2013 (UTC)
So, that was it, I guess. It'll say "title: Title of source page on website" until someone else brings it up again. --82.170.113.123 (talk) 17:28, 1 May 2013 (UTC)
Did you have another suggestion? --Gadget850talk 17:45, 1 May 2013 (UTC)
I'm not a native English speaker, but what I have in mind is something like "Relevant text between the <title> and </title> tags (the title in the browser toolbar), or, if more appropriate and helpful to readers of the Wikipedia article, the ...". Something that: 1. prevents editors from thinking they can inventing a title, 2. more clearly explains what the first option is (not "Title of source page on website" but show the < and > symbols and mention the word "tag" and say it is the title in the browser toolbar), and 3. also makes clear that using the main title on the web page itself is allowed/preferable if the <title> isn't very usable. --82.170.113.123 (talk) 18:45, 1 May 2013 (UTC)
I think it's appropriate to allow editors some discretion in choosing which title to use: it could be the text in the title tags, it could be the text in the most significant titling tag of the rendered page, or it could be some substring of those. I don't think requiring it to match the full title tag text is a good idea. Another case that frequently comes up is that the title tag text includes the name of the larger web site within which the page exists, which we would normally put into the |work= parameter. In that case, what we put into the title should be the rest of the title tag text, not including the work part. (Example: this page should have |title=Archimedean Solid and |work=MathWorld, not |title=Archimedean Solid -- from Wolfram MathWorld.) —David Eppstein (talk) 19:04, 1 May 2013 (UTC)
(edit conflict)Discerning the title of a web page or site may require some judgment. The <title> can be useful but may contain other information. For example, the <title> for this page is "Help talk:Citation Style 1 - Wikipedia, the free encyclopedia". It is obvious that this includes both the page and site titles and a description, but any automated tool is going to have trouble. --Gadget850talk 19:13, 1 May 2013 (UTC)
I can't edit Template:Cite web#Title, plus I'm not a native English speaker. If anyone feels like improving that's on the page, after the discussion in this thread, please go fo it. Thanks. --82.170.113.123 (talk) 17:02, 4 May 2013 (UTC)
Apparently, when you use the template to cite a book that has an author and an editor, it assumes it's a work in an edited volume. Which is to say, in the citation, it puts it as "[Author name]. In [Editor name]..." Is there a way to correct this so it reads "[Author]. Book title. Edited by [Name]", as is the standard per CMoS? Or do I need to scrap the template and just enter it by hand? Parsecboy (talk) 20:22, 5 May 2013 (UTC)
There is some discussion of this issue at this topic above. Can you provide an example of the problem citation?
Ms. Forster wrote the book and Mr. Norris is credited with editing it. Imzadi 1979→ 20:53, 5 May 2013 (UTC)
This comparison suggests that Module:Citation/CS1 acts differently from ((citation/core)). I have tweaked Module:Citation/CS1/sandbox so that if there is a date, 'In' is omitted and 'ed.' or 'eds. is appended. No doubt the way I accomplished this is crude and inept and I expect Editor Dragons flight shudder and then make it right or revert.
Cite book comparison
Wikitext
((cite book|editor-first=Joe L|editor-last=Norris|first=Edith C.|last1=Forster|location=Detroit|oclc=3324319|publisher=[[Wayne State University Press]]|sandbox=yes|title=Yesterday's Highways: Traveling Around Early Detroit|year=1951))
Live
Forster, Edith C. (1951). Norris, Joe L (ed.). Yesterday's Highways: Traveling Around Early Detroit. Detroit: Wayne State University Press. OCLC3324319. ((cite book)): Unknown parameter |sandbox= ignored (help)
Sandbox
Forster, Edith C. (1951). Norris, Joe L (ed.). Yesterday's Highways: Traveling Around Early Detroit. Detroit: Wayne State University Press. OCLC3324319. ((cite book)): Unknown parameter |sandbox= ignored (help)
Imzadi1979, please explain why you want to cite both the editor and the author. Ordinarily only the author is cited. Amazon has a picture of the book cover, and only the author is listed on the cover. What's so unusual about this editor? Jc3s5h (talk) 23:05, 5 May 2013 (UTC)
A library catalog is different from a citation and there is no reason to put every bit of information from a library catalog into a citation. Jc3s5h (talk) 23:21, 5 May 2013 (UTC)
Full date not showing in "cite web" template
The ((cite web)) template automatically generated from the "Templates" menu above the edit box appears to be operating incorrectly. Currently, it has separate provisions for "date", "month" and "year". Entering these as separate elements produces this tag:
<ref>((cite web|title=Golden Era Records|url=https://www.facebook.com/permalink.php?story_fbid=10151597119869292&id=19129474291|work=Facebook|date=6|month=May|year=2013|accessdate=6 May 2013|quote=Head over to Hilltop Hoods page and wish Suffa a Happy Birthday!))</ref>
Unfortunately, only the "date" (i.e., 6) is shown in the citation in the article; the "month" and "year" are ignored.
"Golden Era Records". Facebook. 6. Retrieved 6 May 2013. "Head over to Hilltop Hoods page and wish Suffa a Happy Birthday!"
The solution is to enter the full date (i.e., 6 March 2013) in the "date" field and leave the "month" and "year" fields blank, but this is somewhat counter-intuitive. Why are the other boxes there? —sroc (talk) 02:24, 6 May 2013 (UTC)
Per the documentation:
date: Full date of source being referenced in the same format as other publication dates in the citations. Do not wikilink. Displays after the authors and enclosed in parentheses. If there is no author, then displays after publisher.
OR: year: Year of source being referenced.
month: Name of the month of publication. If you also have the day, use date instead; do not wikilink.
This was implemented this way because of limitations in the #time parser function. --Gadget850talk 02:35, 6 May 2013 (UTC)
Can't really speak to the functionality of the templates tool you are using because it is not part of CS1. The citation that it created acts the way it does because |year= is a synonym of |date=. In the citation, |date= has priority over |year=. If |date= is not present but |month= and |year= are, then CS1 will concatenate them to make a date field.
The |day= parameter is deprecated and |month= is likely to be deprecated. Use |date= and ignore |month= and |year=
Thanks, Trappist the monk. I'm remember this from now on and will try to patch up my past mistakes with this one. I've raised this at Wikipedia talk:RefToolbar#Dates in cite templates so hopefully this can be fixed to avoid others following in my folly. —sroc (talk) 02:48, 6 May 2013 (UTC)
It's happened again! I was beginning to think that the problem of posts to the same thread causing the other's post to disappear had been resolved. Apparently not. I have restored Editor Gadget850's post.
If authors: Authors are first, followed by the included work, then "In" and the editors, then the main work.
If no authors: Editors appear before the included work; a single editor is followed by "ed."; multiple editors are followed by "eds."; exactly four editors will show three editors followed by "et al., eds."
Thanks; I obviously didn't read the documentation carefully enough. Sasata (talk) 16:33, 10 May 2013 (UTC)
Cite episode deprecated parameters
Editor AussieLegend has reverted this edit. With this edit, Editor AussieLegend added new text. This is the text that, prior to its deprecation, described|episodelink=.
|episodelink= was used to link |title= to a Wikipedia article. But, |episodelink=, unlike |serieslink=, has no matching |episodetitle= parameter. If the citation includes |url= and |episodelink= the resulting citation title looks like this (red link because there is no Wikipedia article titled "Episode link"):
((cite episode |title=Episode title |episodelink=Episode link |url=http://www.example.com))
Use of |episodelink= in this manner breaks |title= when a proper |url= is part of the citation and, when |url= is not part of the citation, serves simply as a wikilink to another article. This latter is a misuse of the citation because WP:NOTRS, particularly WP:CIRCULAR. Citations like those in Major villains in Charmed and other articles should be replaced with wikilinks in article text and proper citations to reliable sources. |serieslink= is not required because its matching parameter, |series=, can and should be wikilinked. For these reasons Editor AussieLegend's changes should be reverted and I have done so.
