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. |
Archive 60 | ← | Archive 62 | Archive 63 | Archive 64 | Archive 65 | Archive 66 | → | Archive 70 |
I am posting here as I have just finished an attempt to mediate between two users who were falling out over their strong and opposed positions on whether (some, or all) standalone years should be linked.
I thought the edited highlights of the discussion we had there might be interesting to readers of this page, and so I append them below. I'd like to think that, especially if we are eventually able to remove the necessity to link dates for them to follow users' date preferences, this could allow us to reduce the amount of overlinking currently present in many articles, clarify the current "wooly" MoS guideline, and thus avoid future friction over this.
Specifically, I propose Quadell's remedy (see table below) as a new guideline for linking years. I'd also suggest something along the lines of "If in doubt, don't link" as many, many of these year articles do not provide any useful context and, it seems to me, never can. Anyway, please read it and see what you think. The original text this was condensed from is at User:Guinnog/date linking and User talk:Guinnog/date linking.
--Guinnog 03:42, 12 January 2007 (UTC)
(copied text begins) This page is an attempt to resolve the dispute about the linking of year articles. Perhaps it could serve as a step towards a more general agreement though. Once this dispute is resolved I will paste it into Wikipedia talk:Manual of Style (dates and numbers).
At present we have got:
"Partial dates
If the date does not contain both a month and a day, date preferences do not apply: linking or not linking the date will make no difference to the text that the reader sees. So when considering whether such a date should be linked or not, editors should take into account the usual considerations about links, including the recommendations of Wikipedia:Only make links that are relevant to the context. ...
... There is less agreement about links to years. Some editors believe that links to years are generally useful to establish context for the article. Others believe that links to years are rarely useful to the reader and reduce the readability of the text. Another possibility is to link to a more specific article about that year, for example [[2006 in sports|2006]], although some people find this unintuitive because the link leads to an unexpected destination."
(Wikipedia:Manual of Style (dates and numbers)).
While I think I understand why this compromise version was adopted, I also think the ambiguity leads to some unfortunate friction between users who think all year links, or almost all, should be delinked, and those who regard this as a loss of utility in articles.
I tend to side more with the delinkers than with the linkers. I find that the majority, even a vast majority of year links are adding little or nothing. I can see the merit of the other side's opinion too, and I promise to try my best to mediate here without regard to my own POV. For interest, here is a recent copyedit I made which includes date delinking: [1]. Guinnog 04:23, 3 November 2006 (UTC)
Never link / Can be delinked on sight | Seldom link / Usually delink | Usually link / Sometimes delink | Always link / Never delink | User | Comment |
---|---|---|---|---|---|
Multiple repetitions of the same linked year within a page, especially in lists. Links where the article linked to would in my judgement clearly add no context to the subject of the article. | Recent links, say roughly from 1900 - 2006. "Easter egg links" like [[1933 in aviation|1933]]. Rather use: (See also 1933 in aviation) | 1800 - 1899 | dates prior to 1800 | Guinnog (before) | These are just my own interpretations; I stress I always discuss link removal on the odd occasion someone challenges it. I obviously try to judge each case on its merits. |
Standalone year or year-month links where, in my judgement, they add no context. I rarely see any such value, just confusion as the reader is confronted with a sea of blue years: what is of value to them? What not? | N/A | N/A | Year links in century articles | Hmains | As far as I know, this is the editing I am supposed to be doing as a WP editor. However, I also do not like the loose guidelines and the trouble they cause and said so when they were being discussed; however, in the interest of 'consensus', the guidelines, as written, were accepted. |
