Template talk:Citation/Archive 6

author-link error
Found this while reading an article. First, the reference code: Which results in the visual of this: [|Collins, Andy] (May 2002), "Spelljammer: Shadow of the Spider Moon", Polyhedron (151): 16–65 The author name does link to his website, but as you can see, the format is rather strange and has uneeded characters on it. The extra symbols should be removed to make it more consistant with standard referencing. Turlo Lomon (talk) 19:53, 10 July 2012 (UTC)
 * Ah, that old chestnut. As stated at Template:Citation (which comes from Template:Citation Style documentation/author):
 * authorlink: Title of existing Wikipedia article about the author—not the  author's website; do not wikilink.
 * -- Red rose64 (talk) 20:09, 10 July 2012 (UTC)

Would it be worth adding a check for this? Something like:

---— Gadget850 (Ed)  talk 21:42, 10 July 2012 (UTC)
 * Yes, with the rest of template placed as alternative branch of, so that the error would get rendered instead of the reference, not in the middle. — Dmitrij D. Czarkoff (talk) 22:15, 10 July 2012 (UTC)
 * Would it be worthwhile to add a tracking category for these errors? If there's a significant amount, maybe a bot could monitor the category and remove the offending parameter?  GoingBatty (talk) 17:46, 11 July 2012 (UTC)
 * Not really. The message is clear and it is hard to miss. We shouldn't abuse bots. Furthermore, making editors fix such breakages by hand forces good habit of reading docs. — Dmitrij D. Czarkoff (talk) 18:25, 11 July 2012 (UTC)
 * I would rather put it in the first #if for each authorlink— this will reduce overhead. There are bound to be a number of extant bad uses; we should track and repair these for at least the short term. It only checks for, but I don't see that there would be many other uses. ---—  Gadget850 (Ed)  talk 21:50, 11 July 2012 (UTC)

Documentation for postscript parameter
At present, at Template:Citation/doc, the documentation says:
 * postscript: The closing punctuation for the citation; defaults to a period (.); if the parameter is present, but blank, no terminating punctuation will be used. Ignored if quote is defined.

This is not correct for the Citation template: if the postscript parameter is omitted the effect is the same as if it is present but blank, namely no terminating punctuation will be used. I haven't edited the documentation to correct this because I'm a bit confused by the nested templates and their parameters which make up the documentation. Peter coxhead (talk) 09:36, 17 August 2012 (UTC)
 * Actually it is a bug that should be fixed: unless postfix is specified, the template should render with a period on the end. — Dmitrij D. Czarkoff (talk) 10:14, 17 August 2012 (UTC)
 * That is: line " |PS = " should be replaced with " |PS =  ", if I get this template correctly. — Dmitrij D. Czarkoff (talk) 10:22, 17 August 2012 (UTC)


 * Fixed. Added switch so it now shows: "postscript: The closing punctuation for the citation; defaults to none. Ignored if quote is defined." ---— Gadget850 (Ed)  talk 11:02, 17 August 2012 (UTC)
 * This is not a fix. Citation is a full sentence, so it should end with period unless there is any special case. — Dmitrij D. Czarkoff (talk) 11:16, 17 August 2012 (UTC)
 * This fixes the presented issue in that the documentation did not properly describe how the template actually works. If you want to propose a change to the template, please start a new discussion for clarity. ---— Gadget850 (Ed)  talk 11:23, 17 August 2012 (UTC)

trans_title broken for journal articles
It seems that trans_title does not work for journal articles, and that one has to use trans_chapter instead. This seems counterintuitive and wrong, and besides is not what the documentation says. Example: I'm not entirely sure how to fix this: the template code does pass trans_title along to the core template in either case, but under the wrong core parameter name (the one for the journal name rather than the article title). I'm not sure why the core template then ignores it rather than putting it in the wrong place. We would need to be careful that fixing this case doesn't break the other types of citations. —David Eppstein (talk) 21:24, 20 August 2012 (UTC)
 * With trans_chapter:
 * produces
 * With trans_title:
 * produces
 * With trans_title:
 * produces
 * produces


 * The issue is in citation/core. journal in cite journal is passed to the metaparameter Periodical to citation/core, where a bunch of template magic is performed when Periodical is defined. If journal (or an alias) is defined, then title is treated as a chapter: enclosed in quotes and trans_chapter applied. If journal is not defined, then title is treated as a title: displayed in italics and trans_title applied.


 * I can fix the documentation. It would be nicer if the template did this automatically. The Periodical chunk of the core is very complex (it makes at five changes in formatting that I know of), and I don't have my head it it at the moment. ---— Gadget850 (Ed)  talk 22:20, 20 August 2012 (UTC)

Translation in patents
The keywords language, trans_title (or trans_chapter) do nothing in a patent citation. (Yes, country-code should be a clue that the patent is in German.)





Glrx (talk) 17:11, 23 August 2012 (UTC)


 * Updating the documentation for patents has been on my list. If inventor-surname or some other parameters are defined, then citation/patent is called, not citation/core. citation/patent support a much smaller set of paramters, which does not include trans_title. ---— Gadget850 (Ed)  talk 19:46, 23 August 2012 (UTC)
 * I don't see a problem. The documentation at Template:Citation doesn't mention language, trans_title or trans_chapter. Just because those parameters are provided for the non-patent form doesn't mean that they are necessarily valid for the patent form. -- Red rose64 (talk) 14:40, 24 August 2012 (UTC)

Undocumented parameters: ---— Gadget850 (Ed)  talk 21:32, 14 September 2012 (UTC)
 * publication-number
 * year
 * fdate
 * pridate
 * assign1
 * assign2

No parameter for name of translator(s)?
For translated works, particularly books, the name of the translator is significant information that should be included in a citation. I can't find any fields in the citation templates for providing the translator name. Am I overlooking something, or can this be added? --Orlady (talk) 11:47, 30 August 2012 (UTC)
 * See Template:Citation. -- Red rose64 (talk) 12:04, 30 August 2012 (UTC)

Make postscript default to dot
edit protected Currently postscript parameter defaults to empty string, so that citation (otherwise a proper sentence) is not properly terminated. Please apply this.

As long as it was the long-lasting default until recently, I consider this change to be non-controversial. — Dmitrij D. Czarkoff (talk•track) 15:31, 14 September 2012 (UTC)
 * It is far from uncontroversial. There are huge numbers of citation templates out there that expect no terminator, and where one is provided after the template in the text of the article. This would break them all. I don't know what you mean about "as long as it was the long-lasting default"; I don't remember a time when this was the default. (Maybe you're thinking of cite journal, for which it was and still is the default?) —David Eppstein (talk) 16:20, 14 September 2012 (UTC)


 * I think that this sort of thing needs to be discussed in more detail (in other words I don't think it is non-controversial enough to just be done). So I am going to disable the editprotected request, although the discussion is not closed in any way.


 * Historically, the lack of punctuation was used in articles to add additional items to the end of citation, relying on the fact that the items are separated by commas and there is no period at the end. Many other articles that use the citation put an intentional period after the template. Thus adding a period here would break numerous articles. The lack of a period in this template dates back to at least 2007, it is not a recent change; perhaps a period was inadvertently added recently and then re-removed. &mdash; Carl (CBM · talk) 15:51, 14 September 2012 (UTC)


 * Strongly support no change i.e. no punctuation should be the default. There are two quite different issues:
 * Breaking existing usages: adding a full stop by default would, as noted above, break existing usages, e.g. ", chapter 5 " which would render with ".,".
 * Whether citations should end with a full stop. I think they shouldn't; they aren't sentences. Others clearly think they should. But even if there is a consensus for a full stop, this doesn't justify changing long-established template behaviour.
 * Peter coxhead (talk) 16:46, 14 September 2012 (UTC)


 * Strongly oppose this proposal. The reason I favour the citation template is precisely because it doesn't emit a terminating fullstop. And the reason I favour that is precisely the reason given by Peter coxhead above; citations aren't full sentences. Malleus Fatuorum 16:53, 14 September 2012 (UTC)


 * comment This template uses different punctuation from the templates, including the terminal period. As best I can tell, it was a deliberate design from inception. ---—  Gadget850 (Ed)  talk 17:23, 14 September 2012 (UTC)


 * Strongly oppose. That would break a lot of existing citations.  Also, there are many cases where the automatic period is not desired, and having to include a whole parameter to preclude that quite offsets the dubious advantage of not having to type a single period. ~ J. Johnson (JJ) (talk) 20:57, 14 September 2012 (UTC)

Capitalization of references
Would the source "MLB.com" appear as "MLB.com" in the reference section, or "Mlb.com"? Or are both allowed as long as the article is consistent. As I've come to understand, the rules that are applied to other sources which may appear as all capitals, such as "Life," do not appear in all caps when being referenced (the article is even Life) would make sense to apply to "Mlb.com" as well. Zepppep (talk) 19:45, 1 October 2012 (UTC)
 * Please comment at Help talk:Citation Style 1, not here, in order to avoid split discussion. -- Red rose64 (talk) 20:29, 1 October 2012 (UTC)
 * Exactly what I was about to say. WP:FORUMSHOP seems relevant. —David Eppstein (talk) 20:30, 1 October 2012 (UTC)
 * Well, being as I am the one who started that thread...but thanks. Zepppep (talk) 06:45, 2 October 2012 (UTC)

Template:Cite quick
I see that Template:Cite_quick was added to Julian Assange today (i reverted this). Have things changed about the templates usage? And if so why is the new coding not being implemented into this main cite template?Moxy (talk) 17:57, 27 October 2012 (UTC)
 * The article size is about 148kb and I count over 200 templates which were replaced with cite quick. It appears to me that the use of cite quick in the Julian Assange article was well within the usage envisioned in the documentation for the template.Also, I see that it is currently being ised by theSerena Williams, Barack Obama, Special Court for Sierra Leone, Taylor Swift, Public opinion of same-sex marriage in the United States, 2011–2012 Egyptian revolution, Arab Spring, Suasoria, and Controversia articles. I see taht a discussion here was closed with a result of keep. Based  on this, I've reverted the change as a good faith edit. Wtmitchell  (talk) (earlier Boracay Bill) 23:20, 27 October 2012 (UTC)
 * It changes the formatting (specifically it parenthesizes the publisher rather than separating it unparenthesized from the work by a period) and so violates WP:CITEVAR. And it loses important information: compare "276. ^ El Comercio." (from the quick variant) to "276. ^ (in es)El Comercio. http://www.elcomercio.com/politica/Julian-Assange-embajada-Ecuador-Londres_0_721727907.html" from the original. I think it's not yet ready for prime-time. Re-undoing. —David Eppstein (talk) 00:27, 28 October 2012 (UTC)
 * Taking that specific example:


 * It appears that with, the "work" has been used as if it were the article title. If the title parameter, omitted in both cases, had been used, a more consistent appearance is produced:
 * It appears that with, the "work" has been used as if it were the article title. If the title parameter, omitted in both cases, had been used, a more consistent appearance is produced:
 * It appears that with, the "work" has been used as if it were the article title. If the title parameter, omitted in both cases, had been used, a more consistent appearance is produced:
 * It appears that with, the "work" has been used as if it were the article title. If the title parameter, omitted in both cases, had been used, a more consistent appearance is produced:
 * It appears that with, the "work" has been used as if it were the article title. If the title parameter, omitted in both cases, had been used, a more consistent appearance is produced:


 * But whether title is used or not, there are two items of information lost by use of : the WP:COINS metadata and the language indicator. As an aside, the documentation for shows that language is not intended for a language code such as "es" ("Use the full language name; do not use icons or templates"). -- Red rose64 (talk) 12:46, 28 October 2012 (UTC)
 * The language parameter applies to all of the templates, as well as citation. In the given example, the issue is not with cite quick per se, as the abbreviation was used with the previous use of cite news; there was also no title, which is why the original and updated versions are odd. Indeed, cite quick does not support language.
 * I agree with the above: the quick templates must demonstrate that they render in the same manner as the standard templates. The concept is good, but the implementation is not yet there. ---— Gadget850 (Ed)  talk 13:12, 28 October 2012 (UTC)
 * I wasn't trying to suggest that es wasn't being recognised because it's a code, nor was I implying that it's only recognised by . All my comments refer to and, these being the two templates involved in the double revert concerning ref 276 in that article. The correct usage - whether you're using  or  - would be Spanish - but whether you put Spanish or es, cite quick ignores it and the other two output it verbatim, so you get "(in Spanish)" or "(in es)", respectively. -- Red rose64 (talk) 14:54, 28 October 2012 (UTC)
 * I was responding to the original post. Sorry for the confusion. ---— Gadget850 (Ed)  talk 15:33, 28 October 2012 (UTC)
 * Have the errors been corrected in the template? - as the template is still being implemented in other high profile articles like here at Israel. The explanation for the change leads us to an essay called Wikimedia Foundation error. Moxy (talk) 16:20, 31 October 2012 (UTC)
 * I agree with the above: the quick templates must demonstrate that they render in the same manner as the standard templates. The concept is good, but the implementation is not yet there. ---— Gadget850 (Ed)  talk 13:12, 28 October 2012 (UTC)
 * I wasn't trying to suggest that es wasn't being recognised because it's a code, nor was I implying that it's only recognised by . All my comments refer to and, these being the two templates involved in the double revert concerning ref 276 in that article. The correct usage - whether you're using  or  - would be Spanish - but whether you put Spanish or es, cite quick ignores it and the other two output it verbatim, so you get "(in Spanish)" or "(in es)", respectively. -- Red rose64 (talk) 14:54, 28 October 2012 (UTC)
 * I was responding to the original post. Sorry for the confusion. ---— Gadget850 (Ed)  talk 15:33, 28 October 2012 (UTC)
 * Have the errors been corrected in the template? - as the template is still being implemented in other high profile articles like here at Israel. The explanation for the change leads us to an essay called Wikimedia Foundation error. Moxy (talk) 16:20, 31 October 2012 (UTC)

Removing COINS metadata
FYI: Template talk:Citation/core Kaldari (talk) 00:08, 13 November 2012 (UTC)