—Trappist the monk (talk) 14:33, 11 April 2013 (UTC)
This is not actually an accurate interpretation. The intention of |episodelink= is simply to provide a link to a Wikipedia entry on the episode. Per MOS:TV, specifically WP:TVPLOT, television episodes are acceptable primary sources. We link to the episode article because it's not possible to link to the actual episode that aired. When corrrectly used, ((cite episode)) should ideally include use of either the |minutes= or |time= parameters. For example:
((cite episode|title=Red-2|episodelink=#ep91|series=NCIS: Los Angeles|serieslink=NCIS: Los Angeles|network=[[CBS]]|date=March 26, 2013|season=4|number=19|minutes=02:21)) produces "Red-2". NCIS: Los Angeles. Season 4. Episode 19. March 26, 2013. 02:21 minutes in. CBS. ((cite episode)): Unknown parameter |episodelink= ignored (|episode-link= suggested) (help); Unknown parameter |serieslink= ignored (|series-link= suggested) (help)
The documentation as I modified it simply reflects the way that the template works. It does not change the way that the template works; |episodelink= is still a functional parameter after Trappist the monk's removal of the documentation, which was modified around this time last year to change the description of |title= to "Title of source. Can be wikilinked to an existing Wikipedia article or url may be used to add an external link, but not both. Displays in quotes." (This was how the parameter was deprecated. There doesn't appear to have been any discussion about deprecating either |episodelink= or |serieslink=.) Despite the "but not both" claim, it is still possible both link to an article in the title and use a url. Doing so results in:
This results in the same peculiar output as Trappist the monk's example, where both the url and the wikilink are shown, with the url preceding the wikilink. Argue as you may that it won't, history shows that the likelihood such a citation will be added is high - I'm always fixing incorrect uses of this template. The solution would seem to be simply to modify ((cite episode)) so that |url= overrides |episodelink=; hiding the problem with documentation hacks is not the way to do it. I suspect that the reason this was never done previously is that, being a broadcast medium, there rarely is a url to a particular episode. Other than a recap on an official site, any such url is generally a linkvio and most urls don't actually seem to support the information that is being cited. There is no explanation as to why serieslink has been deprecated. It too is still a fully functional parameter. --AussieLegend (✉) 16:15, 11 April 2013 (UTC)
Stepping back from the issue of what it does output, which is pretty clearly malformed, do you have an opinion on what ought to happen if both episodelink and url are specified? If you do something like this with the new Lua citations you get:
((cite web|title=Episode title |titlelink=Episode link |url=http://www.example.com))
Which is arguably better than the present format, but probably still bad because there is no description on the URL. Dragons flight (talk) 17:07, 11 April 2013 (UTC)
As I said, "The solution would seem to be simply to modify ((cite episode)) so that |url= overrides |episodelink=". In the event that a url is available, you should see "Episode title". ((cite episode)): Missing or empty |series= (help), which is easier to do if we continue to use |episodelink=. --AussieLegend (✉) 17:24, 11 April 2013 (UTC)
Having said that, I've discovered a now archived proposal by 117Avenue that may be an alternative. I haven't checked it out thoroughly, but you can find it here. --AussieLegend (✉) 17:32, 11 April 2013 (UTC)
(edit conflict)|serieslink= was and should be deprecated because:
|series=NCIS: Los Angeles|serieslink=NCIS: Los Angeles
is the same as:
|series=[[NCIS: Los Angeles]]
This works because |series= is not a synonym for another parameter and technically identical to how you implemented |network=.
How about this? Wikilink to the Wikipedia episode article through the |number= parameter. Doing that shows how |episodelink= isn't necessary, reduces duplication, and leaves |title= free for use with |url=. Here is a tweaked version of your example:
((cite episode|title=Title for link to external NCIS: Los Angeles (season 4) site|url=http://www.example.com|series=[[NCIS: Los Angeles]]|network=[[CBS]]|date=March 26, 2013|season=4|number=[[NCIS: Los Angeles (season 4)#ep91|19]]|minutes=02:21))
Per MOS:TV, specifically WP:TVPLOT, television episodes are acceptable primary sources. Well, yeah ... Except that it states in the same paragraph: "Since the episode is the primary source and the infobox provides details about it, citing the episode explicitly in the plot summary's section is not necessary." I think that this supports my position that the citations in articles like Major villains in Charmed should be reduced to wikilinks in the text rather than malformed "citations".
Here is a ((cite web)) version of the tweaked ((cite episode)) citation above:
((cite web|title=Title for link to external [[NCIS: Los Angeles]] (season 4) site|url=http://www.example.com|series=[[NCIS: Los Angeles]]|network=[[CBS]]|date=March 26, 2013|season=4|number=[[NCIS: Los Angeles (season 4)#ep91|19]]|minutes=02:21))
The reason I show you this is because this is how Module:Citation/CS1 treats the case where |title= has both |url= and a competing wikilink. This is the future. CS1 will eventually handle all of the 23 or so separate Citation Style 1 templates.
—Trappist the monk (talk) 17:38, 11 April 2013 (UTC)
|serieslink= is often not the same as |series=. NCIS won't get you the TV series, Hell's Kitchen won't get you to the UK TV series or the U.S. TV series. The statement "|serieslink= is not required because its matching parameter, |series=, can and should be wikilinked" is not entirely correct either. The new documentation only says, "The name of the series the episode belongs to; may be wikilinked" but the old documentation gave more detail: "If the citation is being used in the article about the series itself, this call is not necessary and will in fact create improper formatting." This is more correct - when used in the main article, the series title will be improperly bolded in the citation:
"Red-2". NCIS: Los Angeles. Season 4. Episode 19. March 26, 2013. 02:21 minutes in. CBS. ((cite episode)): Unknown parameter |episodelink= ignored (|episode-link= suggested) (help)
i.e. The series name should not be linked in some circumstances. I know from experience that when re-using a citation from a series or episode article, or episode list, which happens very often, it's a lot more convenient to strip "|serieslink=Series name (TV series)" from a citation than it is to delink [[Series name (TV series)|Series name]]. OK, it's not extremely complicated but I don't see the point in making anything more difficult than it needs to be and this has been working fine for years. |network= rarely requires disambiguation, just linking, which is why there is no |networklink=.
""Since the episode is the primary source and the infobox provides details about it, citing the episode explicitly in the plot summary's section is not necessary." I think that this supports my position that the citations in articles like Major villains in Charmed should be reduced to wikilinks in the text rather than malformed "citations"." - No, I'm afraid it doesn't. The section that you've quoted only applies to episode articles that have an infobox for the specific episode. Episode lists have a different infobox and ((cite episode)) isn't normally used in the plot section of episode lists or articles, although sometimes it is necessary. More often than not it's used in TV series main articles, character articles and even some completely unrelated articles that need to cite a television episode. It provides for the inclusion of times using |minutes= or |time=, which a wikilink doesn't. These parameters point to a specific time in the episode that supports the claim. For example, prior to airing, the NCIS: Los Angeles episode "Red-2" was listed in reliable sources as "Red: Part Two".[11] However, when it aired the title shown on-screen was "Red-2". The citation that I've been using as an example includes the time in the episode ("02:21 minutes in") where this is shown. Unlike a newspaper or website, where a reader can usually do a word search to verify a claim, this is not possible in a TV episode. A time code is needed in order to avoid having to watch a large chunk of most episodes (not all information is in the opening credits and fast-forwarding often misses the information, especially when it is spoken). --AussieLegend (✉) 17:58, 11 April 2013 (UTC)
Some sort of (edit conflict) occurred here that resulted in part of Editor AussieLegend's above post getting deleted. Not sure how that happened; there was no warning at the time I saved my post below. I think that Editor AussieLegend's has been restored,
—Trappist the monk (talk) 19:20, 11 April 2013 (UTC)
It isn't clear to me why there are any |whateverlink= parameters except for |authorlink= and |editorlink= - these because author and editor names are often broken up into given- and surnames. Your argument, as I understand it, is that creation and maintenance of disambiguated wikilinks is an onerous task. I'm not persuaded.
