Archive 45 ← Archive 50 Archive 51 Archive 52 Archive 53 Archive 54 Archive 55

RFC on publisher and location in cite journal

The following discussion is an archived record of a request for comment. Please do not modify it. No further edits should be made to this discussion. A summary of the conclusions reached follows.
This RfC sought to answer 3 questions, each of which is addressed below. The short answer is that there is consensus that Citation bot should stop automatically removing the parameters.

1. Should Citation bot continue to remove these data?

There is consensus against automatically removing the parameters (|location= and |publisher=). However, assuming arguendo that I am mistaken, I am still quite certain that there is not a consensus in favor of automatically removing these parameters. Since the bot policy requires that a bot only performs only tasks for which there is consensus, a no-consensus close would also result in stopping the automatic removal. Accordingly, @Smith609 is requested to disable the bot until the code is updated and compliant with the results of this RfC. Since they have not edited in almost 2 weeks, I will leave a note at WP:ANI asking that the bot be blocked until it is updated to be compliant with the result of this RfC.

2. Is there consensus for this removal, anywhere, by anyone?

This question is a bit trickier. The discussion relating to question 1 discussed the benefits and drawbacks of having the parameters included, and the consensus was that the parameters should not be automatically removed. Thus, there is consensus against the indiscriminate removal of these parameters in the manner of a WP:MEATBOT. However, there is consensus that these parameters can be removed in the same way as any other edit is made (respecting the BRD process, discussing on the talk page, etc.); likewise, they can also be included in the same manner.

3. Should we continue to support these parameters in Module:Citation/CS1 for ((cite journal))?

There was less discussion about this question. But, given that there is consensus against indiscriminately removing these parameters (see questions #1 and #2), it would make no sense to remove the support for the parameters. Accordingly, we should continue to support the parameters for ((cite journal)).
This was a complicated topic; if anyone has any questions about this close feel free to ask me. I will leave notes on the maintainer's talk page, as well as at WP:ANI and WP:BOTN, regarding this close. Thanks, --DannyS712 (talk) 19:50, 22 February 2019 (UTC) (non-admin closure)

Right now, Citation bot removes publisher and location in ((cite journal)). Proponents of this removal state that this is recommended by nearly every style guide under the belief that there is little value to the information.

However:

This RFC seeks consensus for the following questions:

  1. Should Citation bot continue to remove these data?
  2. Is there consensus for this removal, anywhere, by anyone?
  3. Should we continue to support these parameters in Module:Citation/CS1 for ((cite journal))? If we continue to support these parameters, should we:
    1. Support them only in metadata without display;
    2. Display them only without support in the metadata;
    3. Status quo, which is both to support them in the metadata and display them?

RfC added to Template:Centralized discussion and relisted by Cunard (talk) at 07:48, 7 February 2019 (UTC) after an RfC close was undone. --Izno (talk) 15:59, 27 November 2018 (UTC)

