Help talk:Citation Style 1/Archive 64

Access levels and archives
Is there a way to mark that a live URL has a subscription or similar access requirement but the archived version is free to read? This is not uncommon with old news articles? An example: live (subscription required to read complete article) archive (full article available for free.) Thryduulf (talk) 14:11, 17 February 2020 (UTC)
 * Doesn't cs1|2 do that already? Setting subscription causes the subscription icon to follow url in the rendered citation:
 * with live
 * with dead
 * cs1|2 doesn't support archive-url-status because an archived copy of the source's teaser-view of a subscription-only page seems rather pointless to me (though someone thought it a good idea to archive and cite the now defunct HighBeam Research teaser pages ... (example).
 * —Trappist the monk (talk) 15:44, 17 February 2020 (UTC)
 * I have actually found that archiving a page is a good way to get around GDPR paywall restrictions. That said, even if that is the case, I don't think supporting an icon on archives makes much sense. --Izno (talk) 16:57, 17 February 2020 (UTC)
 * cs1|2 doesn't support archive-url-status because an archived copy of the source's teaser-view of a subscription-only page seems rather pointless to me (though someone thought it a good idea to archive and cite the now defunct HighBeam Research teaser pages ... (example).
 * —Trappist the monk (talk) 15:44, 17 February 2020 (UTC)
 * I have actually found that archiving a page is a good way to get around GDPR paywall restrictions. That said, even if that is the case, I don't think supporting an icon on archives makes much sense. --Izno (talk) 16:57, 17 February 2020 (UTC)
 * I have actually found that archiving a page is a good way to get around GDPR paywall restrictions. That said, even if that is the case, I don't think supporting an icon on archives makes much sense. --Izno (talk) 16:57, 17 February 2020 (UTC)

Related question ( Template talk:Cite web redirects here ). The parameter |subscription= is still on the documentation page, but no longer works. Can someone retire it ? TGCP (talk) 08:09, 9 March 2020 (UTC)
 * ✅ for cite web. – Jonesey95 (talk) 17:20, 9 March 2020 (UTC)

Template causes a page to transclude itself?
If you put something like to produce on a page, the page trancludes itself on its own page. Just preview this section, go down to "Templates used in this preview" and see that "Help talk:Citation Style 1" is listed in the transclusions.

This is weird and shouldn't happen. &#32; Headbomb {t · c · p · b} 23:31, 9 March 2020 (UTC)
 * cs1|2 reads the article looking for templates so that it can auto-format dates.  To do that cs1|2 uses the title object's   method which records the page as a transclusion.
 * —Trappist the monk (talk) 00:22, 10 March 2020 (UTC)
 * Any way of making use of the linkless exists hack? I'm guessing a no, since it actually needs the content of the page. &#32; Headbomb {t · c · p · b} 00:33, 10 March 2020 (UTC)
 * No. Neither that template nor the underlying PROTECTIONEXPIRY magic word return article content.
 * —Trappist the monk (talk) 00:42, 10 March 2020 (UTC)

publication-place, place, or location and their proper use (cont.)
Following up on this discussion after seeing this Citation bot posting and this.

As a result of the original discussion, we created which, at this writing contains 822 articles. The code that adds articles to that category does not discriminate between publication-place and place or location having same or different values. When values are the same, cs1|2 renders only one.

I have written an awb script to troll that category and remove redundant parameters. After I run this script, we should have some idea about how multiple publication-place, place, location params are being used.

With regard to Editor Jc3s5h's HELP:CS1 edit, the Prefer "publication-place" over the ambiguous "location" recommendation, if that is what it is, is contrary to how these parameters are used. This is documented in the original discussion. That edit should, I think, be reverted.