I'm not understanding the need to cling to ((cite episode)) and |episodelink= when all that it is accomplishing is a multiple-step link from the article text that the reader is reading: (click) to §References where the reader has to figure out which link in the citation to follow; (click) to the episode article. Tell me why using ((cite episode)) and |episodelink= is better for this than a simple wikilink in the article text where the reader is reading; (click) to the episode article.
Documentation can pretty much always be made more clear just as what we write here in this discussion can be made more clear. Perhaps I should have written "|serieslink= is not required because its matching parameter, |series=, can and should be wikilinked instead."
—Trappist the monk (talk) 18:49, 11 April 2013 (UTC)
(taps himself on the shoulder and reminds himself that the topic of the discussion is deprecated parameters)
This cite does what I think you want yet doesn't use the deprecated parameters |episodelink= and |serieslink=:
((cite episode|title=Red-2|series=[[NCIS: Los Angeles]]|network=[[CBS]]|date=March 26, 2013|season=4|number=[[NCIS: Los Angeles (season 4)#ep91|19]]|minutes=02:21))
→"Red-2". NCIS: Los Angeles. Season 4. Episode 19. March 26, 2013. 02:21 minutes in. CBS.
—Trappist the monk (talk) 19:36, 11 April 2013 (UTC)
"Your argument, as I understand it, is that creation and maintenance of disambiguated wikilinks is an onerous task. I'm not persuaded." - May I ask how many television articles you've edited and had to reference using ((cite episode))? If the answer is what I think, then I can understand why you can't see. As an editor who has edited hundreds of TV articles I can say that it is something that saves considerable time and, as I've said, it's something that has worked well for years, so I don't see what is to be gained by deprecating useful parameters.
"Tell me why using ((cite episode)) and |episodelink= is better for this than a simple wikilink in the article text where the reader is reading; (click) to the episode article." - The episode link adds information. A link to the episode entry isn't really needed in the citation at all, if editors use the |time= or |minutes= parameters, but they are parameters that are often omitted - editors can often cite something from an episode but don't have the time from the episode that something occurred. Of course we could say that most citation parameters aren't needed. All we really need is <ref>http://www.example.com</ref> but we encourage editors to add as much information as possible to citations. Where a detailed episode article exists, a link to the episode article can often provide the information needed, without forcing the reader to have to search for a copy of the episode, especially if it's something that just deals with plot information. The full citation information is still present, but the link often provides a shortcut. It's also not always practical to provide a wikilink to an episode in the prose. Sometimes the episode title just doesn't "fit" - "The warden was revealed to be a man.[10]" is preferable to "As stated in "Caged Fae", The warden was revealed to be a man.") When an episode is referenced multiple times, as often happens, a wikilink in each place generally violates WP:REPEATLINK. An episodelink in the citation avoids that.
"Perhaps I should have written "|serieslink= is not required because its matching parameter, |series=, can and should be wikilinked instead."" - No, because that's wrong too. There are times, as explained above, when | should NOT be wikilinked.
"This cite does what I think you want yet doesn't use the deprecated parameters" - It does, but I've explained the problems with that above, and you haven't convinced me that those parameters should be deprecated, remembering that never was any discussion that lead to them being deprecated. There was just one arbitrary edit that decided they should be deprecated. It was a bold edit, but it has been opposed. --AussieLegend (✉) 13:02, 12 April 2013 (UTC)
I updated the template to ((citation/core)) over a year ago and the documentation a bit later.discussion My reasoning for deprecating, but not removing these two parameters was to encourage editors to use the wikilink method of the other Citation Style 1 templates but retain the parameters for backward compatibility.
As we move forward in updating CS1 templates to use Lua, we have added a lot more error checking. For example, with ((cite book)), if you specify 'url' and use a wikilink in 'title' then there is an error message. These are not yet visible as we want the help system to be in place before editors start seeing a bunch of new errors. This particular check would not fix the issue with 'serieslink' and a wikilink in 'title', but it would give an error and place the page in a category.
I have added a feature request to check to see if a wikilink is to the current page and simply kill the link. --Gadget850 (Ed)talk 16:20, 12 April 2013 (UTC)
What you guess is probably close to right, but the number of whatever-type of article edits I've made has no bearing on whether or not |episodelink= and |serieslink= should be deprecated. Have you considered AWB for these repetitive edits?
Have you considered using ((rp)) to reference a specific event time?
However, when it aired the title shown on-screen was "Red-2".: Red-2 02:21
I included a linked title which may not be the right destination but you get the point.
In your Caged Fae example, does the reader not already know that the subject is the Lost Girl Caged Fae episode? If not, then perhaps the article text needs editing to make that clear. While it might be helpful to the reader to know that the discussion is about the Caged Fae episode, it may be inconvenient or inappropriate to include that information at that particular place in the article text. So then perhaps this:
or something similar. The sentence might need rewording to more fully accommodate the wikilink. I don't think that this violates the tenants of WP:REPEATLINK.
Since I haven't convinced you and you haven't convinced me, here we stand unresolved. So, I've been wondering if an alternate solution exists that answers the need to make reference to an episode article, event time, etc that doesn't use ((cite episode)) at all so there is no need for |episodelink= and |serieslink=. What if we were to create a referencing template ((ref episode))? It might provide <ref group=episode></ref> tags or similar so that references to episode articles would be automatically grouped together under a separate ((reflist|group=episode)). Parameters might include |title=, |number=, |minutes=, etc. No external link conflicts; |title= and |number= are wikilinkable. Not sure how to do <ref group=episode name="??" /> but I'm sure it can be figured out – ((sfn)) does it which might be forked to create ((ref episode)). Here, in this thread though, is not the place to discuss the details of this ((ref episode)) idea.
—Trappist the monk (talk) 16:03, 12 April 2013 (UTC)
If you haven't been through the experience, then it's hard to understand what it means to have the convenience. That's not a criticism of you, it just makes it hard to convey the information. For example, most people wouldn't know the advantage the pedals in a left hand drive Lamborghini Gallardo has over a right hand drive model (just trying to pick an "out there" example"). No, I haven't considered using ((rp)) because it is "is for appending page numbers in Harvard referencing style (or AMA style", not for adding time codes. ((Cite episode)) is a template specifically for citing episodes including times within episodes. We really should use the right tool for the right job, and the right tool here is ((Cite episode)).
"In your Caged Fae example, does the reader not already know that the subject is the Lost Girl Caged Fae episode?" - Yes, of course they should. I was attempting to show you how episode titles don't necessarily fit into the prose - don't overthink it.
""The warden was revealed to be a man."" - That's pretty ambiguous. Wikilinks should be obvious in their purpose. A reader is more likely to ignore such a link, thinking that it's a link to an article related to "revealed".