I think this is inherently a judgement call. If someone thinks it is useful, keep it. If someone doesn't, remove it. Just don't get in the habit of spending all ones time doing either one or the other. | Generally, I don't mind if people cut recent links (say, last twenty years). I disagree with Guinnog about setting a cutoff about 1900 - most of the most useful links, I would argue, are those for historical articles during the 20th century, where providing a broader geopolitical context is often really quite relevant. | Again, I think this is inherently a judgement call. There are many cases where links in the time periods both Guinnog and I mention where date links may be relevant, but there also many where they may not. I don't think setting specified dates is particularly helpful, though I do think there's much less likely to be an issue with removing date links in the last couple of decades. | As above. | Rebecca | I think this really is a matter of both having respect for other points of view and using a bit of discretion. Setting hard and fast rules inevitably leaves useless date links in and leads to the removal of perfectly useful ones. |
Multiple repetitions of the same linked year within a page. | Any year link, except in special circumstances. "Easter egg links" like [[1933 in aviation|1933]]. Rather use: (See also 1933 in aviation) | Century links (most cases) | Year and/or century links that are critical to the understanding of the article | Guinnog (after) | I have learned a lot from this study and thought it would be interesting to summarise that in this table. |
Multiple repetitions of the same linked year within a page. "Easter egg links". | Years mentioned in an article, where the event is important, but not the year itself. E.g., the year an album was released in an article not about the album. | Any year links important enough to the article that it would be appropriate to use them in the main (summary) section, e.g. the founding year for an empire in that empire's article, or the year a swimmer won the Olympics in an article about the swimmer. | Birth and death years, or years in century articles. | Quadell |
(in order of appearance in the article)
*1st century (second instance, but well down a long article from the first)
*408 (second instance)
*12th century (second instance, in image caption)
*9th century (second instance)
*13th century (second instance)
*13th century (third instance)
Century or year | Grade for article quality (1=high, 5= low) (=x) |
Grade for relevance to Flat Earth (=y) |
Grade for utility (=x*y) |
Comment |
---|---|---|---|---|
6th century BC | 3 | 2 | 6 | Decent article, provides some historical context to when Pythagoras lived |
1st century BC | 4 | 4 | 16 | Little here |
1st century | 4 | 4 | 16 | Little here |
2nd century | 4 | 4 | 16 | Little here |
3rd century | 4 | 4 | 16 | Little here |
4th century | 4 | 4 | 16 | Little here |
5th century | 4 | 4 | 16 | Very thin indeed. Nice map. |
7th century | 4 | 4 | 16 | Minimal article, minimal interest |
8th century | 4 | 4 | 16 | |
9th century | 3 | 3 | 9 | Better article |
11th century | 3 | 3 | 9 | |
12th century | 4 | 4 | 16 | Back to an article which is a formless list; hard to see any relevance |
13th century | 4 | 4 | 16 | |
15th century | 4 | 4 | 16 | |
16th century | 3 | 4 | 12 | Longer article but still a sequence of lists. |
19th century | 2 | 3 | 6 | Very interesting article; lots of peripheral historical articles to browse to, if you were finished reading the original article. |
240 BC | 5 | 3 | 15 | Stub article |
245 | 5 | 4 | 20 | Even less, and less of interest |
315 | 5 | 3 | 15 | |
325 | 5 | 4 | 20 | |
329 | 5 | 4 | 20 | |
344 | 5 | 5 | 25 | Nothing here |
354 | 5 | 4 | 20 | |
379 | 5 | 5 | 25 | |
386 | 5 | 4 | 20 | |
394 | 5 | 3 | 15 | |
408 | 5 | 4 | 20 | |
430 | 5 | 4 | 20 | |
480 | 5 | 5 | 25 | |
524 | 5 | 5 | 25 | |
547 | 5 | 5 | 25 | |
560 | 4 | 5 | 20 | |
636 | 5 | 5 | 25 | |
672 | 5 | 4 | 20 | |
700 | 4 | 4 | 16 | |
735 | 5 | 4 | 20 | Year identifies the death of Bede, but article is so poor it provides only two other links, plus the turgid List of state leaders in 735 as context. |
784 | 5 | 5 | 25 | |
1013 | 4 | 4 | 16 | |
1054 | 4 | 4 | 16 | |
1070 | 4 | 4 | 16 | |
1120 | 4 | 4 | 16 | |
1225 | 4 | 5 | 20 | |
1274 | 3 | 3 | 9 | |
1400 | 4 | 4 | 16 | |
1550 | 3 | 3 | 9 | |
1816 | 2 | 2 | 4 | |
1828 | 3 | 3 | 9 | |
1830 | 2 | 3 | 6 | |
1885 | 2 | 3 | 6 | |
1888 | 2 | 3 | 6 | |
1896 | 2 | 3 | 6 | |
1898 | 2 | 2 | 4 |
I will have more to say on this. For now, let me explain what I have just done. I tried to put myself into the role of a reasonably intelligent but non-expert person reading the article on Flat Earth (the subject of the test piece). If we can take as read that the repeated, easter egg and very recent year links are regarded by both editors (and by me) as low-value links, it is interesting to click on each of the links that Hmains would have deleted and Rebecca would not, with a view to finding what information if any related to, or even gave meaningful context to, the events described in the article.