For the record, ((cite journal)) does not emit metadata for |location= and |publisher= because COinS does not support those parameters for journals. For more, see these:
Matrix defining the KEV Format to represent a journal publication
Matrix defining the KEV Format to represent a book
Matrix defining the KEV Format to represent a patent
Matrix defining the KEV Format to represent a dissertation
With respect to the metadata, all cs1 periodical templates are treated as journal templates; all cs2 templates that use a periodical parameter (|journal=, |magazine=, |work= etc) are treated same as their cs1 counterparts.
—Trappist the monk (talk) 16:34, 27 November 2018 (UTC)
Thanks. Amended the RFC. :) --Izno (talk) 16:37, 27 November 2018 (UTC)
Second, conference proceedings are typically bound like a book, but they usually come out every year, so may be thought of as a periodical, and hence, a journal. An editor might be unaware of ((cite conference)) and cite a paper in conference proceedings with ((cite journal)), but include the place and publisher because it is like a book. Just deleting the place and publisher is the wrong way to correct such a situation; the way to correct it is to change from the journal template to the conference template, and leave all the parameters alone. Jc3s5h (talk) 16:52, 27 November 2018 (UTC)
How does one trivially prevent another editor from running a bot, aside from just blocking the bot entirely? ♦ J. Johnson (JJ) (talk) 23:12, 28 November 2018 (UTC)
I was talking about the documentation update, but if you want to know about how to prevent citation bot from touching a citation, simply put a comment on the problematic citation (e.g. ((cite journal<!-- Bypass citation bot -->|...))), or simply add ((Bots|deny=Citation bot)) to the article to tell it to leave everything alone. Headbomb {t · c · p · b} 14:49, 30 November 2018 (UTC)
The problem with "bots deny" is that one might want to be more selective than all or nothing servicing. And I've seen a bot-driver remove the bots-deny on the grounds that "all" is better than "nothing". On the otherhand, is this "bypass" comment simply a note to the bot-driver? Or is it bot detectable, possibly with specific requirements? ♦ J. Johnson (JJ) (talk) 21:44, 30 November 2018 (UTC)
@J. Johnson: As I understand it, the comments just make the bot skip over it by interrupting the string it searches for in the code (e.g., strings like ((cite journal|, presumably). It's still annoying that someone who might not even be aware that some future editor might go ahead and mess up all the citations by running a bot should be expected to go and comment out exceptions in all the fields which Citation Bot makes worse though. Umimmak (talk) 22:04, 30 November 2018 (UTC)
Curious. An undocumented feature? Well, thanks for that info. ♦ J. Johnson (JJ) (talk) 23:39, 30 November 2018 (UTC)
Izno waves his hands around a bit alleging various points (such as "Several recent threads at User talk:Citation bot ..." and "the consensus in Module:Citation/CS1"), but his links don't point to any specific language supporting his alleged points. Perhaps there is something there (like, one comment at Citation bot), but waving one's hands around isn't the same as describing an actual problem. That is a poor basis for trying to generate consensus for multiple questions on a matter of deep significance. ♦ J. Johnson (JJ) (talk) 00:15, 28 November 2018 (UTC)
Comment I'm not sure which discussion Izno has in mind, but a quick search produces the following relevant discussions on User talk:Citation bot: Publisher (Nov 2018); Bug: Publisher weirdness (Oct 2018); Publishers being deleted & specific pages being changed to page ranges... (Oct 2018); I disagree with the Consensus the drives the bot's actions (Sept 2018); Do not remove the publisher (Jul 2018). Umimmak (talk) 01:10, 28 November 2018 (UTC)
^ --Izno (talk) 03:37, 28 November 2018 (UTC)
Nor am I sure which discussion Izno has in mind, lacking any definite statement of his argument. ♦ J. Johnson (JJ) (talk) 21:42, 30 November 2018 (UTC)
Those same discussions show that the users asking for the publisher parameter to be filled could not agree on which name to put in it, for instance in the case of what appears to be a society journal. Those who want publisher names in citations should first come up with a system to choose the name and settle disputes on it. (Do we have exact records for who was the registered/legal publisher of every journal dating back to centuries ago? See also Umimmak below.) Nemo 07:02, 28 November 2018 (UTC)
@Nemo bis: That's a valid argument on an individual article or at WP:CITEVAR. But in terms of enforcing it across articles by way of rules in an automated tool, CITEVAR exists because the community as a whole has decided that such issues need to be settled at each article rather than centrally imposed. As it happens I would prefer there be One True Citation Style for enwp—and would make a strong argument for what The One should be and in favour of strict enforcement—but CITEVAR has stood steady for ages and I see no reason to think the community-wide consensus on this has changed since the last time.
PS. And, yes, for most journals, at least in my field, we know exactly who the publisher of record is over time, and, where relevant, who the actual publisher is (other fields may differ). In some cases the publisher is relevant information (location less so, but sometimes relevant), and some times it's not. That this is difficult to determine in some cases, or that editors agree on the specifics in others, does not really bear on the general issue (there are cases where author is unknown or editors disagree on authorship; it's a problem in the specific case, but doesn't really impact the general case). --Xover (talk) 08:00, 28 November 2018 (UTC)
WP:CITEVAR is about styles for otherwise equivalent information. Just read it: «Editors should not attempt to change an article's established citation style merely on the grounds of personal preference». It's not about giving total freedom to anyone to add whatever they want to citations unless someone objects on each talk page. Otherwise tomorrow someone can start adding fax, phone number, street address, ZIP code and their library's preferred unique identifier in the "publisher" field on the grounds that it's what they find most useful for the specific citation. Nemo 08:30, 28 November 2018 (UTC)
Actually, yes, that's exactly what CITEVAR provides for. It's just that such extreme examples are very unlikely to be actually used by anyone, and it's exceedingly unlikely that such use would find consensus on that article's talk page. There are also technical limitations when using citation templates: for example, |url= must actually contain a valid URL (but these do not apply when not using citation templates). But whether or not to include publisher and location information for journal cites is absolutely within the scope of what CITEVAR addresses. To wit: the arguments for mass removing these refer to external style guides. (Note that there are several good arguments for why this information should not be added in a lot, or even the majority, of cases—some of which have been brought up here—but per CITEVAR these must be decided on an article-by-article basis). --Xover (talk) 10:52, 28 November 2018 (UTC)