"So, I've been wondering if an alternate solution exists that answers the need to make reference to an episode article, event time, etc that doesn't use ((cite episode)) at all so there is no need for |episodelink= and |serieslink=." - Again, don't overthink it. The reality is that, there was never any consensus to deprecate episodelink and serieslink and they remain fully functional parameters. They've only disappeared from the documentation, again without any consensus to remove them. The template is transcluded 7,226 times and many uses include episodelink and serieslink. There is no reason why we shouldn't continue to use them. What we DO need to do is fix the issue caused when the episode title is linked (in whatever form it is linked) and |url= is present. I don't think anyone disagrees with the argument that url should override wikilinking. --AussieLegend (✉) 16:46, 12 April 2013 (UTC)
The parameters are still supported, so use them as desired. The issues can be resolved per my post just above, unless there is an objection to the Lua update. --Gadget850 (Ed)talk 3:31 am, Today (UTC+10)
No need to be patronizing. I'm not a neophyte; like most people I can use my own past experiences to understand things of which I have no direct experience.
If you'd rather not use ((rp)) in a manner different from its intended purpose, try this:
However, when it aired the title shown on-screen was "Red-2".(Red-2 02:21)
The sentence: "The warden was revealed to be a man," was an example with the caveat that it might need rewriting. Do you think that readers give superscript-links to citations more attention than they do wikilinks? I don't know, but I'd be surprised if they did.
I agree that a |url=-type parameter should have precedence over a wikilink when there is a contention for the same title and, when this occurs, CS1 should report an error.
Still, I haven't convinced you and you haven't convinced me, here we stand unresolved.
—Trappist the monk (talk) 18:27, 12 April 2013 (UTC)
You certainly seem to be having issues with deleting content. This edit completely reverted Gadget850's post from more than an hour before yours.[12] In that post, which I've restored,[13] he says continue using the parameters. Since he's the one who originally deprecated the parameters, that should carry some weight. --AussieLegend (✉) 19:29, 12 April 2013 (UTC)
Use them, don't use them- it doesn't matter as long as we get the proper citation. But, the link parameters don't do anything that wikilinking can't do. When we do the Lua update we will just fake it. The doc does need to be edited for style. --Gadget850 (Ed)talk 19:46, 12 April 2013 (UTC)
I don't delete other editors talk posts, ever. That it's happened to me twice and only in this conversation is interesting but not a willful act on my part.
Yep, |episodelink= and |serieslink= are deprecated and should stay that way.
—Trappist the monk (talk) 23:08, 12 April 2013 (UTC)
No, they are not deprecated. One editor can't make an arbitrary decision to deprecate parameters in a widely used template. There needs to be discussion and consensus to deprecate. Ironically, the editor who made the bold decision to deprecate seems to accept this. It's only you who is arguing for deprecation. --AussieLegend (✉) 15:30, 13 April 2013 (UTC)
Actually, an editor canmake an arbitrary decision to deprecate parameters in a widely used template, especially when that editor is eminently qualified to do so. |episodelink= and |serieslink= have been deprecated for more than a year. I suspect that that year, many, many others have read the documentation so you aren't the first to discover that |episodelink= and |serieslink= are deprecated. Consensus by editing, ne?
I guess you and I not only disagree about |episodelink= and |serieslink=, but we also disagree about what Editor Gadget850 believes.
As before, I haven't convinced you and you haven't convinced me, here we stand unresolved.
—Trappist the monk (talk) 16:46, 13 April 2013 (UTC)
I'm sorry, but that is not how Wikipedia works. Whether or not an editor thinks he "is eminently qualified to do so", if his edits are objected to, then you need to form a consensus to make such changes. An editor has made edits, they've been objected to, and the editor who made the edits has said "The parameters are still supported, so use them as desired". That's fairly clear to anyone, apparently except you. --AussieLegend (✉) 12:03, 14 April 2013 (UTC)
"What we've got here is failure to communicate." I did not not write, nor intend to imply, anything about what Editor Gadget850thinks he "is eminently qualified to do..." (emphasis added). Please do not put words into my mouth that I have not spoken. It is my opinion that Editor Gadget850 is eminently qualified. Without the broader Wikipedia community's consent, Editor Gadget850 would not have been allowed to make the significant, and largely under-appreciated, changes necessary to coerce 23 or so disparate CS1 templates into a nicely coherent suite. That Editor Gadget850 did this work across all of the CS1 templates is tacit consensus else, long before now, those changes would have been opposed.
I wonder if there is a conceptual misunderstanding. There is a difference between deprecated and disallowed. I have never argued that editors are now prohibited from using |episodelink= and |seriesslink=. I have only argued that |episodelink= and |seriesslink= are, and rightly should be, deprecated.
—Trappist the monk (talk) 14:02, 14 April 2013 (UTC)
With this edit and then this you've effectively "prohibited" editors from using the parameters by removing them from the template skeletons, which misleads editors into believing they're no longer valid parameters. --AussieLegend (✉) 14:58, 14 April 2013 (UTC)
Hardly. Those edits are entirely consistent with deprecation. Use of |episodelink= and |seriesslink= is not expressly disallowed nor is it encouraged.
—Trappist the monk (talk) 16:30, 14 April 2013 (UTC)
Semantics. The lack of documentation for fields gives the clear impression that they're parameters that either don't exist or shouldn't be used. The documentation does not mention them under the heading "Deprecated", although all those parameters were deprecated after discussion. Please note that the section begins with the statement "These parameters are deprecated and will no longer work", a statement that links the terms "deprecated" and "non-functional" for the average reader. --AussieLegend (✉) 13:33, 16 April 2013 (UTC)
By definition, deprecated parameters either don't exist or shouldn't be used. The documentation clearly marks |episodelink= and |serieslink= as deprecated.
Yeah, the documentation could be better. I've just started work on that because the Citation Style 1 citations are in the process of the Lua upgrade.
—Trappist the monk (talk) 11:11, 17 April 2013 (UTC)
|episodelink= and |serieslink= both exist and there's no reason why they shouldn't be used. Serieslink avoids the need for editors to manually link (semi-automation is always a good idea) and episodelink is needed for 117Avenue's suggestion below. Therefore, by your own definition, neither is deprecated. --AussieLegend (✉) 15:46, 19 April 2013 (UTC)
That outdent and your comments make it feel like we're starting all over again. I have stated why |episodelink= and |serieslink= should be deprecated. I'm not going to repeat myself ad infinitum.
While I might like to take credit for the defining deprecation, I can't, but thanks anyway.
I haven't convinced you and you haven't convinced me, here we stand unresolved.
—Trappist the monk (talk) 18:36, 19 April 2013 (UTC)
Yes, you've explained why you think the parameters should be deprecated, but you haven't shown how they are deprecated and your definition shows that they are not. Since you don't seem interested in continuing, I assume you don't have a valid opposition to restoring the documentation? --AussieLegend (✉) 11:43, 21 April 2013 (UTC)
|episodelink= and |serieslink= exist;
|episodelink= and |serieslink= are not disallowed;
|episodelink= and |serieslink= are deprecated to standardize parameter usage across the CS1 suite of citations;
|episodelink= and |serieslink= are deprecated because wiki markup in |title= and |series= makes them redundant;
|episodelink= is deprecated because when used in a citation with |url=, the rendered citation is broken;
This whole conversation is about what you and I think. Neither of us has convinced the other to think differently. I'm perfectly willing to continue as long as there is something new and relevant to discuss. I'm not interested in discussion about the discussion.
—Trappist the monk (talk) 13:19, 21 April 2013 (UTC)
Lack of documentation gives the appearance that |episodelink= and |serieslink=are disallowed, especially to editors newly using the template, as they won't realise the parameters have ever existed. The argument about |episodelink= being deprecated because of the problem caused when |url= is supplied is specious. If |title= is linked the same error occurs, and it doesn't fail gacefully. Since "Neither of us has convinced the other to think differently", there's no WP:CONSENSUS for your edits that removed the documentation. Per WP:STATUSQUO the removed documentation should remain in the documentation until we can come to some agreement. --AussieLegend (✉) 09:12, 23 April 2013 (UTC)
Leaving deprecated parameters in the template skeletons is contrary to the stated purpose of the deprecation so I removed them. It is intended that new and experienced editors move away from |episodelink= and |serieslink=. Removing the parameters from the template skeletons is a form of gentle dissuasion.