Section parameter edit request on 7 December 2012
Please add a section parameter to the template. Newspaper often are divided into sections (sports section, metro section, etc.) and it would be helpful to have a section parameter to optionally include that information in the cite. -- Uzma Gamal (talk) 13:44, 7 December 2012 (UTC)
 * Red information icon with gradient background.svg Not done: It has been suggested before, and it was agreed that the at parameter serves this purpose, as in sec. C, p. 12, col. 4 -- Red rose64 (talk) 15:14, 7 December 2012 (UTC)
 * Found it. Template:Citation discusses parameter at. Below are two cites:
 * The first cite merely says page 18, but that's only true for the Arts & Books section of that newspaper. So, like most newspapers, the above newspaper cite needs the section info. The second cite gives the correct information. On a different note, if you look at:
 * . (This uses 18)
 * . (This cite uses p. 18)
 * They produce the same result. The page parameter merely is one form of the at parameter. Why isn't page listed as an alias of at? This would encourage using the at, which has more versatility. -- Uzma Gamal (talk) 15:44, 7 December 2012 (UTC)
 * at is not an alias for page. As is documented, when at is defined, then page has no effect. Now, cite news supports department, which may be what you want (see the documentation) and can easily be added here. --— Gadget850 (Ed) talk 16:33, 7 December 2012 (UTC)
 * I got that backwards: the order of precedence is page, pages, at. --— Gadget850 (Ed) talk 06:29, 8 December 2012 (UTC)
 * at is not an alias for page. As is documented, when at is defined, then page has no effect. Now, cite news supports department, which may be what you want (see the documentation) and can easily be added here. --— Gadget850 (Ed) talk 16:33, 7 December 2012 (UTC)
 * I got that backwards: the order of precedence is page, pages, at. --— Gadget850 (Ed) talk 06:29, 8 December 2012 (UTC)

department now in citation/sandbox:

--— Gadget850 (Ed) talk 17:55, 8 December 2012 (UTC)


 * It seems to me that section would a more intuitive alias for at. And the prospect of mixed (e.g.) at[section]=Sports and p=4, which fails because 'at' suppresses 'p', gets back to an earlier discussion. Is it time to re-visit the matter of suppression? ~ J. Johnson (JJ) (talk) 22:33, 8 December 2012 (UTC)
 * department is not an alias for at. There is no alias for at. --— Gadget850 (Ed) talk 23:01, 8 December 2012 (UTC)
 * Online newspaper sources show things like "Section: NEWS", so using "Department" to mean "Section" in Wikipedia cites might be confusing to others. When I made my initial request above, I didn't know about the at parameter. I'm surprised that the at parameter is a catchall for so many things: "page (p.) or pages (pp.); section (sec.), column (col.), paragraph (para.); track; hours, minutes and seconds; act, scene, canto, book, part, folio, stanza, back cover, liner notes, indicia, colophon, dust jacket, verse." To work around the above issue, my current, standard newspaper cite template is, where I will replace the "xx" variables with the correct text. Most newspapers use something like "Section: NEWS", but they also will show the page as something like "35A" where the "A" portion means NEWS section. Personally, I do not like leaving the Section text out of the Wikipedia cite because the Section text makes it much easier to find the article in a real/paper newspaper. -- Uzma Gamal (talk) 12:07, 9 December 2012 (UTC)
 * The primary purpose of a citation is to assist others in verifying your edits. Whether you put News section, p. 35 or Section news, p. 35 (or even p. 35 in News section) is of secondary importance, so long as it's a unique description for the place where the supporting text occurs, and that others can understand it. -- Red rose64 (talk) 13:10, 9 December 2012 (UTC)
 * Yes, I agree. But as some periodicals paginate within a section, an editor trying to use Section X with p might be surprised, and even a bit discouraged, to see that doesn't work.  (P.s.: I did not mean to imply that 'section' is an alias of 'at', but that it might be useful if it was.) ~ J. Johnson (JJ) (talk) 22:28, 9 December 2012 (UTC)
 * section was already used as an alias of chapter in cite manual and is now supported as such in cite book now that manual is merged. --— Gadget850 (Ed) talk 12:07, 10 December 2012 (UTC)
 * With the perspective of the primary purpose - assist others in verifying your edits - that clears everything up. I wasn't looking at the big picture. I'm really impressed at the level of thought that has gone into this template. Keep up the great work everyone. -- Uzma Gamal (talk) 12:05, 11 December 2012 (UTC)


 * I think of "section" as more generally a subdivision of "chapter" (though possible counter-examples come to mind). Well, what's done is done; I'm not going to worry about it. ~ J. Johnson (JJ) (talk) 23:01, 11 December 2012 (UTC)

Retrieved parameter
I'm liking this template more than the news, web, journal, book templates. I just posted Atrium Award. What is the reason that the retrieved parameter shows up in some of the cites and not others? -- Uzma Gamal (talk) 17:53, 9 December 2012 (UTC)
 * ??? There is no retrieved parameter. -- Red rose64 (talk) 20:19, 9 December 2012 (UTC)
 * It's accessdate. It's used in all five citations in that article but for some reason is visible only in two of them. —David Eppstein (talk) 20:32, 9 December 2012 (UTC)
 * Ah, in which case, the answer is simple: only two of the citations have a non-blank url, so those are the only ones for which accessdate is meaningful. -- Red rose64 (talk) 21:27, 9 December 2012 (UTC)
 * See Template:Citation and Template:Citation. --— Gadget850 (Ed) talk 22:13, 9 December 2012 (UTC)
 * Sorry about that. I was thinking of "retrieved" when I subsequently posted this thread. Regarding the request, I always thought accessdate meant the date when I accessed the item, similar to archivedate, which is the date the item was archived, and format, which is the format of the work/item. Renaming accessdate to be urlaccessdate would make the purpose of the parameter clearer. Now I'm wondering why accessdate has been listed (in the cite news, web, book template) as one of the few basic parameters. So long as the URL works, does it really matter when the URL was accessed? -- Uzma Gamal (talk) 14:26, 10 December 2012 (UTC)
 * Yes, because the content of the web page pointed to by the URL may change; this is particularly the case with recent news stories where new developments may cause the page content to alter. It's been discussed many times here and on other talk pages, but comes down to this: regarding the origin of your data, you need to provide sufficient information so that others may verify that what you have written agrees with the source. -- Red rose64 (talk) 15:46, 10 December 2012 (UTC)
 * Chicago advises including an access date only if a publication date cannot be determined. Discussions have gone both ways— some to include it only for content subject to change and others to always include it. The documentation shows how to hide the access date. --— Gadget850 (Ed) talk 22:07, 10 December 2012 (UTC)
 * Thanks. Now I see the value in the accessdate parameter. Perhaps some explanation can be added to the documentation as well, something like: "accessdate - Date when the url was accessed ( see WP:V) in case the content of the web page pointed to by the URL changes." On a different note, has the issue of renaming accessdate to urlaccessdate been addressed before as well? -- Uzma Gamal (talk) 11:58, 11 December 2012 (UTC)

Authorlink parameter
Please add a "yes" condition option to the authorlink parameter that, if authorlink yes, the authorlink wikilink takes on the parameters First and Last (and somehow avoids linking to yes). A "yes" condition option will help save time in preparing a citation template and reduce errors in the First and Last parameters in some cases. -- Uzma Gamal (talk) 14:19, 7 December 2012 (UTC)
 * Red information icon with gradient background.svg Not done: please establish a consensus for this alteration before using the template. This will also require changes to . -- Red rose64 (talk) 15:14, 7 December 2012 (UTC)
 * Regarding the concept of yes, it seems like a fair addition. From a technical standpoint, I'm now fairly sure that we would only need to modify, since does nothing with the parameter - it's merely passed through. This means that the feature would be simultaneously added to the  templates such as , , etc. so there would be no issues of inconsistency.
 * One consideration which could be a "killer" is that the speed of the various citation templates has attracted so much adverse comment that we rarely add new non-essential features. There are now at least five schools of thought: (i) get rid of all citation templates and use hand-crafted references; (ii) write an entirely fresh set of citation templates that are purported to be faster (by up to 10x in some claims); (iii) wait for Lua and then rewrite all the existing citation templates using that; (iv) strip out all non-essential features; (v) do nothing (especially, do not add any new features), and put up with the speed as it currently stands. -- Red rose64 (talk) 17:58, 7 December 2012 (UTC)


 * Huh? Perhaps I am missing something obvious, but it seems to me that authorlink specifies a wikilink, which is to say, some value in article namespace. As "first" and "last" have no meaning in that namespace, it is quite unclear to me what this request is supposed to do. ~ J. Johnson (JJ) (talk) 22:47, 8 December 2012 (UTC)
 * Are you suggesting that you would like to generate King, Stephen instead of King, Stephen?  GoingBatty (talk) 22:55, 8 December 2012 (UTC)
 * That is how I understood Uzma Gamal's request. -- Red rose64 (talk) 23:36, 8 December 2012 (UTC)
 * Overhead would be very low, as the template already does a check for authorlink. I am wondering how many valid links can be made vs. links that have to be disambiguated. The example of  rendering as King, Stephen is perfectly valid, but this won't work for James E. West (Scouting). --—  Gadget850 (Ed) talk 00:34, 9 December 2012 (UTC)

My request is  would generate King, Stephen. If implemented, would generate a red link for King, Stephen, giving a clue to the editor in that, either something is wrong with the first and last parameter or that Wikipedia has no article for Stesrphwen Stesrphwen. When I wrote the recent article Ford assassination attempt in Sacramento, I used 50 references. Collectively, those 50 references took a lot time themselves to write. For each one, I needed to get the person's name correct. Then, to determine whether I can add a name to authorlink, I have to stop everything I'm doing, copy the newspaper author's name, post it in the Wikipedia search box, then look at some of the results and try to figure out whether Wikipedia has an article on that newspaper person. (I was using the cite newspaper, book, periodical, web templates for that Ford assassination article, but have finally decided to comit to using Citation template. They all have authorlink, but I'm now using Citation template in a new article I'm writing and posted my suggestion on this talk page). If there were an authorlink=yes option, I could have checked all fifty authorlink= parameters all at once (lessening the amount of time my writing an article is disrupted and saving me a lot of time that I could used for my writing). Then, I would need to have only either removing the "yes" from the authorlink or revising authorlink to read, for example, "James E. West (Scouting)" when needed. In any event, if an editor used, User:DPL bot would (?) post on the user talk page a notice that the James West link is to a DAB page. -- Uzma Gamal (talk)


 * My understanding of this is that you want either automatic creation of a suitable authorlink, or at least a list of candidate links. I think the first is unlikely, as there is no way to automatically resolve ambiguities.  As to the second, it seems you have a mind an automatic search and presentation of candidate links, with a single click sufficing to create a filled-in authorlink. I don't know how that would work, and am rather dubious about its utility. ~ J. Johnson (JJ) (talk) 22:41, 9 December 2012 (UTC)
 * I'm not sure where you got the list of candidate part. At some point in the cite process, an editor has to manually determine the suitability of the authorlink. Authorlink=yes would create an authorlink, where the suitable of the First, Last, authorlink parameters could be manually determined afterwords by right clicking on the linked author name. J. Johnson, how do you now go about manually determine the suitability of the authorlinks that you use? -- Uzma Gamal (talk) 15:13, 10 December 2012 (UTC)


 * Sorry, I'm having difficulty understanding what you envision. It seems you want something like "auto-completion".  But that would more properly be a function in the editor, not in the citation template. ~ J. Johnson (JJ) (talk) 23:25, 11 December 2012 (UTC)
 * I think the original editor's request is that the link be automatically created from first/last, in the hope that this is most often the right link. However if this links to a disambiguation page, or happens to link to the wrong individual, editors will then manually fix them. I'm suspicious that the automatic linking will not always get manually checked, so we'll have more linked authors and more incorrect links. Rjwilmsi  08:14, 12 December 2012 (UTC)
 * Rjwilmsi, you probably are right. Editors may leave it for others to change the "yes" in yes for incorrect links, so we would end up with more incorrect linked authors then the current authorlink method. Also, I just realized that probably no more than 10% newspaper authors have a Wikipedia article, so it's better to keep things the way they are than to have a yes option that presumes the author has a Wikipedia page. -- Uzma Gamal (talk) 10:18, 12 December 2012 (UTC)

Web-link flag for Author-link
I understand that  is not intended for and does not work for external links. It is meant to be used for wikilinks. There have been several discussions on this talk page's archives here, here, here, and here about trying to use  for external links.

The following crafty workaround (that I found here) seems to work in circumstances when an author of a source does not have a wikipedia article and the editor who included the citation wanted to include an external web page that references the author.

The workaround is clearly a hack and the / /  parts of the Citation template were not intended to be used that way. However, it's clear from the past discussions on this talk page's archives, and my (relatively little) experience, that it would be useful to be able to make  work for external links.

My solution for this is to add a set of  flags to the Citation template which could be used for each. If the  flag is set to , then   should be treated as an external link. Additional weblinks could be added for additional authorlinks: for example,  ,  ,    , etc. This modification would not require changing pre-existing citations and would create new functionality for future edits.

Please advise if my request here is sufficient to build consensus for this change or if we must open a discussion on Templates for discussion. Thanks! -   &#x0288;  u coxn \ talk 04:46, 16 December 2012 (UTC)
 * I don't see the need. The purpose of the citation is to provide a way to verify the content with enough information for an editor to find the source. Links are a convenience. Lets take the long view: what happens to that site when the author is dead ten years? And now that I look, your example author Patricia Monaghan has an article, and recently died. The article has information on the author, including the website link. If there is no article, and the author is notable, then create an article. --— Gadget850 (Ed) talk 11:04, 16 December 2012 (UTC)
 * When we re-enable the COinS metadata, constructs like  as suggested above will seriously screw things up. Use each parameter for its designated purpose, and no other, please. -- Red rose64 (talk) 15:12, 16 December 2012 (UTC)
 * Gadget850, regarding the long view and what happens to a site when the author is dead in ten years... use or switch to an internet archive version.
 * Also, what happens if the author is not notable, or not notable enough for a wikipedia article. How do you link to that author to make it easier to find the source?  Take, for example, Michel Lafon.  Please don't ask, "which one?"  I'm not sure... there are several writers with the same name, many of whom are not notable.  However, the name is referenced in many wikipedia articles, making it even more useful to add a web link to the author, in order to verify the source.
 * Redrose64, I agree with you that each parameter should be used for its designated purpose, and no other. Also, it makes sense that constructs like the above would "seriously screw things up," when COinS metadata is re-enabled.  Those are both reasons why   should be implemented. -    &#x0288;  u coxn \ talk 10:29, 17 December 2012 (UTC)

Editor and translator fields for historical sources
The editor field does not appear to function appropriately for editions of historical (and literary) texts. In such publications a single work is published, but with two creators. One is the author (who lived in the past and had no direct connection with the modern publication) and the editor or translator, who is considered the creator of the modern published work – often because of the interpretative commentaries and notes added to the published edition or translation. The appropriate citation format differs, depending on the information of the title page of the published edition.