Some rather surprising (at least to me) things came up. The century articles seem more valuable for providing historical context than the individual year links for the earlier periods. Many of these early individual year links are so poor in my opinion as not to be worth linking to at all at present. It is also rather hard for me to see how they could ever plausibly be significantly improved. Look at 735 for example.
If you review the links in question, it is hard to avoid the conclusion that many of the year links added nothing at all in terms of direct relevance to the article. I do recognise there are some Wikipedia users who will value the semirandom possibility in, say, clicking on 1885 and thus navigating from Flat Earth to LaMarcus Adna Thompson, the roller-coaster pioneer. I had not been aware though that many of the early year articles are so poor as not to give any real possibility for this activity. I'll think about it some more and try to take this onward this evening. For now, anybody else want to comment? --Guinnog 13:02, 6 November 2006 (UTC)
I thought this might be a good place to consider what we want from this, beyond a resolution of the immediate dispute between these two respected editors. It has become obvious that the existing policy is flawed by its well-meaning ambiguity. I'd like to propose that we aim to come up with a list of examples along the model of the table I made up; something like Always link, Usually link, Seldom link and Never link. I would stress that the criteria I used myself, which largely depended on the period of the date being linked to, were only my own ones, and need not be seen as a model for whatever we will end up with. I hope that the exercise I have asked Hmains and Rebecca to do may allow us to discuss towards establishing such principles. Any comments would be most welcome. --Guinnog 13:25, 5 November 2006 (UTC)
1. I came at this as a modest date delinker ("I find that the majority, even a vast majority of year links are adding little or nothing") but with the idea that "dates prior to 1800" should be linked or left linked, while more modern ones were superfluous.
a) I could see some merit in what Hmains was doing; systematically removing all partial date links in articles he edited, per his understanding of what the MoS allowed. I could also see the annoyance of Rebecca at what she saw as mechanistic edits, against her understanding of what the MoS allowed.
b} It seems obvious that a lot of the problem lies with the ambiguity of the present MoS; I know that has been controversial but the compromise we have ended up with is a fudge that leaves us open to the kind of misunderstanding that happened here.
2a) Overall, I think the results favour the idea that I came into this with, that most year links are pretty worthless. However, I will now look differently on which year and century articles are and aren't worth linking. The 19th century one is rather good for instance. I know that many of the 20th century year articles (that we ruled out in this study) are fairly informative.
b) I know now that many of the year articles from early years are so poor as not to be worth linking to, ever. Furthermore I see no realistic way they can ever be improved. I saw no article for a year before 1000 that was more than stub-quality. I will certainly be much less inclined to leave links to articles like 735 in place than I was before doing this. Even many of the more recent ones are poor in terms of adding understanding. You have to get into the 19th century before the articles get quite informative and well-written. Even then, few of them added real core meaning; it was more of a serendipitous sort of "what else happened that year" ability that became more real as the articles got (modestly) better through the years. (This is the opposite of what I believed before this exercise).
c) The century articles are better than the years. Even so, some quite late ones are remarkably poor (15th century for example). This could be an opportunity to improve these articles; a much more realistic aspiration than fixing up the early year articles, in my view. Meantime, I would still only link to century articles in exceptional cases where it would provide a useful perspective. I would change my view somewhat if the century articles were improved.