And if you want to keep the publisher/location because you have a special snowflake citation, you can do that by 1) not activating the bot 2) inserting a comment in the citation template. However, I've yet to see a valid, on-Wikipedia case, of where publisher/location acts an actual disambiguator, rather than something that was just added automatically by tools, or added by users mistakenly thinking "if there's a parameter, the parameter must be used". But if the publication is somehow ambiguous, and there's a need for disambiguation (e.g. you don't have a DOI), then using the ISSN should be teh go-to solution, rather than figure out what corporate entity was publishing the journal at the time of publication, because that may very well have changed 3-4 times since the article was published. Headbomb {t · c · p · b} 11:53, 28 November 2018 (UTC)
As a side note: that is a very interesting project. ♦ J. Johnson (JJ) (talk) 20:31, 2 December 2018 (UTC)
  • @DGG:, what about locations for modern journals, that are not based at a single university or edited by a single person, like the example that I gave above? --Randykitty (talk) 15:20, 30 January 2019 (UTC)
  • @Nemo bis: I can't find it now, but it's just weeks since I last ran into a pair of journals with the same name and wasted hours figuring out what was going on until I noticed sources giving two different publishers (in the humanities there are a lot of names like Poesia, Caliban, etc.: poetic but ambiguous). In that case the publishers were different learned societies in the same or a very closely related field, most likely they were societies for the same area independently created in the UK and North America (a common occurence, but a lot of them have merged post-internet). Most journals in my field have such a publisher and then use a publishing house (Springer, T&F, MIT Press, OUP, etc.) for the technical and practical bits. Giving the latter isn't completely useless (Springer vs. T&F is still a disambiguator), even if exceedingly generic; but the actual societies and organisations publishing these journals are good disambiguators. I've previously given the example Shakespeare Quarterly which is published by Folger Shakespeare Library, and used the publishing house services of Johns Hopkins University Press until a month ago when they switched to Oxford University Press. Taylor & Francis is also the actual publisher of the journal Shakespeare (in addition to being a publishing house for third party journals). In these examples, JHUP and OUP are basically journal database providers like JSTOR and Project MUSE: they change over time and provide little or no information in terms of disambiguation or assessing reliability. The Folger and T&F, for the respective journals, provide both disambiguation and information to help assess the source.
    If I go looking for examples (rather than ones I've actually run into myself while editing) I quickly find that there are two journals called Africa (one by CUP, one by Istituto Italiano per l'Africa e l'Oriente), two called Agenda (ANU Press and T&F on behalf of Agenda Feminist Media), two American Art Journal (Henry C. Watson and Kennedy Galleries), two American Studies (Mid-America American Studies Association and Universitätsverlag WINTER), two The Art Journal (D. Appleton and CAA, both in New York, incidentally), two The Art News (The Art News Company and Sadakichi Hartmann, also both in NY), two Arthuriana (Society for the Study of Medieval Languages and Literature, and Scriptorium Press for International Arthurian Society-North American Branch), two Asian Perspectives (Lynne Rienner Publishers in South Korea and University of Hawai'i Press). TLDR? That's the ones starting with A. On one database. With no cross-referencing.
    Unlike DGG I can't think of a single actual instance where I needed location for scholarly journals (magazines and newspapers, sure, but not journals). However as a constructed example it's easy to imagine a case similar to the above: Dingus Danglers United forms in both the UK and North America, and both start publishing Modern Dingus Dangling. Neither title nor publisher is sufficient for disambiguation, so you're dependent on location to tell them apart. Then there are Sunday Times (Islamabad, Pakistan), Sunday Times (London, England), The Sunday Times (Perth, Western Australia, Australia) (see The Sunday Times (disambiguation) for more examples). Or the gazillion papers published with The Times on its front page, but always called The Times of London or The Times of Northwest Indiana (note the title on the front page there). For older periodicals that are not modern academic journals (often called a "journal", but closer to what we think of as a magazine or newspaper) there are probably plenty of examples. There were a whole gazillion funnily named publications like The Universal Magazine, The Spectator and The Spectator, The Gentleman's Magazine, or the Monthly Review and the Monthly Review and the Monthly Review. Some of these are sometimes best treated as journals, even though they differ somewhat from what we typically think of as a modern academic journal.
    @Randykitty: This RfC came about because CitationBot is forcibly removing location and publisher parameters from citatation templates in violation of CITEVAR and without seeking consensus or BRFA. Your argument is a good one in terms of editor practice (i.e. "What details should an editor add for their citations?"): most journal citations do not need a location, and publisher needs judgement to decide whether it is relevant (but see above for why that's slightly more often then immediately obvious). However, if merely "the vast minority (1% at most)" legitimately needs either parameter, that is also an argument that 1) CitationBot should not automatically remove these parameters and 2) these parameters should not be disabled/deprecated in cite journal. The conclusions will be very different depending on whether we are discussing what a bot should automatically enforce, or what is best practice usage for editors. When you above write "Remove and do not display" your !vote is likely to be counted in favour of a bot or bots mechanically removing all instances of these parameters in all articles using cite journal—including your "1%" of articles who legitimately need them—and ((cite journal)) being altered such that giving |location= and |publisher= will be an error. Does that accurately reflect your position? --Xover (talk) 20:01, 30 January 2019 (UTC)
  • Actually, it does. I don't have any examples handy and have no time right now to search for them, but I have seen cases where, for example, PubMed uses a disambiguator in a journal title, such as: Journal of Foo (Bristol). Citations in articles to this journal will use that name. So "location" in this case is not a separate field, but part of the journal name. If we had two articles to such a pair of journals here, we'd actually do something similar. For most modern journals, as I argued above, the "location" parameter is impossible to determine. --Randykitty (talk) 21:07, 30 January 2019 (UTC)
The method just mentioned is the usual library method in distinguishing titles. For modern journals, there will be instance where at least the place field will be needed. The need for a publisher field is less frequent, but there will be situation where it too is needed, especially for small journals, neither the place nor publisher is necessarily stable.There are also confusions in the title that even place and publisher will not resolve, especially when a journal ceases and restarts either immediately or many years subsequently with the same title but a different set of editors/publishers/sponsors, in which case the beginning and ending dates of both serials are added to their titles, as Foo (1909-1949), Foo (1950- ) . We could do this also, but very often the user will not be aware of the difference, and will not cite it as a matter of course. I can easily find examples that would defy and one single simple method. The general approach for libraries is to include in the record every pssible ariation, but not necessarily display it. There fore I would never think of reoving the fields from the template entirely, but it takes a human who know the field when to display them. DGG ( talk ) 01:26, 31 January 2019 (UTC)