It is true that using |episodelink= with |url= in the same citation breaks the rendered citation. That truth is not invalidated by arguing that identically broken citations can be created by other methods.
I will accept leaving |episodelink= and |serieslink= in the template skeletons if you will accept the addition of <!-- deprecated --> annotations in both of the template skeletons and deprecated; wikilink title instead and deprecated; wikilink series instead in the Brief instructions / notes column of the vertical template skeleton.
—Trappist the monk (talk) 12:02, 23 April 2013 (UTC)
It's not really a case of you accepting "if". WP:STATUSQUO, which has wide support, says "If you make an edit which is good-faith reverted, do not simply reinstate your edit - leave the status quo up. If there is a dispute, the status quo reigns until a consensus is established to make a change." However, I don't see a problem with the above changes. --AussieLegend (✉) 13:57, 23 April 2013 (UTC)
Trappist the monk, you are the only one saying that the parameters are depreciated, a status that can only be determined by a group of editors who monitor the template. They are useful, and should be fully documented. 117Avenue (talk) 04:48, 24 April 2013 (UTC)
There is a marketing ploy that advertisers use to help get people to buy their stuff. Essentially it says to prospective purchasers, "All of your friends have our <insert cool product de jour>, look how great their lives are now." That doesn't work with me irl and it won't work here.
—Trappist the monk (talk) 12:06, 24 April 2013 (UTC)
Wikipedia is a collaborative project, and templates and help pages like these are used to produce "a consistent look throughout the encyclopedia." (From this page.) "Wikipedia does not have a single house style. Editors may choose any option they want; one article need not match what is done in other articles". However, I feel that other users should not be discouraged to use parameters that many others have found helpful. Listing a template's parameter in the documentation does not make its use compulsory, and I apologize if my comments make it sound like I said that. 117Avenue (talk) 22:36, 27 April 2013 (UTC)
Arbitrary section break - Fix for when both url and episodelink are supplied
I should probably add my comments, since my name has been mentioned. But this conversation has gone on long, so I didn't read it, and don't know who has said what. The citation should link to both an external website, and the episode article when both parameters are supplied. The external website should have prominence, because this is a cite template, and Wikipedia shouldn't be referencing itself. However, if an article exists on the episode, it is additional information that should be in the cite, like a link to the author, work, or isbn in other cite templates. 117Avenue (talk) 03:13, 14 April 2013 (UTC)
I hadn't actually realised the problem existed until I found your proposal (now archived here) during this discussion, but the resolution seems to make sense. Personally, I'd prefer to see both the season and episode number linked (ie Season 4, Episode 19, but what you've proposed is an acceptable compromise. The template has been modified since you made your proposal,[14] so the output is slightly different from what the current template provides but, using the example that I've been using, that output is:
I don't see why this change shouldn't be incorporated into the template. Can you update the sandbox using the current version of the template? --AussieLegend (✉) 12:29, 14 April 2013 (UTC)
This is what the change would look like. It should be noted that the template is currently broken, when both are supplied the open quote mark links to the URL, and the title links to the article. 117Avenue (talk) 06:30, 15 April 2013 (UTC)
That looks good, and fixes the problem. I've checked a number of articles and haven't seen any adverse effects that would be caused. Is there any opposition to this fix? --AussieLegend (✉) 11:25, 15 April 2013 (UTC)
I notice that ((cite episode/sandbox)) doesn't fail gracefully if |episodelink= is defined and when both |series= and |number= are missing or blank.
—Trappist the monk (talk) 12:41, 15 April 2013 (UTC)
I'm not sure that's a bad thing. Bells and flashing lights should go off when |season= (not series) and |number= are omitted. They're not really optional parameters, which is something the documentation fails to address. --AussieLegend (✉) 13:25, 16 April 2013 (UTC)
Yeah, I agree that errors should be reported in unambiguously. The Lua versions of the Citation Style 1 templates do that rather nicely. What happens here is that the template's output is readable; all of the season and episode information is there so the eye simply scans over the wiki markup. Because editors expect the citation to show the season and number and they see the season and number in much the same form as they saw it in the edit window, all is good. This is why authors have editors.
—Trappist the monk (talk) 13:42, 16 April 2013 (UTC)
I believe other cite templates err when there is an odd combination of parameters like this, I think it's acceptable. 117Avenue (talk) 05:14, 17 April 2013 (UTC)
Then please don't fix any more templates. Wikipedia needs excellence, not mediocrity.
—Trappist the monk (talk) 10:26, 17 April 2013 (UTC)
I don't think that tone is appropriate. Please be more WP:CIVIL. --AussieLegend (✉) 10:35, 17 April 2013 (UTC)
is excellence? 117Avenue (talk) 03:49, 19 April 2013 (UTC)
For those citations, the error messages are Editor AussieLegend'sbells and flashing lights [that] go off when [|title=] and [|archivedate=] are omitted. They embody Editor Gadget850's engineer, educate, enforce philosophy. The error messages are engineered to draw the editor's attention to the problem, to educate the editor (the help link) and enforce correct construction of citations using the templates.
So, yeah, excellence.
—Trappist the monk (talk) 10:13, 19 April 2013 (UTC)
While I agree with the outputs shown above, they're not up to my standards for bells and flashing lights. Unfortunately even the most obvious of errors often go unnoticed,[15] so actual bells and flashing lights, with an electric shock or two thrown in are what I'd like to see. However, I don't think 117Avenue's proposal is mediocrity. It's a reasonable proposal. I don't see anyone proposing anything better. --AussieLegend (✉) 15:41, 19 April 2013 (UTC)
It was probably some lawyer at WMF that made them disable ((editor chair zap)).
I'm far from an expert in the arcana of the templating language (if one can call it that). But it appears to me that the episode link section of the template starts producing its output before it knows what inputs are available. If it determined that it didn't have all of the requisite inputs, the output should be ((citation error)).
—Trappist the monk (talk) 18:36, 19 April 2013 (UTC)
I see two options for the output when these unique set of parameters are supplied. A. don't use the episode link, giving the external link prominence, as I suggested at the top of this section. Or B. display an error message using Template:Citation error, as Trappist the monk suggests, placing the article in a maintenance category. 117Avenue (talk) 03:31, 20 April 2013 (UTC)
I think "B" is a better option. It still gives |url= prominence but I think there may be a third option; C. make |episodelink= always tied to |season= / |seriesno= and |number=. If |url= is not specified, then the episode name will just be plain text. --AussieLegend (✉) 11:46, 21 April 2013 (UTC)
I don't think that will work. A user would supply a link to a source (episodelink) and a title, but if there is no episode number, nothing will be linked to. 117Avenue (talk) 23:10, 21 April 2013 (UTC)
Fair enough, "B" it is then. --AussieLegend (✉) 09:20, 23 April 2013 (UTC)
So, are we going to stall the discussion at this point, or are we going to agree to implement the fixes? --AussieLegend (✉) 10:45, 30 April 2013 (UTC)
I still got some more testcases to run it through, we want excellence right? 117Avenue (talk) 03:23, 2 May 2013 (UTC)
Yes we do and I can see that you have made significant progress in that direction. Well done.