3. While there might seem to be nothing wrong with Hmains' edits in removing these (very often worthless) links, I understand the annoyance of editors like Rebecca at what they see as an unauthorised bot-like bulk edit.
a) Although I have no more authority than anyone else to tell Hmains and Rebecca how to act, having thought about it a great deal, I suggest the following compromise might offer a way forward:
b) Hmains should refrain from using automated or semiautomated methods for unlinking dates (pending the clarification to the MoS I propose below). All linking or delinking done should be justifiable; for each and every link added or removed a rationale should be possible and be provided on demand, and the rationale should take into account the value of the date article linked or delinked to the main article being edited.
c) Rebecca should refrain from mass-reverting Hmains' edits; if she perceives a problem with an individual edit, she should merely ask Hmains nicely and he should provide his rationale, as noted above.
d) Neither party should revert-war or be anything but civil with each other. Any problems should be referred to me in the first instance.
4. Finally, if there is no objection, an edited version of this should be copied to the relevant MoS talk page, in case it can help progress formulation of a better policy. --Guinnog 08:02, 8 December 2006 (UTC) (copied text ends)
I propose where we currently have: "There is consensus among editors that bare month and day names should not be linked unless there is a specific reason that the link will help the reader to understand the article. There is less agreement about links to years. Some editors believe that links to years are generally useful to establish context for the article. Others believe that links to years are rarely useful to the reader and reduce the readability of the text."
We substitute:
"Bare month and day names should not be linked unless there is a specific reason that the link will help the reader to understand the article. Links to years on their own are often deprecated as adding little value to articles, but consideration should be given to including year links which add value to the article. If the year link adds little, a century link may be more useful. (Years in century articles, by convention, are always linked.)
In deciding whether or not to include any link, the principle of Wikipedia:Only make links that are relevant to the context should be followed, and the value, quality and relevance of an individual year article being linked to should be considered. The use of linking merely to emphasise important dates should be avoided.
Mass edits which solely remove (or add) date links across many articles are discouraged, as with any mass minor edits (see for example 'Avoid making insignificant minor edits ... because it wastes resources and clogs up watch lists.'(Wikipedia:AutoWikiBrowser)), and especially because different editors will make different judgements about the value of different links.
Easter-egg links of the type [[2006 in sports|2006]] are deprecated. Instead use a transparent form like (see also 2006 in sports), where that article has relevant information to the article."
--Guinnog 03:59, 18 January 2007 (UTC)
I really dislike reading articles with a sea of blue dates. It's almost never useful to click on the date number and the blue links are distracting. However, I can see a value in such links from a point of view of information content. It is very useful indeed to be able to go to a page about a particular year and find all of the pages that link to it in order to automatically find all of the things that happened in that year. So what is truly needed is not links from years but categories for each century, for each year and perhaps even each month and day within each year. Some people have already started doing this - I see Category:1959 introductions for new products released in 1959. So if we are considering changing the standard, I think we'd be better off having some kind of year template that spits out the year (unlinked) into the text of the article - yet adds a category reference for that year.