—Trappist the monk (talk) 18:22, 8 March 2020 (UTC)
 * I believe a parameter name that will still be correct, even if another editor comes along and adds a place from a dateline, is preferable to a name that may need to be revised if information from a dateline is added. Jc3s5h (talk) 19:14, 8 March 2020 (UTC)
 * I, too, have a hard time understanding what consensus or reality of the templates the new suggestion is based on. Nemo 19:16, 8 March 2020 (UTC)
 * Example using location = San Francisco
 * Example using place = San Francisco
 * Example using publication-place = New York and place = San Francisco
 * Example using location = New York and place = San Francisco
 * So it looks like the template just doesn't properly support a story with a dateline in a paper where it's place of publication is included in the publication's title. Jc3s5h (talk) 19:53, 8 March 2020 (UTC)
 * I might be stating the obvious, but usually the documentation should refer to the current status of templates rather than an aspirational situation. Nemo 20:07, 8 March 2020 (UTC)
 * Finished running the awb script against which removed approximately 400 articles from the category.  What remains must be evaluated manually.
 * —Trappist the monk (talk) 14:25, 10 March 2020 (UTC)
 * So it looks like the template just doesn't properly support a story with a dateline in a paper where it's place of publication is included in the publication's title. Jc3s5h (talk) 19:53, 8 March 2020 (UTC)
 * I might be stating the obvious, but usually the documentation should refer to the current status of templates rather than an aspirational situation. Nemo 20:07, 8 March 2020 (UTC)
 * Finished running the awb script against which removed approximately 400 articles from the category.  What remains must be evaluated manually.
 * —Trappist the monk (talk) 14:25, 10 March 2020 (UTC)

where something was written is completely irrelevant to citations. The parameter should be removed. what citation guides recommend mentioning is the location of the publisher, because historically it was important to know where a publisher was so you could order a book from it. It was also useful to distinguish between different journals and magazines and newspapers of different cities that happened to have the same title. Where someone happened to sit down to write words is irrelevant.&#32; Headbomb {t · c · p · b} 20:45, 8 March 2020 (UTC)
 * I was having the same thought, just don't cite the location from the dateline. But then it crossed my mind that online newspapers, The New York Times in particular, tend to change the text of a headline during the course of a day, while the dateline is unchanged (except maybe for time of day) and there are negligible changes in the text of the article. Especially in the case of an article that does not name the authors, the place from the dateline can help a person verifying a citation that they have found the article they were looking for. Jc3s5h (talk) 20:52, 8 March 2020 (UTC)
 * All the examples of this somewhat rare case that I have seen either automatically redirect to the proper (edited) article, or land in a placeholder page with a link to the new url. This is not actionable as far as CS1 is concerned, imo. The change to the documentation should be reverted. CS1 is already too complicated. 108.182.15.109 (talk) 13:55, 10 March 2020 (UTC)

Still goes against all citation guides. Because there's no two distinct articles published in the same newspaper, on the same day, with the same title, but which were somehow written in different locations different locations. &#32; Headbomb {t · c · p · b} 21:11, 8 March 2020 (UTC)
 * Also location should be made the canonical parameter. It's by far the preferred parameter of editors, and is much, much shorter and easier to type. &#32; Headbomb {t · c · p · b} 04:36, 9 March 2020 (UTC)


 * I don't think the information should be removed. I consider it to the part of the properties belonging to the publication, and while it is not absolutely necessary it may help to put a source into perspective and to further research the background of a publication.
 * The problem, if there actually is one, is the ambiguity of the location and place parameters. Perhaps we can find better names, and slowly deprecate the old ones?
 * Specifically for the "written at" location, what about parameter names such as write-location, author-location, authoring-place, written-at-place, written-at, foreword-location, dateline-location, lockout-location?
 * The first few suggestions put the focus on the generic act of writing/authoring, regardless of where that location will be specified in the publication (in the foreword, in the dateline at the top of an article, or in the lockout at the bottom of an article). The last three suggestions focus on where the location shows up in the publication, and thereby may help editors to choose the correct parameter, however, given that, they would have to be supported in parallel although they are, I think, mutually exclusive, so the template would have to make sure that only one of them is actually used in a citation. (Or are there any types of publications which can have more than one out of a location in the foreword, dateline, or lockout?)
 * Further, are there, perhaps, synonyms for the term dateline linguistically emphasizing more on the location rather than the date? (In German, it is called a "Spitzmarke", if that helps to find a better English term for it.)
 * Regarding publication-place, this is already quite descriptive (but long), but it's not completely without ambiguity as well. What about publisher-place instead?
 * --Matthiaspaul (talk) 09:29, 10 March 2020 (UTC)
 * It is not the purpose of a citation to put a source into perspective and to further research the background of a publication. If it is necessary to do that in an article, create an end-note or footnote that has whatever extra information is required.  Creating more, rather esoteric, parameters does not seem to me to be an answer to the basic question which is: do we keep the publication-place and location or place functionality (written at...)?
 * —Trappist the monk (talk) 14:25, 10 March 2020 (UTC)
 * Where stories are 'written' is to give context to story for the reader. If you see "Syracuse, NY A man lost his life Sunday night following a break in..." this tells you the story is about a man in Syracuse, NY. It has nothing to do with any sort of bibliographic information relevant to citations. &#32; Headbomb {t · c · p · b} 18:58, 10 March 2020 (UTC)
 * More to the topic at hand, publication-place and place or location should all be aliases of location. The only place where this really causes confusion is where conference are held, which is normally put into the full title of the proceedings, so there might be room for conference-location. But location is the location of the publisher, and every style guide is quite clear on this. We should not invent conventions out of nowhere because a some editors don't know how to follow style guides. &#32; Headbomb {t · c · p · b} 19:02, 10 March 2020 (UTC)