When the author's name appears as part of the title (e.g., Bede: The Reckoning of Time), the Chicago style manual recommends that the editor or translator's name should appear in the place of the author (e.g. Wallis, Faith, tr.), which can be done for editions (but not for translations) using the present editor field.

Things become stickier when the author and editor (or translator) both appear on the title page: Then the recommended citation format would be: Isidore of Seville, Etymologiarum sive originum, ed. W. M. Lindsay,... The present editor field seems to be designed for editions which are collected works, with editor(s) and title for the collected work and with different authors and titles for each chapter or section. It apparently can't deal with scholarly editions of historical works where a work with a single title has both an author and an editor.

A similar translator field is also needed to deal with the same situation for translated works. Can something be done to emend the editor code to deal with scholarly editions of this sort and to extend or reuse the editor code to deal with scholarly translations. --SteveMcCluskey (talk) 19:16, 12 December 2012 (UTC)
 * Use others:
 * --— Gadget850 (Ed) talk 19:34, 12 December 2012 (UTC)
 * Thanks, that's a good way to deal with an editor where the author appears with the editor on the title page. It doesn't deal with the case where the author does not appear on the title page and the editor or translator is to be listed first.  Do you have any suggestions there?  --SteveMcCluskey (talk) 18:39, 18 December 2012 (UTC)
 * Example? --— Gadget850 (Ed) talk 19:02, 18 December 2012 (UTC)
 * Thanks, that's a good way to deal with an editor where the author appears with the editor on the title page. It doesn't deal with the case where the author does not appear on the title page and the editor or translator is to be listed first.  Do you have any suggestions there?  --SteveMcCluskey (talk) 18:39, 18 December 2012 (UTC)
 * Example? --— Gadget850 (Ed) talk 19:02, 18 December 2012 (UTC)

Agency - when available, I use cite news
This has come up in a couple of archived discussion threads. Currently I avoid using 'citation' if an agency is included as part of the byline (such as 'jj rhythm, staff writer for Associated Press') as there is _apparently_ no support for the rather complicated concept of agency in the general citation template. In the archived discussions, some people mentioned articles which have no by-line but are associated with an agency; I personally avoid these as citations because it is not clear whether the agency association is related to redistribution or composition by anonymous staff. --User:Ceyockey ( talk to me ) 01:13, 24 December 2012 (UTC)
 * You should not be using cite news, and other WP:Citation Style 1 templates, in an article formatted using citation. The formats differ in subtle ways that make them incompatible. —David Eppstein (talk) 01:58, 24 December 2012 (UTC)
 * When using citation, you can include the agency in series. This will render in exactly the same manner as cite news, as it feeds into the Series meta-parameter. --— Gadget850 (Ed) talk 08:41, 24 December 2012 (UTC)

Translated title
Would it possible to port trans_title over?
 * ―Lgfcd (talk) 14:28, 15 January 2013 (UTC)
 * This template supports trans_title as documented, if that is what you are asking. --— Gadget850 (Ed) talk 15:04, 15 January 2013 (UTC)

Template:Infobox person
You guys did such a good job with Template:Citation in covering cite book, web, news, etc. Any chance you would be willing to expand your effort to Template:Infobox person to get that template to cover other Category:People infobox templates? -- Uzma Gamal (talk) 14:11, 23 January 2013 (UTC)

ERIC No.
ERIC # is "a unique accession number assigned to each record in the Education Resources Information Center database; also referred to as ERIC Document Number (ED Number) and ERIC Journal Number (EJ Number)." Does the citation template accept ERIC Numbers, such as EJ986098, to link to a URL, such as http://www.eric.ed.gov/ERICWebPortal/detail?accno=EJ986098? Looks like it also can be referred to as "Accession No." Thanks. - Uzma Gamal (talk) 03:24, 7 January 2013 (UTC)
 * No. The convention is to first create a stand-alone template, such as arxiv, then use that in the id field. If the use becomes great, then we can integrate it into citation/identifier and update the various templates. So, start by creating ERIC. --— Gadget850 (Ed) talk 11:12, 7 January 2013 (UTC)
 * There seem to be over 600 already-extant urls of that form on the en WP, so it would seem to make sense to implement the template, in case the ERIC query syntax should ever change. LeadSongDog come howl!  20:32, 7 January 2013 (UTC)
 * Agree. I will look at this when I get some time. There seem to be at least two methodes of linking. --— Gadget850 (Ed) talk 13:32, 8 January 2013 (UTC)
 * Started.  → . --—  Gadget850 (Ed) talk 12:55, 24 January 2013 (UTC)
 * Template:ERIC looks good. Thanks for creating it. -- Uzma Gamal (talk) 08:57, 25 January 2013 (UTC)

Online Computer Library Cente – redirect?
Nitpick: WP:NOTBROKEN notwithstanding (I think a different standard applies to heavily used templates), wouldn't it be a good idea to avoid the REDIRECT in Template:Citation/identifier for Online Computer Library Center to OCLC? In other words, change  to. -- Michael Bednarek (talk) 13:02, 27 January 2013 (UTC)
 * ✅ I didn't see any issues. --— Gadget850 (Ed) talk 14:33, 27 January 2013 (UTC)

ISBN
Can we remove the link labelled "ISBN" that goes to ISBN per this discussion? Thanks. -- Alan Liefting (talk - contribs) 00:04, 23 February 2013 (UTC)
 * Red information icon with gradient background.svg Not done: please establish a consensus for this alteration before using the template. This is a wide-ranging change and would not just affect  but also all the  templates, because it's in . It's been discussed before, and rejected. Most recent was Template talk:Citation/core/Archive 16 but there have been several others. -- Red rose64 (talk) 07:49, 23 February 2013 (UTC)
 * All of the identifiers link to the Wikipedia article on the identifier; see the table in the documentation for Citation/identifier. --— Gadget850 (Ed) talk 10:18, 23 February 2013 (UTC)


 * I am seeing a consensus at Talk:International_Standard_Book_Number, if not a really good reason for it. -- Alan Liefting (talk - contribs) 19:33, 23 February 2013 (UTC)
 * I'm not. What I see there is an open RFC. Until that RFC closes consensus cannot be determined; even if it's snowballing (one way or the other), it should be closed before any actions are taken which depend upon its outcome. -- Red rose64 (talk) 20:44, 23 February 2013 (UTC)

Semantic HTML
Instead of, this template should use   to format titles of publications, along with proper support in the MediaWiki:Common.css style sheet (i.e., restoring the default italic rendering of the cite element).

For bonus points, each article/section title should be in a cite element too, but in roman text with quotation marks, perhaps as.

Comments? —Michael Z. 2013-02-26 23:03 z 
 * Further reading from W3C: 9.2.1 Phrase elements: ...  ... in HTML 4.01; and 4.6.6 The   element in HTML5. Note the change in purpose. -- Red rose64 (talk) 23:16, 26 February 2013 (UTC)
 * I had considered this. We would need a fallback for IE8 and other older browsers that don't support HTML5. We also support and . And  for RTL text and  for Asian text. We need to consider if there a real use for this. --—  Gadget850 (Ed) talk 00:47, 27 February 2013 (UTC)


 * No fallback is needed. The new definition of  only concerns semantics, and it is a tightening of the old meaning. Any use that conforms to HTML5 also conforms to the looser requirements of HTML 4. My emphasis:


 * Old: “a citation or a reference to other sources”
 * New: “the title of a work [. . .] that is being quoted or referenced in detail (i.e. a citation), or it can just be a work that is mentioned in passing”


 * The only thing all browsers do with  is render it in italics, which hasn’t changed since IE 5.5. —Michael Z. 2013-02-27 18:18 z 
 * You are correct in that is supported by older browsers and only the intent has changed. It still needs CSS; for example ' title ' does not show in italics. --—  Gadget850 (Ed) talk 18:42, 27 February 2013 (UTC)


 * That’s because MediaWiki:Common.css turns off the italics. We used to use  to surround an entire citation in some templates. Don’t know if that’s still in use.

/* Default styling for titles of works, styling for the title of an article within a periodical, or a contribution within a compilation. */ cite, .citation cite.article, .citation cite.contribution { font-style: inherit; } /* Styling for the title of any work within a citation, or specifically the title of a periodical. */ .citation cite, .citation cite.periodical { font-style: italic; }


 * —Michael Z. 2013-02-27 21:29 z 
 * Aha. I think I updated the few template uses, but never used it. Yes, some editors used  to enclose the entire citation, which is probably why the CSS is as it is. --—  Gadget850 (Ed) talk 21:46, 27 February 2013 (UTC)


 * Ah. That practice would be obsolete now. I see that the .citation class is used for that now. I will request that the default styling for be reset to italic. The incorrect italicizing should help to make the incorrect uses evident. —Michael Z. 2013-02-27 22:03 z 


 * See MediaWiki talk:Common.css. —Michael Z. 2013-02-27 22:16 z 
 * The  element is used in some of the quote boxes, e.g. Template:Quote. I still see it occasionally used to enclose a citation template; I fix these up when discovered, . To aid detection of these, I've put  into my Special:MyPage/common.css. -- Red rose64 (talk) 00:24, 28 February 2013 (UTC)

trans_title not working
I added a trans_title parameter to the second Bagina reference in the references section of the Pentagon tiling article but only the title shows. The citation documentation suggests that the template does support this parameter. Is this a bug in the template or documentation? Jason Quinn (talk) 04:40, 30 January 2013 (UTC)
 * Works just by changing from citation to cite journal, so a template bug. LeadSongDog come howl!  05:07, 30 January 2013 (UTC)


 * When you set a periodical parameter (work, journal, newspaper, magazine) then the title is in quotes like a chapter, thus you have to use trans_chapter.


 * Cite journal is set up a bit differently, so it always works, but it doesn't support chapters:


 * In this article, there is a mix of and  templates which should be avoided. --—  Gadget850 (Ed) talk 07:10, 30 January 2013 (UTC)


 * Thanks for the replies. I saw that there was a mix of cite style too and noted it ought to be fix before I got sidetracked by this issue. I usually prefer "cite" templates over "citation" templates myself. I had tried "cite journal" and noticed that trans_title would work with it. But the more pressing issue become the question of if the citation template was working than the proper formatting of just an individual reference. I think the design of the citation template is off if trans_chapter should be used by the translation of the title of a work in certain cases. Using trans_chapter seems like a viable workaround but I think indicates a larger bug with the template design. To the extent possible, the parameters ought to be used exactly for what they call themselves and nothing more. If the design is good, at the very least, the trans_chapter is a bad name for the parameter and something more general sounding ought to be used. Jason Quinn (talk) 16:15, 30 January 2013 (UTC)
 * This is why the templates were created: citation has a number of quirks like this that the CS1 templates essentially work around. --—  Gadget850 (Ed) talk 23:45, 30 January 2013 (UTC)

As a side issue: "should"? I think we recommend that editors not mix templates so that they don't get tangled in the details, but there is no problem mixing them provided the details are properly handled — right? I mention this because I recall instances where editors have insisted that uniformity was required, which is hardly the case. ~ J. Johnson (JJ) (talk) 20:46, 18 March 2013 (UTC)


 * Per Citing sources: "While citations should aim to provide the information listed above, Wikipedia does not have a single house style, though citations within any given article should follow a consistent style." The CS1 templates and citation use different default name separators and terminal punctuation, and do some different layout depending on the parameters used. Thus, the same citation using either style usually looks different. --— Gadget850 (Ed) talk 20:55, 18 March 2013 (UTC)

Cross-reference needed
I wanted to add a footnote citing a TV source, and assumed that this would be one of the innumerable aspects of the "cite" template. I was wrong, of course &mdash; by googling Wikipedia I found that separate templates "cite AV media" and "cite episode" exist, and for all I know there are still others for other media.

Could someone please add a "See also" section at the end of the "cite" documentation linking to these templates and any others that the main "cite" template does not cover? Please be sure to include descriptive words like "film" and "television", since those are what people in my position are likely to do text searches on.

Thanks. --50.100.193.51 (talk) 17:37, 18 March 2013 (UTC)
 * See and the navbox at the end of each CS1 template. Citation is a similar template but uses different styling and should not be mixed with CS1. --—  Gadget850 (Ed) talk 17:52, 18 March 2013 (UTC)

Both pages range and particular page number?
When both from–to and  number are specified, could the result please include "from–to, number" instead of omitting the range? It's not a huge deal, but it makes it easier ordering from some libraries with off-site journal storage, figuring out how much money will be needed for copies, and it looks a lot more professional. 75.166.220.33 (talk) 00:50, 28 March 2013 (UTC)
 * Just put all the page numbers in 'pages' per the documentation:


 * --— Gadget850 (Ed) talk 16:07, 28 March 2013 (UTC)

Pages, page
citation currently has 'pages' override 'page', whereas the templates has 'page' override 'pages'. Seems to me like this is a mistake that should be corrected. --  Gadget850 (Ed) talk 17:59, 1 April 2013 (UTC)


 * Yes. Only (and noting the previous section) I wonder if use of both page= and pages= ought to be flagged as an error. Or at least as a warning that something improper is being attempted. ~ J. Johnson (JJ) (talk) 22:32, 1 April 2013 (UTC)


 * and citation ought to take the same action. Another possibility for what that action might be would be outputting the "pages" field name if is not empty and then, if  is not empty, outputting that content with an appropriately placed comma separator either before or after the  content. Wtmitchell  (talk) (earlier Boracay Bill) 01:25, 2 April 2013 (UTC)
 * Category:Pages with citations using conflicting page specifications. The use of both 'page' and 'pages' is usually when 'pages' is used to indicate the number of pages in the source, which is not an element of CS1 citations. --  Gadget850 (Ed) talk 02:30, 2 April 2013 (UTC)
 * Not all cases are when pages has the total - sometimes, I think it's used for the upper end of a range (page being the lower end), as . -- Red rose64 (talk) 19:37, 2 April 2013 (UTC)
 * 5000+ articles. That's going to take a while to clean up. —David Eppstein (talk) 03:26, 2 April 2013 (UTC)
 * Cleanup will be needed, but I thought I'd asked in CS1 templates for pages to override page, and this had been done, which was what I thought was the pre-Lua behaviour? Rjwilmsi  06:15, 2 April 2013 (UTC)
 * For the old and Lua CS1 templates, the order is 'page', 'pages', 'at'. For some reason citation currently uses 'pages', 'page', 'at'. --  Gadget850 (Ed) talk 09:35, 2 April 2013 (UTC)

Missing title?
The following citation: complains about the missing title, but the title is actually supplied (and displayed too). GregorB (talk) 07:32, 20 April 2013 (UTC)
 * The keyword chapterurl is specified but the chapter title is not given. The link points to a book rather than a chapter - thus you need to use the url=. Aa77zz (talk) 07:42, 20 April 2013 (UTC)
 * Thanks - makes sense, even if the error message is a bit misleading. In this particular case chapterurl should be fixed as suggested, but it could have pointed to an actual chapter, so in that case chapter param would be missing rather than title. GregorB (talk) 07:56, 20 April 2013 (UTC)
 * Yes, the real error is "chapterurl without chapter"; the error message is misleading and needs to be fixed. Peter coxhead (talk) 07:58, 20 April 2013 (UTC)
 * Okay, I'm taking it to Module talk:Citation/CS1. GregorB (talk) 08:10, 20 April 2013 (UTC)

Turning off generation of anchors
I'm using Ucucha's HarvErrors script which, unsurprisingly, gives red error messages next to the references in the Further reading section of articles: Harv error: There is no link pointing to this citation.