This The general approach for libraries is to include in the record every pssible ariation, but not necessarily display it. There fore I would never think of reoving the fields from the template entirely, but it takes a human who know the field when to display them. is a cogent summary of how to approach the issue. It's also, incidentally, a good description of one aspect of CITEVAR: it needs human judgement and is subject to local consensus. --Xover (talk) 06:10, 31 January 2019 (UTC)
Two thoughts:
  • It would be inaccurate to imply that a desire to promote the use of human judgement or to encourage local consensus was a significant motivation for CITEVAR. My main goal was to stop edit wars between humans whose judgement differed. adapting the WP:ENGVAR approach has stopped most of those disputes. I would not necessarily object to replacing it with a rejection of all variation in favor of a single, unified style (which is the other major way to stop edit wars over ref formatting).
  • User:DGG, among the editors I spend the most time around, I think that a publisher might be added for specific purposes, i.e., to differentiate an academic journal by the Society of Respectable Folks from a pseudoacademic journal by Crackpots R Us. A neutral place-based differentiator does not have the same value in that instance. WhatamIdoing (talk) 20:55, 11 February 2019 (UTC)
Yes indeed, WhatamIdoin, thanks for mentioning it. That has become a critical factor, and it's becoming more and more important as predatory publishers are deliberately using names that are almost exactly like those of respected journals. DGG ( talk ) 06:35, 12 February 2019 (UTC)
  • Your comment actually illustrates my point. Springer is absolutely not "a US corporation". It has offices in the US, sure, but also in Japan, China, the Netherlands, and a lot of other places. In origin, it was a German company, but even within Germany it always had multiple locations. I follow your argument that any tidbit of information can be useful in some situations. My point is that this only is true when the information actually is correct. In the case of BG, almost any location chosen will be incorrect and misleading. (BTW, the title of the journal is unambiguous, even though its generic, because there's no other journal with a name like that, there isn't even a Journal of Behavioral Genetics or something similar...). In the very few exceptional cases where city of origin is important in the name of a journal, they are cited (in PubMed, for example) as name="Journal of Foo (New York)", the location being part of the journal name, not something in a separate field. --Randykitty (talk) 15:18, 3 February 2019 (UTC)
  • Which was why I said I would not suggest a |location= be added to citations to this particular journal (also: give me a break. I spent 90 seconds looking at your example. Your "Gotcha!" style of argument based on an irrelevant aside in one sentence of my reply is not constructive. Nothing in my reply was predicated on Springer being a US corporation: it was given as an example of a location too unspecific and tangenital to be relevant.). I also very much doubt there is a significant number of cases where a location is needed for articles published electronically (the vast majority of these have unique identifiers that are far better than any number of bibliographic details, iff correct). Any cases of this will be pre-electronic publishing. However, note that the larger Springer conglomerate being multinational is irrelevant: if the particular arm publishing the journal gives a particular place of publication then that is the relevant location. For example, Oxford University Press titles are always notionally published in Oxford (and sometimes New York and New Dehli, depending on the arrangements for a given title), but OUP is most definitely a multinational. And a lot of older Springer-Verlag titles (books, I mean) were published in (that is, the publisher gave its place of publication in the colophon as) |location=Berlin (caveat that they don't do much in my field and I may misrecall).
    One should most certainly not add a location to ((cite journal))'s |journal= parameter as that would pollute the metadata (in addition to being technically incorrect). Removing a tacked on location or faux "of City" suffix from a journal name, unlike removing a |location= parameter, would actually be correct and a good thing to do! In comparing relative harm, having available a |location= parameter that is only rarely used, and even more rarely is used incorrectly, does essentially no harm ("It offends me!" is not actual harm). Polluting the metadata for |journal= does some harm. Note that I'm not arguing that there is some pressing need for |location= that would justify adding support for this parameter to ((cite journal)) if it didn't already have it. But the support is there; there is even less need to remove it; and it nicely mirrors the same parameter in the other citation templates (I'd have to check to make sure, but I suspect we'd actually need to special-case to prohibit |location= for ((cite journal))). It makes no kind of sense to get rid of a parameter that might possibly be used incorrectly some time by deliberately placing incorrect information in an unrelated parameter. And in no case can a bot mechanistically determine the merits of a given instance of a citation's |location= parameter. --Xover (talk) 17:32, 3 February 2019 (UTC)
Degree of action
⬅ ⊖ ➡
Hinderance No difference Helpfulness
I see in the posts above examples of when the |location= and |publisher= parameters have been helpful or have offered no help. But I don't see where anyone has identified an instance where the parameter has fallen under the hinderance column, in that it made a source more difficult to locate. Providing an example of the parameter's hinderance in locating a source would seem to offer the best argument for removal. Without it, the move towards excluding the parameter feels like it's based on aesthetic qualities alone. Regards,  Spintendo  19:48, 7 February 2019 (UTC)
The parameter's hinderance is that nobody ever agrees on what it should contain, so editors keep quarrelling about what to put in it, which is a waste of Wikipedia's resources. I think the discussion above was able two suggest one or two cases of journals where there might be consensus on how to use the (location) parameter, but we don't yet have a proposal for how we could test those cases and write comprehensive documentation for the users to know how to use the parameter. I don't remember seeing a single example where consensus could be found on what to put in the "publisher" parameter. So for now there is no consensus on putting anything at all on these two parameters. Nemo 07:06, 8 February 2019 (UTC)
No, there is no consensus to support the removal of these parameters. And absent a global consensus that means CITEVAR obtains, leaving their use up to local consensus on each article as for all other such citation issues. CitationBot is now, was previously, and has apparently during this RfC intensified its removal of these parameters, without consensus and in violation of CITEVAR. And bots editing against consensus is sanctionable! Please stop before this ends up at the drama boards (which really shouldn't be necessary to resolve this issue!). --Xover (talk) 09:20, 8 February 2019 (UTC)