Syncing to another wiki failed
A few of us have been trying to update the Cantonese Wikipedia's mirror of Module:Citation/CS1 since 2018 without success. If the most recent version (as of today) is ported, all citation templates on the Cantonese Wikipedia return the following error:
 * Lua error in 模組:Citation/CS1 at line 3799: attempt to index field 'date_names' (a nil value).

The relevant modules and templates on the Cantonese Wikipedia are located at:
 * yue:Module:Citation/CS1
 * yue:Template:Cite web

Any help will be greatly appreciated. --Deryck C. 12:27, 4 March 2020 (UTC)
 * You were getting the lua script error from the 2020 version of 模組:Citation/CS1 because the  table does not exist in your c. 2016 version of yue:Module:Citation/CS1/Configuration.
 * cs1|2 is a suite of eight lua modules and one css page. All of these are interdependent so you can't just update one module in a suite of older modules and expect good results; you will be disappointed.  Because cs1|2 is so complex, the best thing to do is to import all of the current cs1|2 module suite into sandbox modules at yue.wiki.  Get the sandbox version working (most of this work is going to be getting the contents of ~/Configuration right) and then update the yue.wiki live cs1|2 from your sandbox.  Yeah, I know it's a lot of work to update a very old module suite to the current version.  Let me know if you have trouble.
 * —Trappist the monk (talk) 13:08, 4 March 2020 (UTC)
 * I'll give it a go tonight. Looks like only Module:Citation/CS1/Date validation and Module:Citation/CS1/Configuration have been localised at yue.wp; the rest are direct clones of en.wp. If I get stuck I'll self-revert and then ask you again for help. Deryck C. 22:37, 14 March 2020 (UTC)
 * Porting the new version seems to be successful, though I have a question about date formatting. Where in the code is it that the default date formatting is set? The templates seem to be outputting all dates in dd MMMM yyyy which does not make sense in Cantonese (due to month names not being used in Cantonese). The East Asian languages have always needed customised date formatting strings in the format of "yyyy年m月d號". Deryck C. 00:08, 15 March 2020 (UTC)
 * There isn't a 'default' date format. The module suite only validates that the date format written in the raw template is valid according to the date format standards that the wiki decides are acceptable.  At en.wiki those formats are specified at MOS:DATE.  Show me an explicit example.  Where I looked I saw dates rendered as you describe them.
 * —Trappist the monk (talk) 00:32, 15 March 2020 (UTC)
 * If the CS1 modules don't format the dates, then it must be somewhere upstream that has a problem. Sorry about that. Deryck C. 15:49, 15 March 2020 (UTC)
 * Still, show me where you are seeing this. Editors can use dmy-all to render dates in dmy format.  If there are any articles copied from en.wiki that have contain the text   or redirects of that (listed in ~/Configuration) then all dates on that particular page will be rendered in dmy format.  Show me where you are seeing all dmy date format.
 * —Trappist the monk (talk) 15:56, 15 March 2020 (UTC)
 * It was yue:格林尼治 where I saw some "15 三月 2020" style dates. But either my memory was patchy or there was replag; when I checked again this afternoon only one such date was left, and on further investigation it was actually stated as "15 三月 2020" in the Wikitext, so I went to correct it. Now the page has at least 3 different date formats (en-GB, yue, ISO), but at least all of them are correct. Deryck C. 22:48, 15 March 2020 (UTC)
 * It was yue:格林尼治 where I saw some "15 三月 2020" style dates. But either my memory was patchy or there was replag; when I checked again this afternoon only one such date was left, and on further investigation it was actually stated as "15 三月 2020" in the Wikitext, so I went to correct it. Now the page has at least 3 different date formats (en-GB, yue, ISO), but at least all of them are correct. Deryck C. 22:48, 15 March 2020 (UTC)