I find the script very useful but the messages are unsightly. I would like to turn off the generation of anchors for these references. The documentation states that "If an empty |ref= is given, no anchor is generated;". I find that I need to include quotes like this: |ref="" to switch off the generation of anchors (as reported by the Ucucha's script). Perhaps the documentation should be more specific as what is meant by an empty |ref=. Aa77zz (talk) 08:34, 20 April 2013 (UTC)
 * This is a change with the Lua version. I have reported the issue. --  Gadget850 (Ed) talk 08:58, 20 April 2013 (UTC)

Access date
The documentation for this template includes this: "accessdate: Date[n 1] when the url was accessed." That's clear enough. So why does it output "Retrieved on [date]"? If it's an access date we're talking about, wouldn't it make more sense for it to output "Accessed on [date]"? As is recommended by the MHRA Style Guide, for example? Justlettersandnumbers (talk) 23:29, 14 April 2013 (UTC)
 * Because that is how the creators designed it. --  Gadget850 (Ed) talk 23:40, 14 April 2013 (UTC)
 * Accessed and retrieved mean exactly the same thing. All that matters is that usage be consistent within any one article. -- Alarics (talk) 05:50, 15 April 2013 (UTC)
 * Right. So how do I get the template to consistently output "Accessed" instead of "Retrieved", as recommended by the style guide that I personally like to follow, the MHRA Style Guide? Justlettersandnumbers (talk) 00:00, 4 May 2013 (UTC)
 * You don't. You use the style already established in the article. See WP:CITEVAR. --  Gadget850talk 01:14, 4 May 2013 (UTC)
 * OK, I'll phrase that differently; where do I find the series of citation templates that output "Accessed" for the access date, for use in those articles where that style is already established? Justlettersandnumbers (talk) 13:07, 4 May 2013 (UTC)
 * I am not aware of any such templates. You will have to create them. --  Gadget850talk 14:20, 4 May 2013 (UTC)

Missing or empty |title=
In some articles, we have started getting errors of the following type where the citation template is used without the title parameter in the form of a short footnote. This technique is used so that quotes can be included. Can someone take a look at it please?

In:.

Betty Logan (talk) 10:57, 27 April 2013 (UTC)
 * This is an undocumented and unanticipated method of use. Where is this used? --  Gadget850talk 11:25, 27 April 2013 (UTC)
 * This particular example comes from the List of highest-grossing films, but this error has only appeared today I think. I edited the article yesterday and I didn't notice it. It is sometimes used when the Sfn system is in place and editors want to add a quote. Betty Logan (talk) 12:01, 27 April 2013 (UTC)
 * The error check has been in place for a while, but caching may hide it until the page is edited. Does this method have a name? --  Gadget850talk 12:37, 27 April 2013 (UTC)
 * Not to my knowledge, it's basically a hack. It problem isn't a major issue anyway since the book in question has bits written by authors but without an internal title which is quite unusual. It just caught my eye, but I can sort it out at least on this article by writing them out manually. Betty Logan (talk) 12:46, 27 April 2013 (UTC)
 * OK. Use none. This feature was just added. --  Gadget850talk 14:54, 27 April 2013 (UTC)

Lua problem?
I see a problem where citation no longer accepts "More than one of |location= and |place=". This seems to be a Lua change. Is this the best place to discuss this? ~ J. Johnson (JJ) (talk) 23:10, 3 May 2013 (UTC)
 * The template never supported multiple parameters like that. The new version detects it as an error. Example? --   Gadget850talk 23:29, 3 May 2013 (UTC)
 * The location and place parameters are aliases for each other; if both are specified, only place is displayed. Previously, if location was specified in addition, it would be silently ignored; but since the Lua upgrade, we display the message " More than one of  and   specified (help) ". Choose either location or place but don't use both in the same.
 * It might be that location has been misused and that the at parameter should have been used instead. -- Red rose64 (talk) 07:21, 4 May 2013 (UTC)
 * I see. This has come up where I specified the 'location' of a conference where a paper was presented and the 'place' where it was published. Apparently I had not noticed that one parameter was silently ignored. I am inclined to just drop 'place'. Any suggestions as to a better way? ~ J. Johnson (JJ) (talk) 21:46, 6 May 2013 (UTC)
 * The way these sorts of things are often formatted in MathSciNet is (in terms of our equivalent parameters) to have Conference name (Location, date) and then to have the date and location parameters refer to the publisher's address and the actual date of publication (which is often a different year from the conference date). For example:
 * In this example the conference location and publisher location happen to both be "Cambridge", but that's coincidental. —David Eppstein (talk) 21:57, 6 May 2013 (UTC)
 * Now that I understand your intent (examples really help), then I can point you to 'publication-place'.
 * Now that I understand your intent (examples really help), then I can point you to 'publication-place'.


 * If any one of 'publication-place', 'place' or 'location' are defined, then the location shows after the title; if 'publication-place' and 'place' or 'location' are defined, then 'place' or 'location' are shown before the title prefixed with "written at" and 'publication-place' is shown after the title.
 * And it looks like it got lost in the documentation. I will fix that in a moment. --  Gadget850talk 22:24, 6 May 2013 (UTC)
 * I fixed the doc. This is pretty clean in core, but each template has this swap the fields around kludge. --  Gadget850talk 22:40, 6 May 2013 (UTC)
 * Of course "written at" and "presented at" are two different things...for most references to conference papers, saying the paper was written at that location is not right. —David Eppstein (talk) 22:41, 6 May 2013 (UTC)

Chapter-url and archiveurl
Hi. If I've used chapter-url but also want to indicate the archive URL in a citation, is there a parameter for this? (I didn't see one mentioned in the documentation.) I tried using archiveurl, but this parameter works only with url and generates an error message if it is missing and only chapter-url is present. — SMUconlaw (talk) 10:00, 10 May 2013 (UTC)

HTML special characters are not unescaped before being encoded for COinS
I hope this is the right place for this. To illustrate what I mean, see the page field for citation at http://en.wikipedia.org/w/index.php?title=Wendel%27s_theorem&oldid=529017245 The value for page is entered as, which, while unnecessarily escaped, is not incorrect (I think. It's certainly rendered correctly in the references). The COinS that is rendered for the page field, however, looks like, which is not right. Looks to me like the right way to handle this would be to first unescape HTML special characters and only then URL encode the string. That should result in correct encoding. Aurimas (talk) 03:00, 25 May 2013 (UTC)

Chapterurl and Archiveurl
Is there some way we can get these two to work in harmony? See reference 19 here for what I'm describing; it's showing up right, but for some reason it's still giving angry red text. Both urls are showing up exactly where they're supposed to, but the template is still giving an error message. Anyone have some idea of what's going on? The Blade of the Northern Lights ( 話して下さい ) 02:45, 30 May 2013 (UTC)

Short citation - op cit
Sometimes I wish to cite the same article with multiple cite manual or cite web tags, using a different quote or sectionurl for each. Is there a way to provide a single tag with full publication details and to link to it for each quotation or section, but getting standard formatting of, e.g., pages, section? I saw sfn, but it doesn't seem to do the formatting. Shmuel (Seymour J.) Metz Username:Chatul (talk) 14:22, 22 May 2013 (UTC)
 * sfn, which is used in, has a 'loc' parameter. --  Gadget850talk 17:04, 22 May 2013 (UTC)
 * However, as I noted, it is missing other parameters. Also, seems to entail an extraneous wikilink. Shmuel (Seymour J.) Metz Username:Chatul (talk) 03:25, 23 May 2013 (UTC)
 * The wikilink in, like those in the other -related templates, takes you to the full citation. Go to NBR 224 and 420 Classes and click any of the bluelinks; the browser moves to the relevant full citation which should also gain a pale blue background. That's if you have Firefox, Chrome, Safari or Opera; some versions of Internet Explorer don't support the  pseudo-class (probably because it's CSS 3 but not CSS 2.1) so the background doesn't change colour although the movement should still be made. -- Red rose64 (talk) 09:14, 23 May 2013 (UTC)
 * You will have to give us an example of what you need. And what is an "extraneous wikilink"? --  Gadget850talk 10:47, 23 May 2013 (UTC)


 * I'm not sure from the responses above whether the answer to User:Chatul's original query has been made clear; apologies if it has. As I understand it, he wants to provide a different url for each page/section/chapter, etc., which can be done by linking the value of loc, e.g.  . (A common mistake I make is to put "|" here as in an  Wikilink.) This works for a single page if you want the "p." to be included in the link text, which looks better to me, e.g.  . If the "cite" templates are being used, rather than the "citation" template, you must add harv. Peter coxhead (talk) 11:01, 23 May 2013 (UTC)


 * What I'm looking for is a means to get the formatting the does and link to an earlier cite manual for the details on the full publication. I'd like something like the following to make SC24-6073-03 in the second citation a wikilink to the first citation.




 * Shmuel (Seymour J.) Metz Username:Chatul (talk) 19:21, 23 May 2013 (UTC)
 * Shmuel (Seymour J.) Metz Username:Chatul (talk) 19:21, 23 May 2013 (UTC)


 * I updated your example. In the first, set harv to create the anchor. In the second, set #|SC24-6073-03 to create the link. --  Gadget850talk 19:45, 23 May 2013 (UTC)


 * Now that what User:Chatul wants is clearly explained, there are several ways of achieving it (I had a different one, but Gadget850's is neater). Whether you should link through an arbitrary component of a citation is another question, and the second citation above just looks as though fields such as the author and date are missing. Why not use the more standard style:


 * , "When you do not specify either the RECOMP or LABEL option, the disk area is initialized by writing a device-dependent number of records (containing binary zeros) on each track. Any previous data on the disk is erased."


 * Peter coxhead (talk) 20:03, 23 May 2013 (UTC)


 * It's almost universal in the IBM mainframe world to refer to manuals by form code and the use of Harvard notation is unheard of. Historically that was true of several other hardware vendors as well. Further, IBM typically publishes hundreds or thousands of manuals in a given year, many on the same day; the date can not serve as a unique identifier. Shmuel (Seymour J.) Metz Username:Chatul (talk) 04:18, 4 June 2013 (UTC)
 * I am quite familiar with IBM documentation. "the use of Harvard notation is unheard of" Not sure what you mean, as Harvard is well used here. "the date can not serve as a unique identifier" If you check the sfn documentation, you will see that the standard fix is to add an alpha suffix.


 * What I mean is that all o0f the IBM shops and IBM publications that I am familiar with refer to IBM manuals by form code, title or both; I have never seen them use Harvard notation to refer to them. Which is not to say that they don't use Harvard notation to refer to other publications. Shmuel (Seymour J.) Metz Username:Chatul (talk) 18:48, 4 June 2013 (UTC)


 * As best I see, you want to cite the whole book in one place by book title and the chapter by chapter title in another place. Is there a style guide that illustrates this? --  Gadget850talk 11:08, 4 June 2013 (UTC)


 * I'm not aware of such a style guide, but I'm wondering whether a similar issue exists for Army regulations, which I've always seen referred to by number rather than author and date. Likewise BCP, RFC and STD numbers for IETF documents. Shmuel (Seymour J.) Metz Username:Chatul (talk) 18:48, 4 June 2013 (UTC)
 * And I am quite familiar with Army regulations and manuals. It is immaterial how IBM creates citations, as we have our own styles. --  Gadget850talk 00:10, 5 June 2013 (UTC)


 * The key is the CITEREF. In harv (I believe sfn is similar) you are not limited to author(s) and year, you can use your own key (CITEREF), provided you supply the same key in the citation template. ~ J. Johnson (JJ) (talk) 23:54, 4 June 2013 (UTC)


 * But, as Gadget850 rightly points out, the issue is not what effects you can technically achieve, but what style should be used in Wikipedia. We should not use different styles for IBM manuals, army manuals, etc. Peter coxhead (talk) 06:21, 5 June 2013 (UTC)


 * IBM turns out lots of documents (the Army even more so), and if your "style" limits you to strictly author-year then I'd say you're screwed. The purpose of a short citation (such as "Smith 2000") is to identify the full citation, and if "IBM SC24-6073-0" works then use it.  Or as Ed suggested, use the year and a suffix.  (I 2000a, 2000b, 2000c, etc., rather opaque, but suit yourself.) These are hardly differences in style, only accommodation to the exigencies of different kinds of sources.  The point is to not let over-concern with petty consistency get in the way of writing good articles. ~ J. Johnson (JJ) (talk) 18:29, 5 June 2013 (UTC)
 * There's nothing "petty" about consistency in Wikipedia; it's a very serious subject, witness the reams of discussion on MOS talk pages! Peter coxhead (talk) 20:13, 5 June 2013 (UTC)