Removal is the status quo of ten years, so you need consensus to stop it, not to continue it. Are you saying that WP:CITEVAR in your opinion means that I can put whatever I want in the "publisher" and "location" field (say, the name and place of birth of the last author of the paper) and then people need to find consensus on each talk page to change it? Nemo 09:28, 8 February 2019 (UTC)
Whatever CitationBot has been doing for 10 years, absent a valid bot approval and consensus to begin doing anything, no bot is allowed to edit. At all. The burden is very much on CitationBot's maintainers to demonstrate both community consensus and authorization for this behaviour. And please don't resort to strawman tactics. Of course there are certain de facto limitations on what can go in citations in general, and specific such in specific cases. Given more than just a small entrenched group of editors on a given article, any of the more extreme theoretical examples (author's home phone number in |url= say) gets reigned in right pronto. For example, |url= should obviously contain a valid URL absent very good reasons otherwise (I can't think of any). And there are very good arguments that what gets put in |publisher= should be something that can reasonably be interpreted as a publisher, and a location for |location=, and that it would be strongly preferable that these are as correct as possible. However, whether to include or not various bibliographic datum for a given citation is exactly what CITEVAR regulates. Ditto what threshold of "correctness" should be required for this. For example, the practice of shoehorning unrelated information into a citation template parameter in order to get it to display in the desired way is fairly widespread and common. I would wish editors didn't do that, and I've done my fair share of cleanup of such, but if local consensus is in favor of that there's nothing much I can do. CITEVAR means that these questions are subject to local consensus on each and every article (unless you can establish a community-wide consensus to override CITEVAR for a specific issue). And that's completely irrespective of whether the specific citation in question uses a citation template or not. If you want to try overturning CITEVAR and get consensus for One True Citation Style on Wikipedia (good luck! the last attempt went down in flames last autumn if I don't misrecall badly) you have my !vote, but until then nobody gets to impose their preference on such details by automated editing. --Xover (talk) 10:15, 8 February 2019 (UTC)
  • RIght. So WP has this for some reason included in their "cite journal" template. I was talking about the world at large. I don't know of a single academic journal adding publisher or location (or ISSN for that matter) to citations to journal articles. Not A Single One. Go ahead, name just one. --Randykitty (talk) 18:50, 13 February 2019 (UTC)
  • Chicago for one, as has been mentioned earlier: If a journal might be confused with another with a similar title, or if it might not be known to the users of a bibliography, add the name of the place or institution where it is published in parentheses after the journal title. Umimmak (talk) 22:51, 13 February 2019 (UTC)
  • The important part here being "IF a journal might be confused". That is something else entirely than what is being proposed here, namely always including this info. As I have indicated above for Behavior Genetics, a "location" parameter will in 99% of cases be ambiguous or impossible to determine. The remaining 1% are the cases that "Chicago" talks about. I maintain that there is no single style outside of WP that always includes location, publisher (and ISSN). --Randykitty (talk) 11:29, 14 February 2019 (UTC)
  • Nobody here is saying those parameters should always be used, just that a bot should not always automatically delete the information in those parameters when an editor has decided that those parameters would be helpful additions to particular citations. Umimmak (talk) 11:34, 14 February 2019 (UTC)
  • Literally nobody in this entire thread has proposed (or even just suggested or implied) that these parameters should always be added. I even responded to your initial !vote explaining precisely what this RfC is actually about, since you at that point appeared to have possibly misunderstood. Since you then reconfirmed your position (and argued vehemently for it) I took it for granted you had taken the time to make sure you had understood what the question was before !voting. If you are still this confused about the basic premise of this RfC after voting, objecting to a neutral closing (thus dragging this out needlessly), and arguing for a further two weeks, I begin to question your WP:COMPETENCE. The alternative explanation is that this is the second time in this discussion that you have resorted to assigning an obvious and transparent strawman to those you disagree with, which would put your good faith in question. Which is a pity because what seems to be your basic position—that these parameters are needed so rarely and are so easy to get wrong when used, that they should be actively forbidden in all circumstances on Wikipedia—is perfectly valid on its own (I just happen to disagree with it). But when you dare others to "Go ahead, name just one", and they do, and you start pretending the issue is something else entirely without even acknowledging the egg on your face? Not a particularly good approach to convincing other editors to your position, lets put it that way. --Xover (talk) 13:30, 14 February 2019 (UTC)
  • As correctly evoked by Espresso Addict above, if the parameter is in the template, many editors will fill it in, necessary or not, just as they will try to fill in any infobox parameter. And I still maintain that there is not a single citation style that (except in a vanishingly small number of cases) includes publisher, ISSN, and/or location. Having these parameters is asking for trouble. --Randykitty (talk) 16:56, 14 February 2019 (UTC)
  • Thank you, and duly noted. I acknowledge that these are valid arguments. I disagree strongly that the potential trouble of having them outweighs the potential trouble of not having them. Note that the ISSN-related parameters are not at issue in this RfC. Note also that the proper venue for the part of your argument that pertains to citation style belongs at WP:CITEVAR. --Xover (talk) 17:40, 14 February 2019 (UTC)
I'm finding it hard to imagine people add the publisher just because they see there's a parameter for it. ((cite journal)) has almost a hundred parameters, do people fill them all in? The template's documentation doesn't even have any examples with this parameter. I don't know what the reasons might be for any instances of improper use. Maybe people are carrying over their habits from using ((cite book))? Maybe there are referencing tools that automatically populate the parameter? – Uanfala (talk) 22:57, 15 February 2019 (UTC)
Yes, the publisher field is usually populated by VisualEditor or other automated systems which use Zotero. It's not something that editors usually enter intentionally. Nemo 07:26, 16 February 2019 (UTC)
But Zotero doesn't have a publisher field for journals, at least not the version I use. And if it's wikipedia's own tools that do that, then this is something we have some sort of control over, so maybe the efforts should be directed at making sure these tools don't automatically populate that field? – Uanfala (talk) 15:46, 16 February 2019 (UTC)
I am curious if you would therefore agree that removing is okay when a doi, pmid, pmc, or ISSN is present? It seems to follow. AManWithNoPlan (talk) 21:28, 19 February 2019 (UTC)