Done, I enjoyed reading the conversation above. — Martin (MSGJ · talk) 18:47, 12 May 2013 (UTC)
"Googlebooks" parameter
Replacing ((Google books)), could some please add a googlebooks parameter to ((cite book)) (or whatever sub-template is affected), that creates a link to, well, Google Books? Similar to JSTOR and others, it should look like |googlebooks=8jHiEwVmB8MC and produce something like "Google Books8jHiEwVmB8MC", or just "Google Books". Thanks. --bender235 (talk) 09:55, 10 May 2013 (UTC)
I just insert the courtesy link to the item at Google Books into |url= along with the full citation information for the print edition that's really being cited. Imzadi 1979→ 16:41, 10 May 2013 (UTC)
Putting in the whole google url is unsatisfactory in several regards. It tends (by default) to disclose the country tld from which the editor is editing. Also it includes information about his browsing choices, etc. that is at best useless to other readers. The only useful things in the whole (often very long) url are the id and the page numbers. LeadSongDogcome howl! 17:19, 17 May 2013 (UTC)
Should make it as easy as possible to see refs when available. In fact we have a tool to help with the problems you describe above see - Google book tool Coverts bare url into ((cite book)) format . Having sources easily accessible helps with article expansion and increases the credibility of articles. Sending readers on a goose chase to find a book for no reason when there is a link available directly to the source is clearly not better. Moxy (talk) 17:34, 17 May 2013 (UTC)
Ok, I've banged out ((gbook)) to simple render the url given the id and the initial page number. This should be usable for the value of the url parameter within CS1 templates. LeadSongDogcome howl! 04:00, 18 May 2013 (UTC)
Unlink trans_title from URL to allow wikilinks or reftags
Proposal: The wp:CS1 cite templates should be changed back, as in the original Lua revisions, to not have external-linking of the "trans_title=" (plus the "title=") to the URL address, but instead, allow the trans_title to contain wikilinks, or footnotes explaining the translation.
Cite web comparison
Wikitext
((cite web|date=18 May 2013|sandbox=no|title=''Nicht gute Mode''|trans_title=Not good style|url=http://en.wikipedia.org/wiki/Hyperlink))
Sandbox2: ((cite web/new2|title=''Nicht gute Mode'' |trans_title=Not good style|date=18 May 2013|url=http://en.wikipedia.org/wiki/Hyperlink)) [inactivated 2016-01-06 because it causes a Lua script error —Trappist the monk (talk) 16:10, 6 January 2016 (UTC)
There is just no need to link both title and trans_title to the URL, when meanwhile any title can be extended to include extra words if a URL-link seems to be mysterious blue-linked text. Unlink "trans_title=" from "url=" and unlink "trans_chapter=" from the "chapterurl=" parameter. -Wikid77 16:58, 18 May 2013 (UTC)
These all look the same. --Gadget850talk 21:48, 18 May 2013 (UTC)
The position of the second double-quote is different: in the 'Old' row it follows the closing square bracket, whereas in the other two it follows the word 'Mode'. --Redrose64 (talk) 22:59, 18 May 2013 (UTC)
The change in the location of the double quote was intentional. The translated title (usually) isn't part of the original publication, but rather something added later, and style guides (such as the Chicago Manual of Style) generally seem to suggest that neither quotes nor italics should be applied to a translated title that was added after the fact. That said, I don't think Wikid77 is actually discussing the quote marks. I think he is trying to raise some other issue, though I'm not sure what. Dragons flight (talk) 23:50, 18 May 2013 (UTC)
The proposal is to unlink "trans_title=Not good style" as not part of the linked title text, as shown above in the Sandbox2 example. -Wikid77 (talk) 03:45, 19 May 2013 (UTC)
How to give author names in different scripts?
We need a parameter for reporting author name in different script (Chinese, Japanese, Korean, Cyrillic, etc.). This is the best I was able to do, and it is obviously not good enough. Giving author names in latin and originals script seems to me as important as translating the title. Ditto, btw, for publisher. --Piotr Konieczny aka Prokonsul Piotrus| reply here 03:57, 30 April 2013 (UTC)
The wp:CS1 cites allow any text in author, title or publisher parameters: Feel free to show other-language text, along with the related English text, in those parameters of the wp:CS1-style cites. However, the title data can be placed into dual parameters, as "title=" for original and "trans_title=" for the English equivalent (or also "chapter=" with "trans_chapter="), but the author names can be combined as in "author2=Jonéiè Doeyåviçnoыfu (John Doe)" with the option "ref=Doe1998" to set author-year links by any spelling of the author names. Some text can be offset within square brackets "[__]" but perhaps encode them as [ ("[") or ] ("]") to avoid temptations to double-link "[" or "]" in the text. Yet, there has been a design flaw where the trans_title is also linked to the URL, causing wikilinks to be rejected in the trans_title text, and so it should no longer be in the external link [http:...] to "url=" addresses. In general, people have tried to avoid too many new parameter names, which would tend to confuse new users, and also slows the template processing. In fact, the rapid markup Template:Cite_quick (which handles only 50-60 major cite parameters) sometimes runs slightly faster, and allows more templates (omitting COinS metadata), than some of the Lua-based wp:CS1 templates, when numerous parameters are passed into citations. The slowdown seems to be related to the MediaWiki parser having exponentially-slower processing to pass additional parameters, and templates can run faster with data combined into fewer parameters. It is a complex problem, where passing the 99th parameter, seems to require re-processing the first 98 parameters, or such, perhaps to check reuse of prior parameter names, as accumulated slowdown. So, a template with 4,000 parameters runs much slower (4x?) than 2 templates of 2,000 parameters each, such as selecting among lists of thousands of entries. Even changing a template of 200 parameters, into 2 of 100-parameters-each, can run 30% faster by calling 2 rather than 1 template! Anyway, feel free to put multiple-language data into the various current cite parameters. -Wikid77 16:48, 19 May 2013 (UTC)
I've seen some differences in the documentation between Help:Citation Style 1 and Template:Cite web, and hope people here can help clear them up. For example, Help:Citation Style 1#Work and publisher states "The 'publisher' parameter should be included even where it would be the same or mostly the same as the work/site/journal/etc., for example:
However, Template:Cite web#Publisher states that the publisher is "Not normally included for periodicals."
Could someone please make these consistent so everyone knows how to use these templates? Thanks! GoingBatty (talk) 02:25, 16 May 2013 (UTC)
I believe the documentation merely instructs you where to put 'The New York Times Company' if someone wants to put it, and not actually an invitation or inducement to do so. -- Ohc ¡digame!¿que pasa? 06:38, 16 May 2013 (UTC)
It does look to me like a recommendation to do so, and it certainly does need to be reworded. It says "The "publisher" parameter should be included even where it would be the same or mostly the same as the work/site/journal/etc.", and this is completely wrong in the case of newspapers. We specifically do NOT want the publisher included in the case of mainstream newspapers. This has been discussed more than once elsewhere. It is absurd to have a ref. saying that The New York Times is published by the New York Times Company, and in fact we never do so. -- Alarics (talk) 07:16, 16 May 2013 (UTC)
Yes, there are several inconsistencies. Discussions have always been concerned with what appears on the template pages, so that should be considered more authoritative, as those are updated after discussion. This help page needs to be reworked so that is keeps up with the template doc— that has been on my todo list for a while. --Gadget850talk 10:31, 16 May 2013 (UTC)
Hi. What does ((Cite web)) has to do with periodicals anyway? It is for quoting web, not periodicals. This sentence is probably a leftover of copy and paste, especially since I see this exact sentence in other DOC pages too. The statement of Help:Citation Style 1 sounds more logical because it follows the same reason for which User:GoingBatty has come here: Consistency. When I was working on taking Microsoft Security Essentials to WP:FA status, I was required to make the citations consistent. It was deemed unacceptable for one citation to have |publisher= and another not having any. Best regards, Codename Lisa (talk) 11:29, 17 May 2013 (UTC)
Therefore, the instructions at ((Cite journal)) and ((Cite news)) are the same as ((Cite web)): publisher is not normally included for periodicals". My goal is to have the documentation consistent, so we know how to make articles consistent. Thanks! GoingBatty (talk) 20:04, 17 May 2013 (UTC)
Hi. I'm really confused: Why make two templates when both are taken to equally accomplish one task and their differences criticized in the name of consistency? But in any case, I was required to include both |publisher= and |work= in WP:FA, so I will continue to do so. No offense but I do not follow a guideline that does not help me in FA. Best regards, Codename Lisa (talk) 17:51, 18 May 2013 (UTC)
Then the problem seems to be at FA, because they are supposed to follow guidelines. And they do seem to rather slavishly seek adherence to the WP:MOS, which we all know is not policy but a guideline. -- Ohc ¡digame!¿que pasa? 05:38, 19 May 2013 (UTC)
Hi. I am guessing "they" think "we"/"others" are the problem! (Seriously, there is no "they"; it is us all the way.)