Namespace check for Category:Articles with missing Cite arXiv inputs and Category:Pages with DOIs inactive as of 2019 January
There should be a namespace check on these error. Or at least a documentation check on the second one. &#32; Headbomb {t · c · p · b} 20:21, 6 March 2020 (UTC)
 * In general, category errors should probably only be done in Article/Draft/Template spaces. &#32; Headbomb {t · c · p · b} 20:27, 6 March 2020 (UTC)
 * There is a namespace check on all errors. cs1|2 does not categorize:
 * User, Talk, User talk, Wikipedia talk, File talk, Template talk, Help talk, Category talk, Portal talk, Book talk, Draft talk, Education Program talk, Module talk, and MediaWiki talk
 * The category is added because Help talk:Citation Style 1/Archive 8 has  with no which no longer works as it once did (the comparison did not include ).  After the change, &lt;anything> causes  to include the 'old'  in its rendering.
 * For, transcludes Template:cite journal/doc.  Both are in the template namespace so the error categories are expected and desired.  To turn-off categorization in the ~/doc page use yes or one of its aliases.
 * —Trappist the monk (talk) 21:13, 6 March 2020 (UTC)
 * I dealt with the doc page. However, the arxiv categories still need to be taken care of. This also includes Category:Articles with a publisher parameter in their Cite arxiv templates and Category:Articles with a journal parameter in their Cite arxiv templates. &#32; Headbomb {t · c · p · b} 22:03, 6 March 2020 (UTC)
 * Category:Pages with DOIs inactive as of 2019 September also contains Wikipedia-namespace pages, which shouldn't be in there. &#32; Headbomb {t · c · p · b} 22:24, 6 March 2020 (UTC)
 * and arise from  called by  in Help talk:Citation Style 1/Archive 8.  Neither category has been supported by cs1|2 since  was converted to lua.
 * The roster of uncategorized namespaces was determined at
 * —Trappist the monk (talk) 23:02, 6 March 2020 (UTC)
 * The "roster" was proposed when the draft namespace didn't exist. Drafts should be covered. If the issue is caused by cite arxiv/old or cite compare, those should be updated to not emit those categories when in those non-supported namespaces. &#32; Headbomb {t · c · p · b} 23:30, 6 March 2020 (UTC)
 * —Trappist the monk (talk) 23:02, 6 March 2020 (UTC)
 * The "roster" was proposed when the draft namespace didn't exist. Drafts should be covered. If the issue is caused by cite arxiv/old or cite compare, those should be updated to not emit those categories when in those non-supported namespaces. &#32; Headbomb {t · c · p · b} 23:30, 6 March 2020 (UTC)

The arxiv problem seems fixed now, somehow. Thanks to whoever helped. Still the issue of Wikipedia-namespace errors tracking though. &#32; Headbomb {t · c · p · b} 23:28, 15 March 2020 (UTC)