Closure

On 07:13, 3 February 2019, I closed this RfC with a result of "stop removing parameters" in response to a request for closure. On 10:58, 3 February 2019, Randykitty challenged the closure on my talk page:

Hi, I feel that your closure here does not take into account my serious concerns regarding the "location" parameter. In addition, I'm surprised this got closed anyway, as the discussion seemed to be still ongoing.

To allow for further discussion, I've undone the closure. When the RfC is ready to be closed, any editor may request another closure at Wikipedia:Administrators' noticeboard/Requests for closure. — Newslinger talk 11:50, 3 February 2019 (UTC)

@Newslinger: In case you are no longer watching this page, Cunard relisted the RFC (which was probably right to do on his part). Just an FYI. --Izno (talk) 20:07, 7 February 2019 (UTC)
Hi Izno, thank you for the update. I did notice that Cunard relisted the RfC, and I had sent them a thanks (which is unfortunately hard to see). I will probably not monitor this RfC anymore, so please ensure that another request for closure gets submitted if you're interested in seeing this RfC closed promptly after the discussion settles. — Newslinger talk 20:45, 7 February 2019 (UTC)
In the meantime Citation bot has been furiously removing publishers from journals today. Shouldn't this stop, at least while the RFC continues and maybe longer depending on the RFC outcome? —David Eppstein (talk) 07:00, 8 February 2019 (UTC)
To stop while discussion is ongoing would indeed be in keeping with… pretty much every single policy, guideline, and good practice on the project. If removal has intensified today that would constitute actively and directly disruptive editing and violating the bot policy (vs. the neglectful failure to get proper consensus first or of stopping once the task was questioned). However, to stop would mean admitting that CitationBot was removing these parameters without consensus and valid bot authorization in the first place, which I imagine might be too much to ask of its operators. Well, operator singular, strictly; the bot has an operator and two "assistants". But the operator and one of the assistants hasn't participated in this RFC at all, nor ever appear to edit the bot's talk page. The last time the actual operator edited the bot's talk page was 5 months ago (and the bot's user page was 17 months ago), and that was a bug report apparently effectively addressed to the "assistants". The other "assistant" hasn't edited the bot's user page in 54 months, and the talk page in 16 months. The actual operator of the bot gives no indication that they are aware of this discussion, which in itself is a violation of the bot policy, and rather negligent given they are actually responsible for the bot's edits. --Xover (talk) 09:20, 8 February 2019 (UTC)
If you look at the journal cover here you will see that it clearly indicates "Routledge". I think the journal itself is a better source than PubMedCentral. Library databases are notoriously unreliable for info like this and more often than not severely outdated. Just one example: the reputed National Library of Medicine's catalog indicates as publisher for Genes, Brain and Behavior "Oxford: Munksgaard" (see here). This is wrong in many senses. Munksgaard never was based in Oxford, it was a Danish publisher based in Copenhagen. It was bought by Blackwell (one of whose main offices was indeed in Oxford), who operated it for a time as an independent inmprint, after a while renamed it "Blackwell Munksgaard", then absorbed it completely, and then was taken over by Wiley to form Wiley-Blackwell. An amateur would all too easily populate the publisher and location fields with "Munksgaard" and "Oxford", the former being most certainly incorrect and without any way of knowing which of the many Wiley offices around the world actually is handling the publication of this journal. To get back to your example, Routledge's journals are all hosted on the T&F Online site, as they (and other publishers) belong to the same parent company (Informa), which operates T&F and Routledge as separate imprints. Similarly, Elsevier's ScienceDirect online platform not only houses journals published by Elsevier and its imprints, but also some journals that are published by other companies. Nevertheless, every page on ScienceDirect has "Elsevier" at the bottom of the page, so if you're an amateur and do not know about this stuff, you'd mistakenly put "Elsevier" on all those journals. --Randykitty (talk) 17:01, 16 February 2019 (UTC)
The bigger issue is that journals routinely change publishers over the years. They get acquired, moved from imprint to imprint, their publishers rebrand themselves and change names and themselves get acquired, and so on. If you take something like Nature, over the years, for the same article you'd be landing on a page that declared the publisher to be Nature Research, Nature Publishing Group, Springer Nature, or Holtzbrinck Publishing Group. Maintaining who publishes the journal now, or deciding amongst this corporate shell game who the publisher is, or tracking back the 1953 published issue to say who happened to publish it at the time, deciding if we mention the original or current publisher, etc... it's pointless work. That's why no citation styles say to mention who the publisher of a journal is. If you care, follow the DOI and figure it out. Headbomb {t · c · p · b} 17:10, 16 February 2019 (UTC)
@Randykitty: Thanks for your amateur advice, but it doesn't alter the fact that Europe PMC indicates the publisher's website for the content supporting the article text is Taylor & Francis. Anybody following the citation can see that. So, what's your justification for claiming that T&F isn't the publisher of the information that I cited? --RexxS (talk) 18:39, 16 February 2019 (UTC)
It's the journal cover itself which prominantly displays "Routledge" and does not mention T&F at all. Or do you think that the journal itself is confused about who is publishing them? --Randykitty (talk) 18:50, 16 February 2019 (UTC)
If you assume that it should be T&F since Routledge is owned by T&F, then the publisher becomes so big to be useless, and furthermore then the publisher should be listed as Informa, since they own T&F. Either way T&F is wrong. AManWithNoPlan (talk) 18:54, 16 February 2019 (UTC)
Guess what the source is for the image of the journal cover that Randykitty relies on. Answer: https://www.tandfonline.com/action/journalInformation?journalCode=gspm20 - the Taylor & Francis website.
I know that the publisher is Taylor & Francis for two reasons:
  1. Europe PMC says it is;
  2. The article text states "Paulev and Zubieta have created a new conversion factor in order to make any sea level dive table usable during high altitude diving in 2007"
    The source that verifies that statement is https://doi.org/10.1080/15438620701526795
    That turns out to be the webpage https://www.tandfonline.com/doi/abs/10.1080/15438620701526795?journalCode=gspm20
    and there is absolutely no doubt that the publisher of that webpage is Taylor & Francis. It's all over the page in big letters for anyone to see.
    Only in some Bizarro universe would anybody try to claim that the publisher of the source that verifies the article text would be "Routledge". This is what happens when editors who don't write content try to impose their preconceived notions on articles. You're simply wrong. Get over it. --RexxS (talk) 20:46, 16 February 2019 (UTC)
[1] makes it pretty clear that the publisher is Routledge. Taylor and Francis is involved through some corporate shell game / distribution / imprints arrangement "Taylor & Francis Group journals are published under the Routledge and Taylor & Francis imprints". Headbomb {t · c · p · b} 21:01, 16 February 2019 (UTC)
None of that is relevant to the fact that the source supporting the article text is Taylor & Francis' website. There's no mention of Routledge on that webpage. From our article on Taylor & Francis:

In 1998 Taylor & Francis Group went public on the London Stock Exchange and in the same year the group purchased its academic publishing rival Routledge for £90 million ... Taylor & Francis merged with Informa in 2004 to create a new company called T&F Informa, since renamed back to Informa ... Taylor & Francis Group is now the academic publishing arm of Informa ... Taylor & Francis publishes more than 2,700 journals ... It uses the Routledge imprint for its publishing in humanities, social sciences, behavioural sciences, law and education and the CRC Press imprint for its publishing in science, technology, engineering, and mathematics ... The company's journals have been delivered through the Taylor & Francis Online website since June 2011.