 * You misstate my words. I did not say anything like "consistency in Wikipedia is petty"; I said: to "not let over-concern with petty consistency get in the way of writing good articles." If your concept of "style" is so rigid that you can't show some flexibility then I think it will get in the way of writing.  ~ J. Johnson (JJ) (talk) 20:47, 5 June 2013 (UTC)
 * Oh, I'm generally an advocate of flexibility in styles; indeed I've had long debates with those who support the MOS dictating a single style where I think that flexibility is better. If I've mis-understood your use of "petty", then I apologize; we can agree that "over-concern with consistency should not get in the way of writing good articles". Peter coxhead (talk) 22:04, 5 June 2013 (UTC)
 * No problem then. Getting back to the point of interest, where I suggested that if the "author-date" (though more typically "authorlastname-year") is inadequate as a short-citation then it is reasonable to use something like "organization-document#". This certainly is not like the conventional "Smith 2004", but I would argue that it conforms to the basic idea of a short-cite pointing to a full citation.  ~ J. Johnson (JJ) (talk) 20:32, 6 June 2013 (UTC)
 * There are examples of working satisfactorily without an author at NBR 224 and 420 Classes - several instances of SLS 1970, and one instance of Gradient Profiles 2003 - in the first case I used the initials of the publisher; in the second I used a shortened form of the book's title. Both were achieved by means of ref in the  - the  got no special treatment. -- Red rose64 (talk) 23:53, 6 June 2013 (UTC)
 * Which work, and look good. Nice work. ~ J. Johnson (JJ) (talk) 00:51, 7 June 2013 (UTC)

Misplaced language in journal citation
The citation yields (or, expanded out in case the template changes): Note the "(in Latin)" between the journal name and journal volume number. It makes no sense for the language to be there; it should be after the article title. —David Eppstein (talk) 06:54, 27 June 2013 (UTC)
 * Euler, L. (1736), "Solutio problematis ad geometriam situs pertinentis", Commentarii Academiae Scientiarum Petropolitanae (in Latin) 8: 128–140.

type param repeats information
Is it a bug that "Audio interview" is rendered twice in the output from i.e. 03:55, 3 July 2013 (UTC)
 * [problem no longer visible since the template has been fixed now -- I guess for the record I should have subst-ed it to preserve the issue] -- EEng (talk) 20:42, 21 July 2013 (UTC)


 * I'm getting the same thing where type="Television broadcast". I'd rather not use "cite interview", it would be better to use, but I can't find a way to fit an interview subject in there. Dementia13 (talk) 17:39, 21 July 2013 (UTC)

--  Gadget850talk 17:55, 21 July 2013 (UTC)


 * Gadget, you've pissed me off in the past, but I do appreciate your taking care of this. EEng (talk) 20:42, 21 July 2013 (UTC)

subjectlink parameters do not work
Although the documentation for Template:Cite interview include various subjectlink parameters, the template code appears to only support authorlink parameters. Should the template be fixed or should the documentation be changed? Thanks! GoingBatty (talk) 01:35, 4 July 2013 (UTC)

--  Gadget850talk 17:55, 21 July 2013 (UTC)
 * Thank you for fixing the template! GoingBatty (talk) 00:57, 22 July 2013 (UTC)

Template data
Template data was recently added. The problem is that this template has two modes: patent and citation. Patent mode is switched by recognizing the inventor parameters. With a fully populated template data, editors are going to be presented with both sets of parameters, with resulting confusion. --  Gadget850talk 20:59, 1 August 2013 (UTC)
 * I'm still not sure its got all the parameters, it was based on the last non-lua version. Template data is not smart enough to cope with overloaded templates. Personally I think the best option is to split the patents into a new template, there are only 2000 transclusions so it would not be too hard to change.--Salix (talk): 22:02, 4 August 2013 (UTC)
 * The deprecated parameters were picked up, and the parameters from the module were not. --  Gadget850talk 23:31, 4 August 2013 (UTC)
 * Where can we find what the current parameters are? There is a list in Module:Citation/CS1/Whitelist but I'm not sure how accurate this is, or how specific it is for Citation.--Salix (talk): 19:10, 7 August 2013 (UTC)

Memento
I just discovered Memento, a standard that makes it possible to send an HTTP request with an associate timestamp, asking for a version of that URL around that time. I see that there is an extension that would allow MediaWiki to process memento links; and there was an RfC on en:wp supporting running a pilot.

How can this be implemented within the citation template? – SJ  +  22:24, 15 August 2013 (UTC)
 * How about the existing archiveurl archivedate parameters? -- Red rose64 (talk) 12:56, 16 August 2013 (UTC)
 * Memento is not installed, so there is nothing we can do here. --  Gadget850talk 10:30, 28 August 2013 (UTC)

I created the document Memento Capabilities for Wikipedia that describes areas in which the Memento protocol could be leveraged to add end-user value to Wikipedia. One of the described capabilities relates to the link rot problem and to expressing the archivedate and archiveurl parameters in a machine-actionable manner in HTML so that Memento-capable clients can use that information to access archived version of cited resources available in web archives. I will propose a Wikiproject that uses the document as a starting point. Hvdsomp (talk) 22:06, 19 September 2013 (UTC)

Lastname, Firstname
Why do we display the author as "Lastname, Firstname"? This only seems useful if you are going to sort the references by the last name. Why don't they display as "Firstname Lastname" which is the way you pronounce them. --Richard Arthur Norton (1958- ) (talk) 23:11, 10 January 2014 (UTC)
 * I guess one answer is that this is a recognized citation style (along with many others) and Wikipedia allows any valid citation style. I personally strongly prefer this style. Peter coxhead (talk) 18:51, 11 January 2014 (UTC)
 * I feel that names should only be in the "last name, first name" format in bibliographies and other lists, and not in footnotes where they should be in the usual "first name[,] last name" format. If you want names to appear in the latter format, use the author parameter instead of last and first. — SMUconlaw (talk) 19:15, 11 January 2014 (UTC)
 * It doesn't matter which, as long as it is consistent within any one article. -- Alarics (talk) 20:36, 11 January 2014 (UTC)


 * In text ("prose") names should be in "normal" order, which in English (and most European usage, but not other cultures) is "first+last". But for bibliographic purposes the "lastname" (more precisely, the surname, or family name) is the accepted key for identifying (and finding) an author, so it comes first in citations. As to being useful only if a list is sorted: well, bibliographic lists should be sorted. But even if not — especially if not — it is easier to find an author if the surname comes first.
 * For isolated "full" citations in footnotes (i.e., not in a list) there is not such a problem of sorting or finding, so "normal" order is okay. But lacking some kind of toggle to tell the citation to not invert, we should not misuse the author parameter to achieve a desired result. It should be restricted to instances of authors with single names, or group authorship, where the name is not composed of "personal name" and "family name". Where first and last name are stuffed into author it is invariably poor usage. Which is useful in a way: it can be a warning of shoddy editing. ~ J. Johnson (JJ) (talk) 22:47, 11 January 2014 (UTC)
 * There isn't any guideline on the citation page or elsewhere stating that author should not be used in the manner I described, so I don't see why such a usage should be deprecated. Neither do I see any need for such a guideline. — SMUconlaw (talk) 07:06, 12 January 2014 (UTC)
 * The metadata produced by the citation templates is more complete if the author parameter is divided into last and first names, and so is more useful to people using Wikipedia pages as an automated source of references. Value to readers is more important than the preferences of editors, so although editors aren't to use the last and first parameters, they should accept other editors changing author to last and first (as I regularly do). Peter coxhead (talk) 10:32, 12 January 2014 (UTC)
 * Perhaps I am terribly ignorant about how this works, but can you explain how Wikipedia pages are used "as an automated source of references", and how this information (metadata?) is extracted from citation templates? — SMUconlaw (talk) 16:29, 12 January 2014 (UTC)
 * A very good question there from SMUconlaw, and one to which I too await the answer with interest. I should also like to know: is this metadata business of any relevance to the general run of ordinary Wikipedia readers? -- Alarics (talk) 21:52, 12 January 2014 (UTC)
 * If you visit any Wikipedia page that uses or the  templates, view the HTML source for the page and search through that for the place where the expanded form of the template is given, you should find that it's enclosed in a  element, where   is given as one of the classes. That span contains two main sections: the first is the reference text as plain text (portions of this, such as links, being within simple HTML markup); the second consists of another, with   and a   which begins  . That span element is the COinS metadata. -- Red rose64 (talk) 22:02, 12 January 2014 (UTC)


 * "Metadata" is sometimes used in the context of a source, referring to informaton about the source. However, in this context — in a citation template — where we are contemplating usages such as "first=Joe" and "last=Smith", the metadata is the information about the type of each datum.  E.g., the text string "Joe" is a "first" (name) type of datum, and "Smith" is a "last" name.  (If the work is self-titled we could also have "title=Joe Smith".)
 * I don't know that this is of direct interest to the general reader (perhaps Peter can comment on that?), but it certainly is for editors. E.g., there are many names where the surname (family name) is compound, and lacking familiarity with a possibly foreign language and culture it can quite difficult to parse the correct terms to use for sorting and searching. Where someone has taken the effort to label such data there is greater accuracy and reduction of possible error, as well as greater confidence that someone has done careful, accurate work. ~ J. Johnson (JJ) (talk) 22:53, 12 January 2014 (UTC)
 * But that surely isn't going to work with e.g. Cinese or Korean names, where the family name comes first (without a comma). -- Alarics (talk) 08:07, 13 January 2014 (UTC)
 * It's even more important to know which is the family and which the personal name in such cases. The order of names in citations is determined by the style adopted (see e.g. 14.2c here) not by the practice in the person's home country. Peter coxhead (talk) 10:14, 13 January 2014 (UTC)
 * In that case you are going to end up with a comma that shouldn't be there, for example, "Lee, Kwan Yew" instead of Lee Kwan Yew. This looks silly and simply wrong, whatever Harvard might claim. -- Alarics (talk) 11:36, 14 January 2014 (UTC)
 * Silly, perhaps, but not wrong. Or not as grievously wrong as getting the names in the wrong order. This seems quite tolerable. ~ J. Johnson (JJ) (talk) 22:32, 14 January 2014 (UTC)
 * I disagree, the comma there is completely wrong. This seems to be a case of trying to bend the reality to fit the system, rather than the other way round. Same with "Ban Ki-Moon". Nobody calls him "Ban, Ki-Moon" and it looks absurd. The problem is instantly solved if we are allowed to put "|author= Ban Ki-Moon". Which is what I always do anyway and will continue to do. -- Alarics (talk) 22:59, 14 January 2014 (UTC)

In addition to better identifying the data, using separate last and first parameters is essential for creating correct internal links using the harv templates. Mixing harv and author does not work well. —David Eppstein (talk) 04:54, 13 January 2014 (UTC)
 * Exactly. Harvard style citations are always based on surname (family name). So it's essential to know which part of a name is which. Peter coxhead (talk) 10:14, 13 January 2014 (UTC)
 * In that case, I support the original poster's suggestion that there should be some parameter that can be added to citation to instruct it to display names in a "FIRSTNAME LASTNAME" format where required. — SMUconlaw (talk) 11:16, 14 January 2014 (UTC)
 * (To David Eppstein, Peter coxhead, SMUconlaw) That can be dealt with in the Citation and CS1 templates via ref, where author and year match the parameters used in the associated harv family templates. Wtmitchell (talk) (earlier Boracay Bill) 16:25, 14 January 2014 (UTC)
 * Sorry, I don't really understand. Can you explain? I don't use harv at all, only citation. — SMUconlaw (talk) 19:17, 14 January 2014 (UTC)
 * See Parenthetical referencing. Basically if you have a citation template somewhere (in one footnote, or in a separate references section) and a harv, harvtxt, sfn etc template elsewhere (in another footnote, in the text of the article, etc), then the harv template will create a short reference like, with a special wikilink that, when you click it, causes your browser to highlight the full citation. The last name is used to create the wikilink address, so if you use author instead of last the link won't work unless you go through some extra complications involving ref. —David Eppstein (talk) 20:52, 14 January 2014 (UTC)
 * Oh, quite right, and double-plus important, about the automagical CITEREF generation failing if last (or its synonym, surname) is not used. I am stunned I forgot that!! ~ J. Johnson (JJ) (talk) 22:37, 14 January 2014 (UTC)
 * Er, it doesn't fail: if last and surname are absent, but author is present, that is used to construct the harv link. What it can't do is to separate out the surname to make that link, so if you have a with e.g. Doe, John 2014 this makes a link like  and so the harv templates need to be constructed as  - unless you are willing to give the  a custom ref parameter of course. -- Red rose64 (talk) 22:59, 14 January 2014 (UTC)
 * Of course. But if author is used with first+last the citeref is misformed, and does not match what Harv expects. I consider that a failure. ~ J. Johnson (JJ) (talk) 23:59, 14 January 2014 (UTC)
 * I didn't say that you had to use a comma; I also stated "unless you are willing to give the a custom ref parameter". Here's how:  mentioned Ban Ki-moon, and that's a good example to use. If we were using one of his works as a ref source, we would put Ban Ki-moon into the ; if that also has e.g. 2014 this makes a link like  and so in the absence of any customisation, the harv templates need to be constructed as . But if you want to use harv templates formed like, then in the  you would put ref and this makes a link like  -- Red rose64 (talk) 08:05, 15 January 2014 (UTC)

Jumping back in with an explanatory example; removing indent for example wikitext reflist: [...]Some article content about Singapore. Some more content about Singapore. [...] [...]