While we are at it, the significance of metadata my not be obvious today because we think "Well, who doesn't know [Insert Journal Name Here]? Let's not include |publisher=" The catch is: Five, ten or fifty years from now, the situation may be completely different. Even now, |work= and |publisher= have helped me revive dead links while otherwise, it would have been impossible.
To conclude, I support "Help:Citation Style 1"'s current statement. It is the only way of achieving guaranteed inclusion of metadata, true consistency and FA status.
Best regards, 09:54, 19 May 2013 (UTC)
csdoc and maps
Recent changes to Module:Citation/CS1 caused me to review ((cite map)). There, in that template's documentation, |section= and |inset= are grouped with Identifiers. Use of |section= and |inset= is described part of In-source locations. Is this right? Shouldn't |section= and |inset= be grouped with In-source locations and not in Identifiers?
—Trappist the monk (talk) 22:59, 19 May 2013 (UTC)
Fixed Not sure why I did it that way. --Gadget850talk 23:11, 19 May 2013 (UTC)
Proper style for PDF page mismatch?
There are some documents in [bitsavers] that are scans of paper documents, where the PDF page numbers do not match the numbers printed on the page. In that situation, should ((cite manual|page=)) match the numbers from the PDF or the numbers on the actual pages? Or should I show both, e.g., IBM. "Virtual Access Methods". IBM System/360 Time Sharing System System Logic Summary Program Logic Manual(PDF). IBM. p. 56 (PDF 66). GY28-2009-2. The direct access volumes, on which TSS/360 virtual organization data sets are stored, have fixed-length, page size data blocks. No key field is required. The record overflow feature is utilized to allow data blocks to span tracks, as required. The entire volume, with the current exception of part of the first cylinder, which is used for identification, is formatted into page size blocks.((cite book)): Cite has empty unknown parameter: |sectionurl= (help); Unknown parameter |separator= ignored (help)Shmuel (Seymour J.) Metz Username:Chatul (talk) 19:40, 23 May 2013 (UTC)
In general: go with the pagination of the original version. However, it is often a great convenience to have the pdf page numbers. In such cases I add them in square brackets (using "[" and "]"). ~ J. Johnson (JJ) (talk) 20:40, 23 May 2013 (UTC)
This has been discussed before without much in the way of resolution. Unfortunately, I'm just on my way out the door and away from the internet for the next week so won't be able to participate in this discussion until I return. I'd still like to see the tracking categories renamed as I have suggested.
—Trappist the monk (talk) 16:36, 28 May 2013 (UTC)
Embedding a page-specific url within Cite_book when the book has its own article
As far as I can tell, for ((Cite book)) refs I have no choice as to whether it's the book's title or its page/pages specification that is used as link text when an url is supplied. Here's an example of something generated by ((Cite book)), based on a Google Books page-specific url:
The latter has two advantages: the page-specific URL is linked to the page number itself, and I am able to also add a wikilink to the article on the cited book. As far as I know, ((Cite book)) doesn't support this (and I think it should) 72.244.204.222 (talk) 01:08, 30 May 2013 (UTC)
If you use the |at= parameter instead of |page=, you can do this:
I've just tried those links. They don't take me to page 287, nor indeed to any part of the book's text. Instead, they go to a page about the book, including some reviews. This link therefore fails the requirements of the |url= parameter so it is hard to see why it is satisfactory for a particular page. --Redrose64 (talk) 07:26, 30 May 2013 (UTC)
No, it's just the same. There is the usual Google nav frame containing a scrollable page, the content of which (images omitted is:
Extended content
Widdershins
Front Cover
Charles de Lint
91 Reviews
Tom Doherty Associates, 12 Jun 2007 - Fiction - 560 pages
Jilly Coppercorn and Geordie Riddell. Since they were introduced in the first Newford story, "Timeskip," back in 1989, their friends and readers alike have been waiting for them to realize what everybody else already knows: that they belong together. But they've been more clueless about how they feel for each other than the characters in When Harry Met Sally. Now in Widdershins, a stand-alone novel of fairy courts set in shopping malls and the Bohemian street scene of Newford's Crowsea area, Jilly and Geordie's story is finally being told.
Before it's over, we'll find ourselves plunged into the rancorous and sometimes violent conflict between the magical North American "animal people" and the more newly-arrived fairy folk. We'll watch as Jilly is held captive in a sinister world based on her own worst memories--and Geordie, attempting to help, is sent someplace even worse. And we'll be captivated by the power of love and determination to redeem ancient hatreds and heal old magics gone sour.
To walk "widdershins" is to walk counterclockwise or backwards around something. It's a classic pathway into the fairy realm. It's also the way people often back slowly into the relationships that matter, the real ones that make for a life. In Widdershins Charles de Lint has delivered one of his most accessible and moving works of his career.
At the publisher's request, this title is being sold without Digital Rights Management software (DRM) applied.
« Less
What people are saying - Write a review
User ratings
5 stars
29
4 stars
27
3 stars
25
2 stars
10
1 star
0
I'm a sucker for a happy ending. - Goodreads
I don't like the writing style very much. - Goodreads
I like my prose buttery! - Goodreads
This crowdedness moves the book's plot. - Goodreads
The ending is a fairytale ending, of course. - Goodreads
Review: Widdershins (Newford #16)
User Review - Nanci - Goodreads
Charles de Lint has a gift for creating stories that blend the magical and mundane worlds. "Widdershins" is the tale of two of his best-known characters from his Newford series, Jilly and Geordie, and ... Read full review
Review: Widdershins (Newford #16)
User Review - Abby - Goodreads
Widdershins was the second book I read by Charles de Lint, and it was a good book, but I found myself disenchanted with the ending. (view spoiler)[I know that the stories are kind of fairy-tale-esque ... Read full review
All 91 reviews »
Related books
The Onion Girl
Charles de Lint
Spirits in the Wires
Charles de Lint
Forests of the Heart
Charles de Lint
The Ivory and the Horn
Charles de Lint
Moonlight & Vines
Charles de Lint
Someplace to Be Flying
Charles de Lint
Trader
Charles de Lint
Jack of Kinrowan
Charles de Lint
Tapping the Dream Tree
Charles de Lint
Muse and Reverie
Charles de Lint
Memory and Dream
Charles de Lint
From a Whisper to a Scream
Charles de Lint, Samuel M. Key
Other editions - View all
12 Jun 2007
No preview
16 May 2006
No preview
About the author (2007)
Charles de Lint and his wife, the artist MaryAnn Harris, live in Ottawa, Ontario, Canada. His evocative novels, including Moonheart, Forests of the Heart, and The Onion Girl, have earned him a devoted following and critical acclaim as a master of contemporary magical fiction in the manner of storytellers like John Crowley, Jonathan Carroll, Alice Hoffman, Ray Bradbury, and Isabel Allende.