So please don't tell me that Taylor & Francis didn't publish https://www.tandfonline.com/doi/abs/10.1080/15438620701526795?journalCode=gspm20 . You're clearly wrong and you need to just accept that my edit was correct. --RexxS (talk) 21:27, 16 February 2019 (UTC)
"You're clearly wrong" is not an especially constructive attitude. Considering you're currently the only person advocating for that interpretation, in a group of people with considerable experience about scholarly communication, maybe you could accept that this field is more intricated that you initially admitted. Nemo 21:31, 16 February 2019 (UTC)
So after you've got all the rest wrong, your last resort is an ad hominem on my attitude?? Comment on the content, not the editor. I'm the only one, am I? and according to you that makes me wrong? LOL - it's strength of argument, not numbers that count and all your arguments miss the point by miles. Finally you make an "appeal to authority". Well, I'm also an experienced editor with considerable content experience and I know how to find the publisher of a website without being patronised, thanks. The facts of T&F's relationship to Informa, Routledge and CRC Press are laid out for you above. It's not rocket science, and anybody can see that Taylor & Francis Online is the publisher of the source used for the article. They are the publisher of all of that group's online content as is made clear at https://web.archive.org/web/20130509172503/http://resources.tandfonline.com/documents/library-faqs.pdf - maybe you could accept that this particular corner of the field is nowhere near as complicated as you'd like to pretend? --RexxS (talk) 21:48, 16 February 2019 (UTC)
  • RE "ad hominem", somebody (was that perhaps you) wrote above "This is what happens when editors who don't write content try to impose their preconceived notions on articles". Comment on the issue, not the editors, indeed. Putting a PDF online is only the last step in the long process that is academic publishing. In that sense, "Taylor and Francis Online" does not publish anything. It's the online access platform for journals published by Informa and its subsidiaries, which includes T&F and Routledge (and perhaps others, I didn't look closer into that). The link you're citing above (this one) has at the bottom a statement: "Copyright © 2018 Informa UK Limited". Does that now mean that the journal is published by Informa? --Randykitty (talk) 23:02, 16 February 2019 (UTC)
Elsevier does not exist either. It is part of another company, and there are in fact a lot of divisions called Elsevier that are not the same "publisher". Unless you know which one is the right one, then you are wrong. Based upon the above idea, one could just as easy say that sciencedirect is the publisher? Crannking my saltiness up to eleven, does that make the hosting firm the publisher? AManWithNoPlan (talk) 19:23, 17 February 2019 (UTC)
It turns out we do have over a hundred pages with "ScienceDirect" in the "publisher" parameter. Now let's start a hundred talk page discussions with the respective editors to make sure whether they think it's a "bizarro world" to think they might have misunderstood something. Nemo 20:34, 17 February 2019 (UTC)
I thought this was a discussion about closure of the RFC about whether to stop automatically removing publisher and location from references using the cite web template, not an argument about what the publishers of a particular journal are called - these sorts of arguments about who is the publisher are equally applicable to thinks like books, where (I believe) no-one is arguing for removal, so do not seem particularly relevant to a discussion about Cite Journal. This discussion needs closure by an experienced, uninvolved editor rather than just a series of shouting matches. Please keep the discussion on topic.Nigel Ish (talk) 22:18, 16 February 2019 (UTC)
  • The difference with books is that almost every citation style in the real world that I know about includes publisher and location. The same cannot be said for journals. And the shouting match above indicates how complicated these issues can get for journals (especially when dealing with somebody who apparently does not understand what an "imprint" is... Which also would be the case in many many articles where editors unfamiliar with academic publishing practices include references to academic articles. --Randykitty (talk) 23:02, 16 February 2019 (UTC)
Actually for books, once a book is published, the publisher doesn't change. Every printing of A. Pickering (1984). Constructing Quarks. University of Chicago Press. pp. 114–125. ISBN 978-0-226-66799-7. has been published by University of Chicago Press. Reprints by other publishing houses would have been printed in a different year, but there is no question or ambiguity about who published the original copy. And in the case of books, the publisher is often the disambiguator: You will have several books with the same title. You will not, however, have two articles of the same name, in a journal of the same name, where you'd need a publisher to distinguishing between the two. Headbomb {t · c · p · b} 02:33, 17 February 2019 (UTC)
  • This is Wikipedia, an online encyclopedia and we use online sources at least as often as printed ones. The distinction between different journals, books, magazines, etc. blurs when we are using an online source.
    You're wrong again: I understand perfectly well what an imprint is in the UK, and it's not a link to a dab page (check the links, LOL). The problem I have is that I'm dealing with someone who apparently does not understand what a publisher is. To be helpful: "publisher: the name of the organization that actually published the source." So the question becomes "what organisation actually published the source?" For a webpage, you often have that information in its metadata. In the case of the source used, https://www.tandfonline.com/doi/abs/10.1080/15438620701526795?journalCode=gspm20 you will find: <meta name="dc.Publisher" content=" Taylor & Francis Group " />. You're confusing copyright owner (Informa, the parent company) with the publisher of the webpage (Taylor & Francis). Hope you've got that straight now. --RexxS (talk) 00:11, 17 February 2019 (UTC)

So, since the web publisher and the print publisher are different; does that mean we should just delete publisher? Also, does that make jstor the publisher? What about journals that publish on multiple websites? Lastly, the publisher of the website T&F is transient and prone to change obviously, unlike the print version which is written in stone. AManWithNoPlan (talk) 00:44, 17 February 2019 (UTC)


Build the web does not apply to references and in fact overlinking is discouraged in references. AManWithNoPlan (talk) 02:52, 16 February 2019 (UTC)