References

Bibliography (end of example -- edit to examine wikitext) Wtmitchell (talk) (earlier Boracay Bill) 08:39, 15 January 2014 (UTC)
 * Yes, this is a working demonstration of my Ban Ki-moon suggestion. -- Red rose64 (talk) 14:36, 15 January 2014 (UTC)

It seems to me that two different issues are being mixed up here.
 * 1) Whether it's ever better to use author rather than last and first.
 * 2) How names should be displayed in citations when the norm for the country of origin is family name first.
 * It's a very bad idea to allow (2) to influence (1). Separate parameters preserve maximum information and allow kind of formatting to be used.
 * Since we don't have a single prescribed style for citation in the English Wikipedia, then editors should be free to use any reasonable and consistent style within an article. The solution for those who want "Ban Ki-moon" alongside "Smith, John" is to add an additional parameter for each last + first pair to suppress the comma, to require author to achieve this effect. Peter coxhead (talk) 16:57, 15 January 2014 (UTC)
 * Do you seriously think people will (a) know they are supposed to do that, and (b) know HOW to do that? -- Alarics (talk) 18:58, 15 January 2014 (UTC)
 * It doesn't matter if some people don't know something. On a wiki the idea is that those that do know will improve on the work done by those who don't know. As long as we have the information available from last and first, the template can be re-coded to display that in whatever form is required; author has less information and hard codes the first and last names into a fixed format, losing flexibility. Wherever possible we should preserve last and first as that contains more information than author. If there is sufficient support for the OP's request for the template to display "Firstname Lastname" in a citation, then it could be made available by coding a switch into the citation templates. Although I have to say that purely cosmetic changes like that wouldn't have a high priority for my time, especially as I'm used to seeing "Lastname, Firstname" in citations in external sources that I read. YMMV. --RexxS (talk) 20:34, 15 January 2014 (UTC)
 * Peter: your first issue needs to be phrased better. There are instances where author is better, because the name of the "author" does not resolve into firstname/lastname (e.g.: "IPCC"). The specific issue initially presented here is the use of author to force "Firstname Lastname", with the disadvantages just mentioned by RexxS. And which I think should be forbidden. ~ J. Johnson (JJ) (talk) 22:27, 15 January 2014 (UTC)

Similar
There's a similar data granularity issue, still unresolved, under discussion at Template talk:Infobox book. Your contributions and suggestions for fixes would be appreciated there. Andy Mabbett ( Pigsonthewing ); Talk to Andy; Andy's edits 12:11, 16 January 2014 (UTC)

Ebook In-source locations
Kindle and Calibre ebook readers use "location" or "loc" in place of page numbers and in future a growing number of references will be from ebooks where contributors don't have a paper version. I don't know about other devices/apps but I suspect that they do the same. Would it be possible to have an additional template identifier per the above that would address the issue or is there a better way to tackle this? ► Philg88 ◄ star.png 05:43, 9 February 2014 (UTC)
 * The at parameter is provided for cases such as this. -- Red rose64 (talk) 15:37, 9 February 2014 (UTC)

Broken PMID, PMC and DOI identifiers. How to identify and then fix?
Due to an error in the coding of an editing tool, some citation and cite journal references appear to have been added to en.wikipedia between 6 July 2013 and 20 March 2014 with the following malformed styles of PMID, PMC and DOI data:

I have manually fixed these three occurrences to:

What mechanisms are available to identify the remainder of these and similar errors scattered throughout an unknown number of articles?

Is there an automated process that can be invoked to find and fix these? -- 79.67.241.255 (talk) 02:31, 21 March 2014 (UTC)
 * I've fixed a couple of those, too. If I remember correctly, they are included in Category:Pages with citations using unsupported parameters. Perhaps a subcategory could be added for these, for a bot to fix? —PC-XT+ 17:18, 21 March 2014 (UTC)
 * I misunderstood. These are not flagged as errors, and I don't remember fixing any of them, but a bot still could. —PC-XT+ 17:23, 21 March 2014 (UTC) (My original response was for code such as, but I don't think a tool adds these. I move Template:Doi into doi when it is near a link I convert to cite web, but I don't remember working with one in id. I'll be on the lookout, though.) —PC-XT+ 19:17, 21 March 2014 (UTC)
 * See Bots/Requests for approval/DOI bot 2 which does just that. Since that bot approval has become .  LeadSongDog  come howl!  20:15, 21 March 2014 (UTC)
 * Here's some examples showing and some . -- 79.67.241.255 (talk) 23:09, 21 March 2014 (UTC)
 * The Yutu examples are deprecated, but should still work. They are of the second form listed above, which is the original way of doing identifiers, still valid.  See the history of cite journal/doc from five or six years ago... LeadSongDog  come howl!  01:02, 22 March 2014 (UTC)

The coauthors attribute is deprecated
Although the coauthors attribute is deprecated, a large number of editors continue to add this parameter to new and existing citations. A number of commonly used editing tools also appear to allow or encourage this usage.

Is it time to start alerting editors when they attempt to add this parameter and to perhaps suggest the lastn firstn syntax be used instead?

I have no preference whether an alert appears as the edit is made or a note is added to the user's Talk page, or appears in some other way via some other mechanism. -- 109.176.236.32 (talk) 17:39, 11 March 2014 (UTC)


 * Yes! While I can conceive of one possibly valid use of coauthors, I have yet to see an instance of that. Invariably it is used to paste in a string of author names, which the editor often has not taken the trouble to cleanup and parse. I am inclined towards making it an editing alert.
 * If "commonly used editing tools ... encourage this usage", perhaps we should try to identify them, and see if they can be revised. ~ J. Johnson (JJ) (talk) 21:57, 11 March 2014 (UTC)
 * REFLINKS is one of them -- 109.176.227.77 (talk) 18:17, 13 March 2014 (UTC)
 * I would support revising the tools, and then providing an alert. —PC-XT+ 17:47, 21 March 2014 (UTC)
 * As of 20 March 2014 the MakeRef tool is compliant. -- 79.67.241.255 (talk) 18:21, 21 March 2014 (UTC)

This is a case where forcing to use the lastn|firstn syntax is not good for the bulk of casual editors, and I would think that deprecation of the parameter should come in the form of hidden-category flagging of these. It takes a lot longer to add lastn|firstn syntax than it does to add coauthors with a lot of values bundled. My thought is that this should be a cleanup target and not a proscription against use of coauthors. Further, the conversion of coauthors to lastn|firstn should be among those specifically flagged items which would not violate the 'keep references in the _consensus_ format' principle (which I've butted up against more than once). Finally, might be able to add some guideline about how to use coauthors to support bot-assisted dissection as another alternative to full deprecation; for instance, if the format added is "xxx, yyy; zzz, qqq", then this could be parsed by a bot to lastn|firstn format; likewise, if the format added is "yyy xxx, qqq zzz", this could be parsed by a bot to authorn format ... in principle and with the caveat that they should probably be segregated to a hidden-cat flagging them as bot-attempted-parsing-of-coauthors until well established heuristics could handle variations on the themes. --User:Ceyockey ( talk to me ) 00:13, 25 March 2014 (UTC)
 * Variations of name, with prefixes such as Dr. and Sir, extra words in the middle such as von and de la, plus suffixes such as III and Jr make the job of a bot that much harder. Like many things in life, this is another case where doing things properly initially takes longer. The benefit is that the result is immediately usable. Perhaps some simple tools could be developed? Editor enters a long string of names, selects whether these are authors or editors, and the tool returns the code to paste into the template. The editor can and should check the data before pasting it in. -- 79.67.243.69 (talk) 13:30, 25 March 2014 (UTC)

Periodicals and the publisher param
In Template:Citation does Not normally used for periodicals. mean (1) that periodicals usually go in the "work" parameter instead of the "publisher" parameter, or (2) that when citing a periodical, the extra information of the periodical's publisher (and its related "publisher" parameter) should be omitted altogether. Original discussion here. czar ♔  12:20, 11 March 2014 (UTC)
 * The name of the periodical goes in work journal newspaper, whichever is most appropriate. The publisher parameter is never used for the name of any work, whether it's a periodical or not - it's for the name of the person or company that is legally responsible - the entity that produces or issues the work concerned. -- Red rose64 (talk) 17:11, 11 March 2014 (UTC)
 * Right, so is there anything wrong with using both, e.g., All Things Digital Dow Jones & Company? (line 113) czar  ♔  22:16, 11 March 2014 (UTC)
 * I would say that it depends primarily upon two things - whether there is one All Things Digital or several; and if it is unique, how well known it is. -- Red rose64 (talk) 23:36, 11 March 2014 (UTC)
 * The publisher should also be given if there is a connection with the subject; in your example, for instance, if the article is about a Dow Jones subsidiary or employee; or a rival service. Andy Mabbett ( Pigsonthewing ); Talk to Andy; Andy's edits 10:14, 1 April 2014 (UTC)

Read online but no URL
In this edit User:Dennis Bratland decided he needed to mention that he read an online version of an article, but didn't (or couldn't) provide a URL.

I sympathize; I also have access to the Gale databases. If you just copy the URL while viewing the article and then go to a different computer and try to view it, you will be interrupted to enter your name and userid. But the library I use doesn't tell me my name and userid, instead, I go to Gale from a special link in the state library website. So giving a URL is pretty much useless.

So, should the online database be mentioned at all (when the full information about the paper copy can be gleaned from the online copy)? If so, what parameter(s) should be used. Jc3s5h (talk) 16:40, 1 April 2014 (UTC)

retrieved vs. Retrieved
Accessdates show as "retrieved" (lower case) when using citation templates and as "Retrieved" (upper case) when using cite templates. Is this an accident or by design? -- 86.146.133.157 (talk) 18:49, 7 March 2014 (UTC)
 * By design. Compare
 * Notice that separates the various items with commas, whilst  uses a period. A capital letter follows a period; a comma is followed by a small letter (unless it's a name or the title of the work). BTW  is a redirect to  so I have assumed that by cite you actually mean  etc., which are collectively known as the  templates. -- Red rose64 (talk) 18:57, 7 March 2014 (UTC)
 * Thanks for the explanation. I missed those subtle details before. -- 86.173.249.198 (talk) 15:57, 15 March 2014 (UTC)
 * So why commas in one case and periods in another? Andy Mabbett ( Pigsonthewing ); Talk to Andy; Andy's edits 10:10, 1 April 2014 (UTC)
 * That's one of the primary differences between WP:CS1 and WP:CS2. -- Red rose64 (talk) 17:42, 1 April 2014 (UTC)
 * So why commas in one case and periods in another? Andy Mabbett ( Pigsonthewing ); Talk to Andy; Andy's edits 10:10, 1 April 2014 (UTC)
 * That's one of the primary differences between WP:CS1 and WP:CS2. -- Red rose64 (talk) 17:42, 1 April 2014 (UTC)

author-mask
This is what the "⸻" symbol is for. Can we just use this rather than allowing editors to specify an arbitrary number of dashes? ⇔ ChristTrekker 18:33, 3 March 2014 (UTC)
 * No, thanks; the symbol you have used is not present on my fairly standard Windows XP/ Google Chrome setup, and renders as an empty square. (I would be content with the "arbitrary number" of dashes being hard-coded instead to one or two. Andy Mabbett ( Pigsonthewing ); Talk to Andy; Andy's edits 18:53, 3 March 2014 (UTC)
 * For me it shows as a square containing the cryptic characters 2E3B. Googling around, I find that it's supposed to be a U+2E3B Three-Em Dash. -- Red rose64 (talk) 20:08, 3 March 2014 (UTC)
 * My "fairly standard" (though dated) Firefox doesn't implement the Unicode 3-em dash either. But I don't why the "author-mask" should be an arbitary number of unspecified dashes; the printing convention is that it is (as named) three "em" dashes. And for fixed-width typewriter fonts I  seem to recall four hyphens used as the equivalent. ~ J. Johnson (JJ) (talk) 22:53, 3 March 2014 (UTC)
 * It's a function of the available typefaces more than the browser, but I understand what you're saying. The character only became standardized in the past few years. Most fonts in a default OS install are several years older than that, and thus wouldn't have it available. I keep hoping the situation will change. ⇔ ChristTrekker 21:43, 31 March 2014 (UTC)
 * I see no need to keep the parameter to define the length of the dash. Three "ems" seems long to me, but debating that would be a case of WP:BIKESHED. Andy Mabbett ( Pigsonthewing ); Talk to Andy; Andy's edits 20:22, 1 April 2014 (UTC)

Accessdate
I've noticed an epidemic of accessdates in the last few months, and even quite a few cases where "date" has been removed and replaced with "accessdate," even on items like newspaper stories. Is there some automated tool that is doing this? I assume "accessdate" is still only used for items with no date or that are subject to change? Kendall-K1 (talk) 23:37, 24 March 2014 (UTC)
 * I hope it is not an automated process run amok but rather some unfortunate editing. I tend to add access date to web site entries where either a) it is subject to change or b) it cannot be archived (e.g. robots.txt blockade).  --User:Ceyockey ( talk to me ) 00:02, 25 March 2014 (UTC)
 * If the source has a publication date - not necessarily a full date but maybe a month & year or just a year - then this should not be removed from the citation under any circumstances. Access dates are supplementary to, not a replacement for, publication dates. Publication dates are particularly important for news sources. If you see people removing these, ask them why. -- Red rose64 (talk) 07:30, 25 March 2014 (UTC)
 * The Reflinks tool adds an accessdate to every reference it touches, even where the tool is populating the publication date information at the same time. My understanding is the accessdate should be specified where the publication date isn't specified and cannot be determined. -- 79.67.241.76 (talk) 15:15, 28 March 2014 (UTC)
 * Not necessarily. The accessdate applies to the url and is an indication of when the material was available on the web. Stuff which is expected to be fixed in content and permanently available online (e.g. it has a DOI) doesn't need an accessdate, but other references might well do – some sources only have recent editions online, others might post corrections. Peter coxhead (talk) 16:37, 28 March 2014 (UTC)
 * That is one consideration, yes. However, active URLs are not a prerequisite for many sources such as newspapers and books -- as long as the editor has provided sufficient information to find the source in the physical world (e.g. page number and oclc id, for which isbn and issn are indirect surrogates).  For purely online content, there is much more of a need to include accessdate _if_ there is no archived version.  If an archived version cannot be made, as in the case where robots.txt blocks indexing by Internet Archive, then accessdate is pretty essential.  These days I will almost always check to see if a source I am touching is represented or can be represented in Internet archive and indicate the result in an editorial note as part of the citation (hidden note bracketed by  ), typically following the hidden parameter deadurl=no .  I might be a bit odd in doing this, but I think it is helpful to editors while not unnecessarily extending the visible citation footprint. --User:Ceyockey ( talk to me ) 00:37, 2 May 2014 (UTC)

Need for documentation
I'm finding it problematic to piece together a rational approach to overhauling citations/references in some articles. The basics aren't so bad, but there are always the more unusual cases. There is a long list of the complete set of parameters to this template, most of which lack any description whatever. Many are obvious synonyms or replications for multiple authors, editors, etc. Some are just obscure in meaning. For example, what is "type" for? It could be almost anything. How about "format"? Is there a useful list of values it can take? Then, is "publication-place" a synonym for "place"? All these would be nice to know with regards to advance expectations of what footnote text will be generated. If you're editing just an article section, you can't use preview to get a look before saving. Throw us ignorants a bone, please. A short entry into the description field of the parameter table would be a big help.