comma space error
did a search on the archive and found this which fixed the |page= option, but it also happens on the |issue= option. Dave Rave (talk) 00:47, 15 February 2020 (UTC)
 * A real-life example of what you mean will help us help you. Example please?
 * —Trappist the monk (talk) 01:12, 15 February 2020 (UTC)
 * any page with the option on it would show it to you .... you could sandbox it ?? oh, let me just hold your hand for a bit Dave Rave (talk) 01:16, 15 February 2020 (UTC)
 * Not helpful. If there is a problem with a citation template, please provide exact information. Vague statements about archived discussions involving unspecified options in "pages" will delay resolution of the issue if there is any. You could have stated that issue does not render properly if it includes a comma as in 5-digit numbers. That would have moved the discussion faster. 199.102.115.203 (talk) 15:03, 15 February 2020 (UTC)
 * , the cite compare template is very useful for demonstrating simple test cases for all of the CS1 templates. You can see examples of its use above. – Jonesey95 (talk) 16:25, 15 February 2020 (UTC)
 * what is not useful ? the archive on the pages which has the same problem ? the let me show you page where you can see all the refs with comma space, or is it just too hard ? Dave Rave (talk) 20:51, 15 February 2020 (UTC)
 * You (sarcastically) pointed us to a whole article and, before that, to a long talk page discussion that covered multiple topics. You are asking us to scrub through both of those things, trying to figure out what you are talking about. If you actually want help, be clear and concise in your request for help. If you just want to argue and accuse us of being lazy, please choose a different venue. – Jonesey95 (talk) 21:29, 15 February 2020 (UTC)
 * that might be the comma space error purported to in the title, oh, heavens, it's like talking to a wall. Must be hard being administrative, like being on a help desk. Dave Rave (talk) 00:06, 16 February 2020 (UTC)
 * As a somewhat off-topic sidenote, these extremely high 5-digit issue numbers were added in this edit. While I can imagine issue numbers reaching these heights for daily newspapers, I must admit that I have never seen any such high numbers printed on newspapers myself (typically they are divided into yearly volumes and thus the issue numbers remain much smaller). Therefore (and since the original contributor can't be bothered to answer this question, unfortunately), can someone confirm that these numbers are genuine and real? Were they actually printed as such on these newspapers or are they some calculated or otherwise derived numbers found in some later (online) index? If they are not true issue numbers but other index numbers, should they perhaps be moved into the id parameter instead? Thanks. --Matthiaspaul (talk) 09:26, 16 March 2020 (UTC)
 * Todays The Times (London) has No. 73,108 after the date on the front page (see here near middle). —  Jts1882 &#124; talk 11:41, 16 March 2020 (UTC)
 * I see, thanks for the pointer. Much appreciated.
 * --Matthiaspaul (talk) 14:27, 16 March 2020 (UTC)

Cite ebook
redirects here, but I can't see where the info is. I guess the important thing is what replaces page numbers, and whether it works with harv referencing. —— SN  54129  14:23, 19 March 2020 (UTC)
 * was created as vandalism in 2010; now just a redirect to.
 * —Trappist the monk (talk) 14:48, 19 March 2020 (UTC)
 * As far as differences in citation style go, the general guidance is the same as it always is: include as much information as you can. Some ebooks are clearly and reliably paginated, while others are infinite scrolling nonsense. If there's page numbers in your edition, cite the page numbers. If there aren't any or they're not a reliable indicator, use chapter or at in citations and loc in harv footnotes to get as close as you can. --AntiCompositeNumber (talk) 14:52, 19 March 2020 (UTC)
 * Even in print books, pagination can vary between editions, or even printings. Citing a chapter as well as a location or page number in an ebook, if you are willing to provide both, will help WP readers verify claims in articles. – Jonesey95 (talk) 15:18, 19 March 2020 (UTC)

Add support for title-url
This would be very useful to indicate which url is meant when you have title/chapter/contribution/etc... &#32; Headbomb {t · c · p · b} 23:54, 15 March 2020 (UTC)
 * So url would be an alias for title-url? Kanguole 00:26, 16 March 2020 (UTC)
 * Yes. Or the other way around. However aliases work, as long as they are synonymous. &#32; Headbomb {t · c · p · b} 00:34, 16 March 2020 (UTC)
 * Or like author-first/last are aliases for first/last for symmetry with editor-first/last? --Matthiaspaul (talk) 09:35, 16 March 2020 (UTC)
 * This question was meant for clarification. As I prefer specific and non-ambiguous parameter names (even if they are longer), I would support the addition of this alias.
 * --Matthiaspaul (talk) 04:14, 20 March 2020 (UTC)

Possible additional parameter check for list of authors/editors etc.
I don't know if this is a frequent error, but over the years I occasionally ran into citations doubling some of the authors or editors in longer lists. This was probably down to copy&paste errors during citation composition.

This condition could be detected if the template would check the list of (recombined first+last) author names (and likewise the list of editors) for duplicates and display a warning in edit preview.

I'm not sure if such a test would be too expensive to be performed outside edit preview as well, but if not, we would probably need some method to override the test for the (rare) case of multiple people of the same name contributing to a publication.