Bibliographic information
QR code for Widdershins
Title Widdershins
Author Charles de Lint
Publisher Tom Doherty Associates, 2007
ISBN1429911344, 9781429911344
Length 560 pages
Subjects
Fiction
Windows XP Firefox 21. --Redrose64 (talk) 17:45, 30 May 2013 (UTC)
There is another reason not to use the technique described by GoingBatty: even though the |at= parameter has been used, it still ends up being described as a page number in the COinS metadata:
it's the bit from &rft.pages= to &rft.pub=, which since it contains Wiki markup, is therefore invalid (see WP:PAGELINKS). --Redrose64 (talk) 07:36, 30 May 2013 (UTC)
Wrapping long code example
A side issue, but is it possible to force line-wraps in a long code example, such as the one in the above section? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:04, 30 May 2013 (UTC)
My original objective was to have a way of linking a page-specific URL to the cite book page number text, so I could add a wikilink to the title text for the cited book:
That URL does not display page 287 of Widdershins. It used to, at least for me, but it doesn't anymore. Since Widdershins is under copyright, there are limits as to what Google Books lets us see. I think Widdershins was Limited Preview (see https://developers.google.com/books/docs/dynamic-links), and I thought pages of Limited Preview books could be reliably viewed again and again if the url included working values for ots, sig, sa, ei, etc. I thought I had gotten it to do so in the past, but I clearly didn't in this case.
Here's a working example of a page-specific URL for a different book:
The URLs all work for me on Windows 7 Firefox 21 as well - I can read the text on page 287. The Google book tool doesn't meet the requester's requirement to have the reference include a wikilink to a Wikipedia article and an external link to the page in question. GoingBatty (talk) 02:22, 31 May 2013 (UTC)
All they have to do is add the link to the page section and wikilink the title as mentioned above - takes a second. But the fact is the tool gives a stable URL.
((cite book|author=Charles de Lint|title=[[Widdershins]]|accessdate=30 May 2013|date=12 June 2007|publisher=Tom Doherty Associates|isbn=978-1-4299-1134-4|page=[http://books.google.com/books?id=Lze6oI8YnJ8C&pg=PA287 287]))
Putting any form of wikimarkup into the |page= parameter causes bad COinS metadata. --Redrose64 (talk) 10:13, 31 May 2013 (UTC)
This should be fixed - we have thousands and thousands of refs like this. Should be more concern with verifiability and accessibility then reference management software.Moxy (talk) 13:32, 1 June 2013 (UTC)
I agree with Moxy, we have thousands and thousands of page specific urls. Fix the software. --Bejnar (talk) 19:49, 1 June 2013 (UTC)
CoinS
I have three questions about COinS metadata:
Where is it documented that CS1 templates emit the page parameter as metadata?
Where are the specifications about what form a page parameter would have to take for the metadata to be valid?
What applications make use of the metadata, that is, why should we care?
I use Zotero, which in turn uses COinS, to extract reference details to a local library, which I can then drag into other articles. The links in this comment will tell you more. Generally speaking, marking up our content in a semantic, machine-readable manner like this helps us in our mission to disseminate knowledge as widely as possible. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:16, 31 May 2013 (UTC)
Use 'chapter' and 'chapter-url' here. There is no guarantee that different versions of these templates use the exact same markup. --Gadget850talk 07:29, 23 May 2013 (UTC)
Wikibooks is a different project. Some templates do exist on more than one project; often they are copied from one project to another. But although the copy might be identical to start with, there is no formal link so any changes made in one are not reflected back to the other.
The |section= and |sectionurl= parameters were added with this edit, and the template was copied from Wikipedia to Wikibooks on 28 April 2010 after which the two templates took different paths. The |section= and |sectionurl= parameters continued to be valid on Wikipedia after ((cite manual)) was merged into ((cite book)) with this edit and this one. The next edit to ((cite book)) was the Lua migration, so it does seem that these parameters were lost at or after that point.
I'm not willing to track down what happened in Module:Citation/CS1 because the module editor is nasty; moreover, whilst Ctrl+F works to find text when viewing the page (which is supposedly the module's source), it won't find the same text when you go for "edit". --Redrose64 (talk) 10:07, 23 May 2013 (UTC)
I was just hunting this down. These parameters were supported by ((cite manual)) and were ported to ((cite book)) before the merge. Looks like they got lost in the Lua update. But, Module:Citation/CS1/Configuration is all the aliases, and easy to update. This should now be fixed. --Gadget850talk 10:32, 23 May 2013 (UTC)
Is it? I think that you've correctly made CS1 act like the old ((cite manual)). That doesn't answer what I think Editor Shmuel (Seymour J.) Metz Username:Chatul is asking for. If I read it correctly, the request is for the ability to link to a section, a chapter, and the title all in one citation. Because |section= is an alias of |chapter=, this request is not answered. That is something that should, I think, be added to Module talk:Citation/CS1/Feature requests.
—Trappist the monk (talk) 12:12, 23 May 2013 (UTC)
I don't know the difference between 'section' and 'chapter'. ((Cite manual))/section and ((cite book))/chapter both fed into the same meta-parameter, thus had the same functionality. ((Cite manual)) never supported both parameters separately, nor does the Wikibooks version. --Gadget850talk 12:43, 23 May 2013 (UTC)
My initial concern was that previously valid ((cite manual)) tags were being flagged as errors. However, it would certainly be useful if chapter and section could be used in the same citation, with associated url specifications. Is there a case for something more general, e.g., section-n, sectiontype-n and sectionurl-n, where the types might be, e.g., Part, Chapter, section, subsection? Shmuel (Seymour J.) Metz Username:Chatul (talk) 19:01, 23 May 2013 (UTC)
I would have to see an example where multiple sub-chapters/sections and links would be useful. --Gadget850talk 19:38, 23 May 2013 (UTC)
My first example only needs two levels. My second example needs four, but since it is for a PDF it can't use the URL links.
Chatul's desire seems to be to state the title of a manual, the name of a chapter in the manual, and the name of the section within the chapter, and have a separate URL which links to each level in the hierarchy. This is not what one usually sees. Usually a URL is only provided to the lowest level in the hierarchy for which a specific URL is available.
A variation of this case is when a work is cited several times in an article, each cite being to a different part of the work. In this case there could be a reference in the bibliography, and then a short citation to the specific part in the "Notes" section. A common way to do this is with the ((sfn)) template, which provides a link to the bibliography entry. However, it could be done manually and the link in the notes could lead to the appropriate section of the work. Unfortunately, it would be cumbersome for the short citation to provide links to both the bibliography entry and the section of the work itself. Jc3s5h (talk) 10:47, 4 June 2013 (UTC)
Done. Actually, this needed to be added to Template:Cite book/doc, which is unprotected, so you could have done it yourself. I have made the edit for you, however. Best — Mr. Stradivarius (have a chat) 05:07, 10 November 2012 (UTC)
Fixed Added 
Done Added as an alias to series and version. --— Gadget850 (Ed) talk 19:25, 4 December 2012 (UTC)
Not done: Sorry for the delay in answering this. I can see how this might be a good idea, but I think it would be better to fix this problem for all similar TfD nominations, not just this one. Maybe you could suggest this at WT:TFD, and if there is support for the idea we could update ((Template for discussion)) and/or ((Template for discussion/dated)) to allow renames to be specified by the proper template. Let me know if you need any help with the coding. Best — Mr. Stradivarius (have a chat) 14:06, 20 December 2012 (UTC)
Fixed --— Gadget850 (Ed) talk 19:35, 29 March 2013 (UTC)
(Seriously, there is no "they"; it is us all the way.)