Ah, one more thing. Put a notation of some sort there also to indicate "deprecated" to indicate parameters you'd like to dispense with in the future. That way, editors will be discouraged from creating even more uses of it to be hunted down and eliminated later. Evensteven (talk) 20:11, 5 May 2014 (UTC)
 * All of the things you request are in the documentation for the template. Scroll down past the table to Template:Citation for details on how each parameter should be used. – Jonesey95 (talk) 21:12, 5 May 2014 (UTC)
 * If there are any undocumented parameters, have a look at or  - most parameters have the same meanings. -- Red rose64 (talk) 22:13, 5 May 2014 (UTC)
 * Thanks, both of you; I appreciate the pointers. Still, the "Description" section does not cover all of them, and the uninitiated will not know to try the other citation templates. I was about to suggest that "cite journal's" common tabular descriptions simply be copied over to "citation". I see many indications in the articles that editors are only using the information they can find fast. I would hazard that most do not want to spend time hunting, but it can leave the articles jumbled. Small holes can matter. It's basically a technical writing issue here, a thing I'm familiar with from my professional activities, and one that's often given too little attention, by engineering departments, and by sales or marketing. But customers notice, and gripe to customer service, which then is ignored by the rest of the company. I'm not trying to lay all that dead weight at this door, but just to point it out for what it's worth. Search time to find info counts for a lot. Again, thanks for the help. Evensteven (talk) 23:16, 5 May 2014 (UTC)


 * Allow me to do the obvious. I'm one of those engineering types.  I suck at documentation and I complain mightily when documentation for a commercial product falls far short of the mark.  All of the CS1 documentation, as you rightly point out falls too short of what we would desire.  All of it needs to be redone to eliminate inconsistencies and contradictions; to make is somewhat readable so that editors will read it.  If you have the skills necessary to help us out of this mire, please do.  I stand ready to assist in whatever ways I can.


 * —Trappist the monk (talk) 23:43, 5 May 2014 (UTC)


 * I sympathize, and also recognize what you say. Please take my current comment as suggesting only the lightest touch to the doc for the one template. Every engineer can recognize that a bug fix does not require a system redesign (usually ;) ).


 * Obvious replies first: "busy", "other priorities", "I'd need to understand it all myself, first" (all of which are true enough), and on to "gee, I'd like to ...", "sure, someday", and finally to the silent "how do I get out of here as fast as I can?" (all of which are natural enough). But I can't say that a doc system redesign is not called for here, or that it wouldn't be fairly high impact for WP. I do have at least some of the skills to help, and some of the willingness. But I can't make an immediate start. The best I can promise is another look when I clear my plate a little (not the whole plate, but enough to give some room and cause for reassessment). I've done both software and tech writing, and understand well how and why engineers hate making docs: it's not their line, for one thing. When (not if) I reconsider my next moves on WP, I'll contact you on your talk page. By then I'll also have collected some more experience with this stuff. Evensteven (talk) 01:16, 6 May 2014 (UTC)
 * undertook a major rewrite of the docs for almost all of the and  templates, beginning in December 2011. The idea was that since most of the parameters functioned similarly, and that most of the underlying code was shared, it made sense to have shared documentation as well. Hence, we now have  and all of its subpages, and the documentation for e.g. the last parameter of  is similar to that for  - both come from . -- Red rose64 (talk) 09:58, 6 May 2014 (UTC)
 * "There is a long list of the complete set of parameters to this template" If you are referring to a specific template, then note that all the CS1 template talk pages redirect here. Please let us know which particular template you are referring to. --  Gadget850talk 11:05, 6 May 2014 (UTC)
 * Er, they don't. Most redir to Help talk:Citation Style 1 ([//en.wikipedia.org/w/index.php?title=Special:WhatLinksHere/Help_talk:Citation_Style_1&hidelinks=1&limit=500 full list]). The list of those that redir here [//en.wikipedia.org/w/index.php?title=Special:WhatLinksHere/Template_talk:Citation&hidelinks=1 is quite short], so I assumed that the OP is referring to . -- Red rose64 (talk) 11:24, 6 May 2014 (UTC)
 * You are correct. --  Gadget850talk 12:13, 6 May 2014 (UTC)
 * Gadget850, "this" template refers to the template whose talk page we are talking on. While documentation "sharing" (modularization) is a good engineering principle, it's more difficult to implement in documentation than it is in software, because documentation is not created dynamically, whereas running a program is naturally dynamic. I'm not saying that's a wrong approach generally, but it does require constant maintenance by the documenter(s). In addition, the end product must place documentation for "X" with "X", not with "the thing whose engineering underlies X". The document modules do not precisely parallel the engineering modules. Finally, documenting each engineering module is not sufficient, because all you have then is a reference manual. That's only good so far as it goes, but you need overview material and "how to" descriptions at a higher level that point you to all the right engineering modules. At that level "engineering module" means "template", not "implementer of template". The documentation that exists is inconsistent as to which engineering level of module is being addressed, and to which audience it is directed to, and levels are mixed within a "document". It's all needed stuff, but by different people at different times. And different levels of expertise are assumed at different points. There's a lot of organizational work still required, and it's not clear just where all the needed material (not all of which exists) should be put (or how it should be shared).


 * Let me illustrate some items from my point of view as an editor who recognizes engineering but knows (essentially) nothing special about the engineering of this stuff. I see various mentions of COinS and "ids" strewn around the documentation. I also see examples that illustrate the citation text produced from a template call (excellent: that's critical), but interspersed with stuff about metadata for whatever type of citation "engines" external to WP (that I care nothing about until I care about them - if I ever do - and I'll be one of a small minority if I ever do). I see that the templates seem to be doing two things at the same time: producing citation text for an article, and storing metadata for some obscure purpose I can't make out and don't care about (which is why I never bother to try to make out what it is). Unless my guess is way off, what I suspect is that the documentation needed by the editors trying to create a citation is mixed in with the documentation needed by template and engineering module builders creating and maintaining the templates themselves. This is a real no-no, and needs fixing.


 * I can see with an engineer's eye that I should use these templates in a way that doesn't create extraneous unneeded metadata that lies around using up resources and providing no benefits. The editor documentation should simply say "generally, don't include parameters that have null values when you use a template" (that is the way to fix it, I guess, yes?). Never a word should be said there about metadata; that information belongs elsewhere. Ids hook up various stuff with wikilinks and external links. Each needs its own description (including which type of link it uses), purpose, how to create, how to put to use, etc. Here's where you engineer the documentation. Each id has a separate doc page or section for itself (within a page devoted to the set of all ids). Each id usable from a template (or the set of ids) is introduced in a one-liner in that template's documentation, which provides a wikilink to the id page or id set page (or section). User access to required information is evident, complete, and immediate without re-creating the same text in multiple places and requiring continual maintenance checks for consistency (which would never be perfect either). Well, I could continue, but this paragraph is too long already, and you have a couple of ideas here. Those are a couple of the fundamental rationales and directions I'd pursue. Evensteven (talk) 19:56, 6 May 2014 (UTC)


 * Hmm, a couple of other things. (1) The rewrites from 2011 are no doubt helpful, but additional writing is needed in the right areas. (2) More critical is the overall organization of material, restructuring. (3) Complementing the restructure is "separation of audience": who needs what for which purpose? (4) Modularization of documentation is much less problematic on WP than in other places because we can use wikilinks to great advantage.


 * Another idea. Has anyone thought of creating a template for assisting with WP-internal documentation? Its sole initial purpose would be the transclusion of material from some page (or documentation source we create) directly onto another documentation page. This would be a substitute or alternative for using wikilinks to structure and modularize documentation. Evensteven (talk) 20:20, 6 May 2014 (UTC)


 * You know, usually when I suggest to someone that their skill set would be a valuable contribution to some particular project, all I get back is the echo of my own voice. Not so this time!


 * If I understand your last question, then I think the answer is sort of. There are a pair of tags,   and   that do direct transclusion.  I used it in Help:CS1 errors in the documentation, such as it is, for the error messages rendered by Module:Citation/CS1.  The parser function   copies whatever text is between the section tags in one page into another.


 * So, for an example: at Help:CS1 errors in §Check date values in: |param1=, |param2=, ... there is help text bounded by  and  .  At,   transcludes the text from Help:CS1 errors.  That puts the help text in both places where I thought it would do the most good and I only have to maintain one copy.


 * —Trappist the monk (talk) 20:46, 6 May 2014 (UTC)
 * Template documentation was originally monolithic, i.e. one page of text. Over time, the documentation for the various templates drifted as the text for one parameter on one page would be changed but not the 20+ other pages. My introduction of citation Style documentation made all the documentation pages consistent. I did some reorganization, but really did not change the content. But take a look at the source of the documentation at Template:Citation/doc. You will see a lot of calls to csdoc which is a redirect to citation Style documentation. Each entry is a call to a subtemplate of citation Style documentation. Take a look at citation Style documentation}— each entry has a  link so you can tweak the existing documentation and it is propagated across all the CS1 templates. There has been some discussion here about updating the documentation. I have been considering a simplified overview of the most used parameters. And I am always open to constructive criticism. --   Gadget850talk 21:04, 6 May 2014 (UTC)

Ok, one thing at a time. Trappist, thanks for the vote of confidence. Yours are the pieces of the underlying implementation of what I was suggesting, exactly. So now, translate the context to me posing as "typical editor", and you get "what's a parser function?". Even I don't really know, although I have a solid idea what it ends up doing. But I still don't know how it does it, how to use it, etc. So what is needed is still a template based on that underlying technology. A template is an interface that "typical editor" is already familiar with, or should come to be familiar with. And in this case, we're not really talking "typical editor", but "typical editor who is interested in contributing to documentation": a "documenting editor". And the documentation might be for the citation templates, but expand the context to the internal documentation for "typical editor" readers. And then let's think a bit bigger to other places in WP where internal documentation is useful, and see if the template can be generalized to support them all. Now you have a template that's a tool for any editor wanting to contribute to creation and maintenance of internal documentation on WP, for any WP reader. But let's be careful not to make it too complex (the "kiss" principle and all that). Focus the design to give it practical limits. (1) Text transcluded in one call ought to go into a single target section and should not contain subsections within itself. (Basically, I'm saying the transcluded material should be "section body" only, with no section-level structuring within it.) This defines the limits of the basic "document unit" or module. "Documenting editors" will then create and maintain these separate document modules and use the template to transclude them wherever they are useful. Many modules will be small. But consider the length of one that lists all the authoring identification options of a citation template! I'm suggesting no section-level structuring within, because different levels of section may be desirable in different transclusions, and that just adds all sorts of unnecessary engineering and usage complication. (2) How about recursively-defined transclusion, and to how many levels? Personally, I see two levels as just about max, if any is really needed. Go too far, and you make the documentation source itself so complicated it becomes nasty to maintain. (3) Other ideas? I'd like to see this documentation template be so simple that it requires at most two parameters, defining markers for the "begin" and "end" of text to be transcluded. Better yet, require those markers to have identical names (except for "begin" and "end") and give the template one parameter. We're talking use by "documenting editors", not engineers themselves. At any rate, start there and see if more power is called for later.

Gadget850, your templates are targeting small items that construct the bits and pieces within what I would call a "documentation module". Perhaps they have been useful to you, but then you wrote and used them yourself. I'm a "documenting editor", and see them, and react with "learning curve repulsion". Once the principle of writing a documentation module once and transcluding it from there is established, any constructs internal to a module that may themselves be repetitive don't need that level technical "support", which rapidly becomes overbearing. As "documentation editor", I would simply go to the same or similar construct in another module, copy the source text, and modify to my current needs. In the end, it's simpler, and it only needs doing the once. And by then it's also another example the next editor can use the same way. A whole module can be used to pattern and adapt a new module where appropriate. This approach can go a long way without need for tools. If any really prove to be necessary, I would suggest a larger set of really tiny specific tools rather than a megalithic tool with a ton of options or parameters. When each specific use is small, so should the tool be small and specific. No sledgehammers to pound a finishing nail. (I hope this feels like constructive criticism; I'm not meaning to pound on you.)

In case it's not entirely clear, I should say that my vision of this thing is that the raw documentation modules themselves are never displayed on their own, at least outside the working area of the documenters themselves. They are strictly raw source material. They're always seen widely only after they are transcluded, in template headers or the like. Documenters might like to have a working area for doing and testing drafts also, like engineering implementers of templates must have. So thought should be given to the structuring of documentation source pages and work areas so everyone's clear on how to proceed safely with the interim work without adversely affecting the end users. Evensteven (talk) 22:35, 6 May 2014 (UTC)


 * A template died abirthing because of .  It is apparently a simple fix but ...


 * —Trappist the monk (talk) 22:58, 6 May 2014 (UTC)


 * Ok. I don't even want to understand the nature of the bug. What I do want to comment on is about the design of Section, strictly from the interface. (0) Poor name, "Section". Nonintuitive. (1) Why so complicated? It's doing two jobs: putting the markers in, and using the markers to transclude. Small tools for small jobs = two tiny templates with minimal interfaces. No parsing of intent or context. No confusion of application. (2) Well, what's left?


 * Template "markDocmodule" |title=titleText. First question: is this even needed? Maybe, but are you sure? Prove it's necessary before introducing complications. Here's an alternate suggestion. Define a documentation module as "a level-2 section on a documentation module source page". The parser can look for a section on a page and just transclude the section content, no section heading or title. No subsections; no full-page transclusions; no special markers. Does that work from an engineering standpoint? Tell me if I'm missing something. Prove it's necessary to transclude subsections, sections, or whole pages. Tell me if I'm missing something. If I'm not, it's engineering overkill. So hopefully this template goes away.
 * Template "transcludeDocmodule" |page=WPpage |section=sectionTitleText. This template transcludes the content of the identified section at the location where it is placed. It does not transclude a section heading or its title. No section structure results from the transclusion. Section definition is strictly left up to the receiving document to define for itself. Tell me what's missing.


 * I really think that document modules as restricted in definition as this provide a huge amount of flexibility in structuring content within their transclusion environments, and a single module's content can be far from small even without subsectioning. If once you have to define section levels within a module, then the module itself becomes restricted to transclusion into places where that leveling is appropriate. The idea is not to have all the documentation you want to transclude in one module, but to put all related modules together on one source page. Then divide it into sections for multiple transclusions at the appropriate destinations. If several destinations have the same or a repeated set of transclusions, that's not really much of a problem to replicate or to maintain, because the need for the shared documentation is relatively unlikely to change much. If that is a problem, it may be that the modules themselves were not properly identified or divided, and should themselves be restructured or redivided. So there's a skill set that should belong to the "documentation editors". Evensteven (talk) 00:09, 7 May 2014 (UTC)


 * Since is non-functioning, no point in discussing it here.


 * Re: The idea is not to have all the documentation you want to transclude in one module, but to put all related modules together on one source page. Isn't that basically what we have now?   is simply explanatory text with transclusions of the various documentation module templates. Each independent CS1 template doc page uses these same documentation modules.