Nonsense. You have it exactly backwards. Of course "Build the web" applies to references just as it does to all article text that is displayed to readers – for exactly the same reasons. Overlinking is discouraged everywhere, but a sensible amount of linking is good in references, and more overlinking is tolerated there than in normal article text. --RexxS (talk) 15:24, 16 February 2019 (UTC)
Irrelevent and pointless I vote this discussion to be pointless now that the bot has stopped removing when there is not a unique identifier. This eliminates most of the arguments against removal. Since only Citation Bot (and no other bot) has ever had consensus for removal, I doubt any other bot would ever be able to get this going again. AManWithNoPlan (talk) 22:33, 16 February 2019 (UTC)
1) CitationBot did not "have consensus" for this task, it has merely not been challenged on it before (which counts as consensus in only the weakest possible sense), and bots need actual bot authorization through BRFA in any case. After this RfC it seems likely to have actual consensus against this task (which is why, I suppose, its proponents are so determined to prevent closure). 2) Stopping in some cases but not in all cases is not sufficient when consensus is against bot removal: that you and Headbomb decided that was probably good enough on the bot's talk page does not supercede community consensus here. Bots editing against consensus do not become more acceptable by trying to avoid getting challenged on it. 3) The proponents of bot removal make one very good point: determining the correct (and thus determining what counts as incorrect) publisher is often complicated and metadata in bibliographic databases is frequently incorrect or misleading: they just don't realise that this is a strong argument for why a bot cannot perform this task and discussions are needed to determine each specific case (that "T&F" is incorrect in the example above does not mean "T&F" is thus always wrong: that's an actual fallacy). 4) I concur with Nigel Ish: this discussion has devolved into entrenched positions and ad hominem and should be closed (by an experienced closer). --Xover (talk) 07:21, 17 February 2019 (UTC)
No, it simply shows that there is no consensus on filling the publisher parameter, as we were not able to find a single case where the location and publisher would be uncontroversial. The automated removal allows to counter the indiscriminate automated addition and to make sure that thousands of cite journal calls have consensual content. It would be illogical to conclude that there is a consensus to multiply the lack of consensus. Unless we come up with precise and actionable guidelines on how to determine what names must be picked to fill the location and publisher parameter, we'll see thousands of pages suffer edit wars and the same discussion as above (on T&F vs. Routledge vs. Informa etc.). Nemo 07:52, 17 February 2019 (UTC)
You appear to be quite confused. The previous close, and the most likely conclusion in the next one (up to whoever closes it, but it appears to be the most likely outcome), is that there is no consensus for the automated removal of these parameters. At that point CitationBot will be making automated edits against consensus (if the relevant function is not disabled). Which will get it blocked. Which is not unlikely to get its operators sanctioned since that would be an instance of deliberately and knowingly using a bot to edit against consensus. Which is pretty dumb given all the uncontroversial cleanup the bot could be doing, to great benefit for the project, instead of trying to impose the WP:IDONTLIKEIT preference of a small number of editors. Oh, and there is no indescriminate automated addition of these parameters (that was someone's idle speculation) and the "thousands of pages" do not currently "suffer edit wars", mainly due to CITEVAR, except insofar as CitationBot's proponents created such conflict by attempting to impose their preference on them, in violation of CITEVAR. Rule one for making automated edits on Wikipedia is to make darned sure you have consensus for the changes before you begin, and to be humble, responsive, and sensitive to challenges to them afterwards. CitationBot gives every appearance of operating on the complete opposite assumption (echoes of a certain editor using a bot to enforce their idea of NFCC compliance back when dinosaurs roamed; which didn't end well for anyone involved and caused irreparable harm to the collaborative environment along the way. Really. We still see conflicts and bad blood that started there.) --Xover (talk) 09:55, 17 February 2019 (UTC)
It's a bit rich to cite IDONTLIKEIT when the people arguing for removal of these parameters provide solid reasoning and clear examples, whereas those who argue for maintaining this don't get any further than ITSUSEFUL or ILIKEIT and apparently know so little about academic publishing that they think the act of putting a PDF online is what publishing is about... --Randykitty (talk) 10:51, 17 February 2019 (UTC)
Arguing that our stopping some removals means that we think we are wrong or think we have lost is to violate Wikipedia policy against reading minds. Being kind and letting someone get their way from time to time does not mean they are right or that you agree with them. While I find the ITSUSEFUL argument to largely invalid since the information is so often wrong or so generic as to not be useful. Nonetheless, the changes made to the Bot have eliminated the ITSUSEFUL arguement. AManWithNoPlan (talk) 17:02, 17 February 2019 (UTC)
@AManWithNoPlan: Presuming you are replying to my previous message a bit up the page regarding the "mind reading", I see I owe you an apology. In hindsight I realise that what I wrote comes across as not just excessively snide, but also as if I were "daring" you to follow consensus here and that that would "prove" that it had previously happened without consensus. That's the plain and most obvious interpretation of what I wrote: it was clumsy and careless, and for that I apologise. Mea culpa!
Not that it matters much this far after the fact, but the meaning I intended, but failed badly, to convey was simply that "A" would of necessity imply "B" (in a strictly logical sense), and since "B" is a presumably undesired implication, it would be unreasonable to expect anyone to be particularly motivated to do "A" without a compelling reason (i.e. the point was more that it was "too much to ask"). It was not intended to suggest any deficiency in you or anyone else, and you may, of course, also disagree with the very premise that "A implies B". I did, however, hope to see you, or another of CitationBot's proponents, respond to say that such concerns would not have been to the detriment of disabling this function if the assessed consensus here was against it. I still hope that is the case.
As for the rest of your (plural) points, I'm inclined to let you have the last word: the discussion here is now simply running in circles with ever increasing levels of vituperation, and I would very much like to see this end before people start dragging each other to the drama boards for issues unrelated to the matter actully at hand. --Xover (talk) 06:57, 18 February 2019 (UTC)
Nemo said:

It's not about giving total freedom to anyone to add whatever they want to citations unless someone objects on each talk page. Otherwise tomorrow someone can start adding fax, phone number, street address, ZIP code and their library's preferred unique identifier in the "publisher" field on the grounds that it's what they find most useful for the specific citation. Nemo 08:30, 28 November 2018 (UTC)

You joke, but this is actually much more common than you would want to believe; I have fix hundreds of pages and thousands of citations like this by hand. AManWithNoPlan (talk) 17:06, 17 February 2019 (UTC)
If you want to make it explicit, then that's something you mention in prose, not implicitly through referencing. Likewise, if you want to 'warn someone' that the information is sourced to a predatory journal, you do that in prose, or better, you don't cite that predatory journal in the first place. Headbomb {t · c · p · b} 21:56, 19 February 2019 (UTC)
Yes, that's one way to do that. – Uanfala (talk) 22:12, 19 February 2019 (UTC)
It is the only acceptable encyclopedic way to do it. Headbomb {t · c · p · b} 22:24, 19 February 2019 (UTC)
There are a couple of very good reasons for the bot to leave publisher and location alone by default, even though most style guides do not recommend including this data. These reasons are due to the different editing processes between WP and academic journals, and the different audience for each publication. A style guide such as Chicago provides advice for authors writing articles that will be edited to ensure consistent referencing (among many other things), and then published in journals read by specialists in the same field. There's a reasonable presumption that readers will be familiar with all common journals in their field, and that the copy editor will ensure that journal names are listed consistently. In those circumstances, publisher and location are redundant information.
In Wikipedia, we have a far broader audience and pool of editors. There's no reason to think that citations will be referenced consistently; there's no consistent editor ensuring that style is rigorously followed (and going back to the author to supply missing info); and many readers will be entirely unfamiliar with the journals cited for a particular article. As an example, I'd challenge anyone to track down the journal Anatolian Archaeology without knowing that the publisher is the British Institute at Ankara. Since yesterday, a WP user is now able to follow the redirect I created to find out that this journal is now called Heritage Turkey. Prior to that, tracing the two references in Wikipedia that cite Anatolian Archaeology was very difficult without knowing the publisher. Now, if User:Citation bot removes the publisher that's a definite loss of useful information.
I can live with automated deletion of publisher and location if the citation has a persistent reference, such as a DOI. Absent that, I'm sure that the bot is taking a large number of confusing cites that can be tracked down by using publisher and location parameters and rendering them useless. That should be halted and reversed. Rupert Clayton (talk) 01:51, 22 February 2019 (UTC)
That's a lot in line with what I have been saying. Linking the journal's to a Wikipedia page is much better than including publisher/location. It eliminates the need to search. Considering that in your examples, the publisher has changed names, just listing that as text is not that useful. Oscure journal that it is. AManWithNoPlan (talk) 04:42, 22 February 2019 (UTC)
Rupert, challenge accepted: search the name in the most popular catalog in the world, click the first result. Time: 20 seconds. Usefulness of the publisher name: 0. Nemo 13:40, 22 February 2019 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
More