As for the 'preferences' thing of rendering dates in alternative styles - that leaves me cold. Linking dates as a way to make the MediaWiki recognise them as such is an ugly mechanism. If we must do that, there should be some other kind of markup that doesn't result in an ugly link. However, providing the month name is spelled out in letters (ie not as a number), there is really no ambiguity and readers who may prefer "10th January 2007" are unlikely to be confused by "Jan 10th 2007". On the other hand, we must strongly deprecate using month numbers because "01/10/07" is utterly ambiguous. To an American, it means "Jan 10th 2007", to a Brit, it mean "August 1st 2007" and to a Dutch reader, it would mean "2001 August 7th". If we would only require the use of month names ("Aug" or "August", not "10"), I think we could simply dump the user-preference date presentation. After wall, we don't have user-preferences for regional spelling ("tyre" vs "tire", "colour" vs "color", etc) - or word choice ("lorry" instead of "(big) truck", "ute" versus "(pickup) truck") - and that can be much more disturbing for the reader! We aren't trying to fix "rubber" (meaning "condom" to Americans and "pencil eraser" to a Brit) or "Durex" (being a brand of condom to a Brit or a brand of adhesive tape to an Assie). Compared to the potential for confusion there, the date issue is very, very minor indeed! SteveBaker 14:43, 12 January 2007 (UTC)
I think if we can reduce conflation of linking and formatting, per the proposal above, it would be good, because then we could discuss when dates actually ought to be linked. Based on the analysis, and on my own gut feeling, my view is most of the time they ought not to be linked. I think a case needs to be explicitly makeable to link them, and that by default they ought not to be. I'd support a change to the MoS that moved in that direction. How did I arrive at that conclusion? At least in part thanks to Guinnog's thorough analysis, above. ++Lar: t/c 05:33, 13 January 2007 (UTC)
I created a stub article today and for some odd reason, I linked the Month/Date and then the year...which is something I rarely have done in the past. I look at the stub now and see little improvement by linking to the dates actions mentioned in the article may have occurred. The cite templates also do this, linking to year/month/date and I can't see what the purpose of this is either. Articles that are discussed in the year articles or date articles might be benefited by themselves linking to those dates, but I can't see how any improvement is gained by having this be part of MoS.--MONGO 05:42, 13 January 2007 (UTC)
Guinnog, I see you keep refining the proposal, but I still haven't seen any sign that it's going to gain any consensus. Stephen Turner (Talk) 11:05, 15 January 2007 (UTC)
Guys. This is a perennial problem, in my view, and it may need more worrying at to resolve neatly. Last year's compromise may have helped, but that doesn't mean that trying to get a proposal that is crisper and that does have consensus isn't worth pursueing. I do not in any way see Guinnog as being arrogant! Rather I see him trying to take on an rather difficult and contentious topic at considerable personal cost, for the betterment of the wiki, and if the consensus is not in favour of moving in a particular direction, so be it. Terms like "slam your opinion through regardless", and "Rebecca's clawing and scratching any weight" might not be quite as helpful as they could be, don't you think? Let's work together collegially if we can. ++Lar: t/c 13:53, 15 January 2007 (UTC)
My view is that the elimination of YEAR in X|YEAR links would be a mistake. If you liken Wikipedia to a print encyclopedia, the YEAR articles are like the yearbooks--here's everything that happened in 1982. Referring someone to the entire book is not particularly helpful--much more helpful would be pointing them to a particular chapter of the yearbook where they are most likely to find the context they are looking for. If the user wants a broader context, it will always be easy to find 1982 from 1982 in table tennis, but the reverse is often not true.
The suggestion sometimes made that YEAR in X links should be replaced with parenthetical (See YEAR in X) statements strikes me as contrary to the spirit of Wikipedia, where links are supposed to be unobstrusive and expendable. Nareek 19:05, 15 January 2007 (UTC)
Guinnog said:
Easter-egg links of the type [[2006 in sports|2006]] are deprecated.<!-- The last sentence is also at WP:PIPE, so if it's changed here it should be changed there too -->
I don't see that at WP:PIPE; instead, I find:
Am I looking in the wrong place? Nareek 07:26, 18 January 2007 (UTC)