 * —Trappist the monk (talk) 10:15, 7 May 2014 (UTC)


 * Yes, it sounds like the first piece is already in place. But the doc system would also require a functioning template such as I described. And the material on the doc page would need to be formatted into segments that serve as clear and well-defined documentation modules, making the doc page sectionalized if my idea is followed. Then the transclusions of modules to the targets would need to be implemented via template use on the target pages instead of whatever has been done until now. It's that last step that provides a big maintainability benefit, because then a change to the underlying doc page would automatically transclude to all its targets. It's the design of the doc modules and their implementation on the doc page that is the underlying writing work that supports the scheme. (That's why careful definition of what constitutes a "doc module" is so critical.) There's also new writing for additional modules (new material) that's needed, but the page for them exists already.


 * Eventually, it may be that the doc page will be divided into several such pages, each devoted to levels of documentation targeted at different audiences. But that ought to wait until the design needs to become that sophisticated. One step at a time. Right now, which is the audience to focus on? I'd say "typical editor". But then, we do certainly need to consider "template engineer", so we might get two documentation pages out of the one very quickly. Actually, that might be very good for the project development itself, because we'd run into issues like "should this item be on the editor page or the engineer page?", and the answer might sometimes be "both, but written differently for the two audiences". It's this principle of audiences that's fundamental to all tech writing but so completely absent in what there is now. This is an avenue that opens up the issue directly and keeps it front-and-center in the writing where it should be. It's consideration of that technique that begins to answer the problem of providing documentation that delivers what "this reader" needs to do "that job", and without introducing unnecessary learning curves.


 * And we don't want to introduce unnecessary learning curves for the ones providing the documentation either. We have to keep an eye on "documentation editor" too. Evensteven (talk) 19:01, 7 May 2014 (UTC)


 * So I guess I'm suffering from some sort of brain failure. I'm not understanding how what you are proposing is fundamentally different from what we have now.  Template:Citation/doc uses templates to transclude information that is common between this template and the CS1 templates.  Example text and other Template:Citation-specific information unique to  is not templated.


 * Can you illustrate by example perhaps?


 * —Trappist the monk (talk) 22:05, 7 May 2014 (UTC)


 * A refinement of the template "transcludeDocmodule" idea: "transcludeDocmodule" |page=WPpage |section1=section1TitleText |section2=section2TitleText |section3=section3TitleText. Provide up to n modules to be transcluded in the sequence entered, for some reasonable "n" to be decided. It would simplify the target pages by telescoping adjacent transclusions into one template call. Easy to alter also. But definitely a fine point of implementation. Evensteven (talk) 19:43, 7 May 2014 (UTC)
 * Do you have a demonstration? The only way I know to do this is with labeled section transclusion and you cannot use LST in a template, thus the current uselessness of section. --  Gadget850talk 23:43, 7 May 2014 (UTC)

Need for documentation (continuation)
My mistakes. I got too far into implementation details without knowing what exists here already, and now I see that there is a technical problem with LST transclusion I was unaware of too. I know nothing at all about engineering WP templates. So let me drop that path of inquiry into all of your more capable hands, especially as we're on the same basic wavelengths, barring these technical misconceptions of mine. My main point with all of that line was just to be really basic and simple - to the max.

So, to back up a little, into the writing arena again, I'm suggesting basically that the documentation needs some engineering too. Primarily, it needs to be modularized according to a careful definition of "module", and organized onto pages in some way that clearly distinguishes who the target audiences are, separating text for each. For the latter, I suggest that the clearest separation is most easily maintained by having separate pages for each audience, and that doing so forces the writer to consider better who needs what information. I'm sure there's a lot of material that can be used in its present form, some that will need some reworking, and definitely some that is yet to be written. But it's the pages' and modules' jobs to organize the raw doc source clearly. That's the engineering side of the documentation itself.

The rest is mostly writing skills. What exists is almost all low-level reference-type material, some of it for "typical writer", some for "template engineer". There needs to be more "how to" with examples: that's intermediate level. There also needs to be some high-level: introduction to the whole set of citation templates, "what you want to use for this" and "for that", concepts (minimal), what kinds of choices you ("typical writer") are going to have to make, and how to make them ... that kind of stuff.

Example (a tidbit I just picked up): template "citation" separates items with commas (as in most referencing outside WP), while the various "cite xxx" separate them with periods. This calls for a documentation paragraph that tells that, and illustrates with, say, a book citation, identically implemented with "citation" and "cite book", showing the different resulting citation texts. "Use citation for look A, and cite book for look B." Few details are at this high level, just pointers that lead the reader into lower levels of documentation, making it clear where to go for that. This level tends to be topical rather than functional too. "Citing for bibliographies", "citing for footnotes", "parenthetical (Harvard-style) citing", "using short footnotes with separate citations (ref=harv)": topics like that. But these topics are umbrellas, not separate manuals. They don't explain so much as they mention. For example (using the last topic): "when you create the underlying citation, use the ref=harv parameter; then create the short footnote to the citation with the sfn or harvnb template". Now the reader knows what to look up in the detailed low-level documentation. Basically, it's zoom-out/zoom-in, but you need to see something useful at each setting. That's the writing job.

Now you have to go back an reconsider the doc engineering also. How do you go from high- to mid- to low-levels? Where do you transclude those levels to? Basic thought: a separate help page or pages for high- and mid-levels. Low levels pretty much where they are now, attached to each template (reference documentation). Then, a sentence or two of introductory connecting text at each level with a wikilink up or down. "For an overview of citation creation on WP, see here." In other words, pointers/glue.

It's not so much a mystery as it is figuring out what all the pieces ought to be. It's an indeterminately-sized multi-level jigsaw. But you want to design it so that any user can find his desired tic-tac-toe line through the volume quickly. There's the trick. I hope this helps for now. Evensteven (talk) 03:29, 8 May 2014 (UTC)
 * Example? --  Gadget850talk 10:10, 8 May 2014 (UTC)

Template "cite"?
I can use to create a citation from a wikisource that renders as "". But Template:cite redirects to Template:citation, and using "citation" in this way for a wikisource does not work. So, is "cite" actually another template? And regardless what it really is in the engineering underneath, where is its use documented? Thanks. Evensteven (talk) 20:10, 25 May 2014 (UTC)
 * Everything between the {{ and the | is part of the template name. So {{tl|cite Catholic Encyclopedia}} is a completely different template than {{tl|cite}} and should have its own documentation. You can't just substitute part of the name (such as citation for cite) and expect it to work. —David Eppstein (talk) 20:24, 25 May 2014 (UTC)
 * Thanks; makes perfect sense now. Not sure why that didn't occur to me. Just wasn't parsing like a computer, I guess. Evensteven (talk) 01:53, 26 May 2014 (UTC)

Translator Field
Is there anyway that we could get a "Translator" set of fields that mirror "Editor" and has (trans) instead of (ed) or (eds)? I do a lot of work with non-English sources, and either have to use the "others=" field, which puts it in the wrong spot, or I have to put them as authors and add (trans) to their first name.ReformedArsenal (talk) 16:50, 18 February 2014 (UTC)
 * That is one of the purposes of the others parameter. A param that is dedicated to the translator has several times been asked for, not just here (the most recent being Template talk:Citation/Archive 6 and Template talk:Citation/Archive 6) but at several of the other discussion pages which concern the citation templates. -- Red rose64 (talk) 17:17, 18 February 2014 (UTC)

I understand that, but when you use the other= field it places it near the end, rather than in the author/editor section prior to the title of the work. It creates an improperly formatted citation. ReformedArsenal (talk) 14:18, 19 February 2014 (UTC)
 * The purpose of citations is to enable a reader to correctly locate the source document. We're not here to build a bibliographic database-see the Open Library for that mission. Are there cases where not having an explicit parameter for the translator would impair their ability to do that? LeadSongDog come howl!  17:48, 18 February 2014 (UTC)

The purpose of citations is not just to correctly locate the source document... if that were the case all we would need to do is put the ISBN/DOI/Whatever and we can get rid of every other field. There are many cases where there are multiple translations and not including the translator can cause confusion (Institutes of the Christian Religion has SEVERAL different translations), there are even some cases in which different translators have been used for different editions of the same primary source and therefore have the same ISBN, publisher, and title. It doesn't seem like this would be terribly difficult to do, and wouldn't detract from anything. I just don't have any experience with programming of this nature. ReformedArsenal (talk) 14:18, 19 February 2014 (UTC)
 * The purpose of citation is not simply to provide the minimal information to locate the source, it should include all bibliographic data that describe the source. Who translated a work (especially with multiple editions of multiple translators) can be just as important as who originally wrote it. Having to add "(trans)" to a name field perverts the meta-data, and having it in the "wrong" (i.e., non-standard) place (per "|others=") implies that we don't know or don't care about standard citation practice. Unless there is some strong reason for not doing so, I support having some kind of "trans" parameter. ~ J. Johnson (JJ) (talk) 19:27, 20 February 2014 (UTC)
 * Any chance that this can be addressed? This is still a very frustrating thing for those of us working on articles with ancient primary sources involved. ReformedArsenal (talk) 01:15, 12 March 2014 (UTC)
 * I would support a trans parameter per J. Johnson. Maybe it should have a trans-first and trans-last form, if that separation could be useful. —PC-XT+ 19:07, 21 March 2014 (UTC)

Still nothing? ReformedArsenal (talk) 13:41, 8 May 2014 (UTC)
 * Two options:
 * Add explicit fields for translator, illustrator, producer, director and any other needed descriptors
 * Add an author-type field as free form author descriptor
 * --  Gadget850talk 13:55, 8 May 2014 (UTC)

You wouldn't want it to be Author, because then it places it incorrectly in the citation. Ideally there would be an option for Translator that produces "#First Name# #Last Name# trans." from translator= trans_last= and trans_first= a trans_link= would be helpful as well. Basically the same functionality as the editor field but for translator (the other ones would be helpful too, but I'm not as concerned about those). ReformedArsenal (talk) 14:31, 8 May 2014 (UTC)
 * Any movement on this? ReformedArsenal (talk) 18:56, 21 May 2014 (UTC)
 * For what it's worth, I see the same needs here as ReformedArsenal, and would have an immediate application for it if it were available. I can only get something passable using the others= parameter. Evensteven (talk) 00:47, 22 May 2014 (UTC)
 * Oh, and ancient primary documents are not the only application. There are more modern documents in other languages that can have ready application to various fields of study, wherein identification of specific translators is essential to a citation. Evensteven (talk) 00:53, 22 May 2014 (UTC)
 * Translator and similar descriptor are used in Chicago and other styles. In Chicago, the translators are placed immediately after the title and enclosed in parentheses. With CS1, 'others' is also placed after the title:


 * So the location seems correct. There is no COinS key for translator, so this metadata is not an issue. --  Gadget850talk 15:02, 25 May 2014 (UTC)
 * The problem is that it doesn't actually allow for proper formatting, for example, if there IS no author. In such a case, translator would function just like editor does, taking the place of the Author. There is no way to accomplish this using the current template. ReformedArsenal (talk) 17:31, 25 May 2014 (UTC)
 * Also, |others= displays before the title if there is also a chapter or section, just like an editor. Evensteven (talk) 21:33, 25 May 2014 (UTC)
 * To clarify, it seems my previous comment only applies when |work= is present. If absent, then "others" is transported later in the citation to a place after volume(number) and before "edition". Evensteven (talk) 03:30, 26 May 2014 (UTC)
 * It seems pretty clear that the template as it stands does not properly handle translators, and it doesn't seem that there is any draw back to adding a translator field. Is there some kind of draw back that I'm not seeing? ReformedArsenal (talk) 22:10, 25 May 2014 (UTC)


 * Part of the reason that CS1 citations are the way that they are is because editors have willy-nilly added parameters to meet some specific need without considering what might be best in the long term. This is just another reason for my suggestion that CS1 needs its own style guide so that we can then standardize all of the CS1 templates to the guide.  Any additional parameters must then first be fit into the style guide before someone leaps into the code and just adds enough support to make the new parameter work.  The style guide also serves to document why others, for example, is placed where it is in the rendered citation.


 * First we must think about what it is that we want from every existing parameter, think about possible new parameters when existing parameters are found wanting, and think about how each CS1 template should use (or not use) the variety of available parameters. Then and only then should we move to implement what it is that we want.  If we do not do this then I fear that Module:Citation/CS1 will degenerate into a quagmire of code that will never be as good as it could have been and will become more and more difficult to maintain.


 * —Trappist the monk (talk) 22:10, 25 May 2014 (UTC)


 * Truly, I sympathize with this clearly reasonable, even necessary, approach. But until such a massive project can be completed, what to do? Engineering is always in this bind, wherever it is done. And I think the question regarding translators just begs that your very suggestion be followed for the establishment of that one parameter. I think your idea of a CS1 style guide has merit, and would serve to help address some of the high-level documentation problems I have brought up in another section. The underlying insufficiency is indeed an overall lack of vision and direction, which some agreed-on writing can provide. So, we have the engineering problem of visualizing and designing version 2 all the while supporting and maintaining version 1. Sometimes it is possible to make a start at version 2 by amending version 1 through an update, which also permits field testing and feedback. The need for translator is present now. Unless there is already a project afoot to really plan version 2, then it will never happen. This question can also serve to drive a beginning for that planning. Starting with creation of a "translator" section of the new CS1 style guide gives practical grounding, and it can develop from there. Feedback you already have from here, and you can always find more of that. The one thing we must avoid is paralysis because the tasks are too big. Here is a smaller task to start from. Let it develop into a plan, both for version 1 and version 2. Evensteven (talk) 04:22, 26 May 2014 (UTC)
 * You'll have to forgive me, but this really doesn't seem that it is that complicated of an issue. I know there is no formal style for Wikipedia, but it seems like it should be relatively easy to adopt a basic style guide's formatting (the citation template appears to follow Chicago/Turabian). Why not simply make fields that match the options for Chicago style (Author, Editor, Translator, etc). I don't quite understand why there would be confusion over who needs what since there isn't a specific style required. All things considered, it seems like there is a direct benefit and need for the translator field as proposed, so it isn't exactly willy-nilly. ReformedArsenal (talk) 21:48, 26 May 2014 (UTC)
 * Wrt the translator field, I must agree. Trappist was bringing up larger project issues re CS1, and that is sizable. To translate, I was answering that by saying one doesn't have to solve the whole larger project in order to do the translator field. In other words, a very standard engineering situation is not a drawback to its being done. There is no stated technical barrier to overcome. It is basic project management. Evensteven (talk) 02:12, 27 May 2014 (UTC)