--Matthiaspaul (talk) 04:44, 20 March 2020 (UTC)

Add support for series-editor parameters
Some publications distinguish between authors, editors, and series editors. In order not to have to lump together the two types of editors, I would welcome if we had a series-editor* range of parameters in addition to the existing editor* range of parameter variants. They would be treated almost identical to normal editors, but listed after the authors and editors. Where normal editors are indicated by "(ed.)"/"(eds.)", series editors would be indicated by "(series ed.)"/"(series eds.)". Since AFAIK there is at present no separate class for series editors in metadata, they should be classified as editors there. --Matthiaspaul (talk) 04:28, 20 March 2020 (UTC)
 * others is also available in the meantime. If series-editor* existed, would it require series to exist? – Jonesey95 (talk) 04:38, 20 March 2020 (UTC)
 * Good point! In the examples that come to my mind right now, the name of the series was given as well, so it could (and should) be specified as well. (Not the other way around, I am aware of examples where a series name was given, but no series editors.)
 * Does someone know of an example where series editors are specified without also mentioning the name of the series? Depending on if such examples exist, a missing series parameter should throw a warning in edit preview or an error in the article, IMO.
 * --Matthiaspaul (talk) 06:20, 20 March 2020 (UTC)
 * What style guide even recommends that series editors are mentioned in a citation? I'd be against adding such pointless information. &#32; Headbomb {t · c · p · b} 08:19, 20 March 2020 (UTC)
 * Citations are meant to provide the minimum information needed to locate the original, not to give credit to everyone involved in its production. The citation templates are complicated enough already. Peter coxhead (talk) 09:23, 20 March 2020 (UTC)

Suggestion to add support for SBN parameter
Hi, I'd like to suggest to add support for an sbn parameter. Right now, editors have to convert SBNs into ISBNs and use the isbn parameter instead. However, ISBNs in pre-1968 citations look odd and are historically incorrect. Also, some editors might not know how to convert SBNs into ISBNs (although it is easy) and as a consequence not mention the SBN at all. Since the conversion is just adding a "0-" prefix to the SBN, adding support for an sbn parameter would be easy, as the template could internally do the conversion for number validation, to implement the underlying ISBN link and to feed metadata. The only difference would be that in the rendered citation, SBNs would correctly display as



instead of incorrectly as


 * ISBN 0-356-02201-3

--Matthiaspaul (talk) 11:12, 5 March 2020 (UTC)
 * At first blush, this seemed a simple thing to do. Alas, the first blush was more of a rash.  But:
 * —Trappist the monk (talk) 16:45, 6 March 2020 (UTC)
 * Cool, thanks! :-)
 * However, I would suggest to internally prefix with "0-" rather than "0" only. If the SBN number already contains hyphens (356-02201-3), "0-" blends in naturally (0-356-02201-3), and if it contains just a number (356022013), the "-" still makes sense to let the embedded SBN stand out a little (0-356022013). Looks slightly better to me than the other way around, however, is only cosmetically.
 * I would also suggest to deliberately route the SBN link in the prefix through a redirect (also per WP:NOTBROKEN), either the normal Standard Book Number or even better the Standard Book Number (identifier) one, so that "Standard Book Number" rather than "International Standard Book Number" shows up in the tooltip when hovering over the link. I think, the fact that this will cause the "Redirected from ..." note to be displayed at the top of the "International Standard Book Number" article is useful as well in order to avoid any confusion why the user lands on this article. Going through the (identifier) link would have the additional advantage that "normal" article links to the topic "Standard Book Number" can be easily distinguished from links for specific books with SBNs. This helps to avoid clutter in "What links here" (potentially thousands of incoming links from books intermixed with "normal" article links), enables filtering for one or the other type, and thereby improves reverse lookup of pages using these templates, effectively easing maintenance and research.
 * IMO, the same (identifier) scheme could/should be applied to ISBN etc. links as well - having the template links grouped under a specific redirect (which isn't used for other purpose, except for by an occasional accident, perhaps) would significantly improve the reverse lookup situation of normal links to the "International Standard Book Number" article.
 * --Matthiaspaul (talk) 11:31, 7 March 2020 (UTC)
 * Without we rewrite how internally-linked identifiers are handled, the cosmetic hyphenation will not be done.   assembles precomposed link parts stored in ~/Configuration to create the final rendering.  The precomposed parts are static and cannot be modified.  Because we want the 9-digit version to be displayed and linked using a 10-digit version and because   has no support for such a combination, and because this is merely cosmetic, I'm not inclined to rewrite working code to support this unique case.
 * I have switched the label link to Standard Book Number (identifier) and rather like the idea of doing the same for the other identifiers. Before I do so, are there any reasons that we shouldn't do that for the other identifiers?
 * —Trappist the monk (talk) 13:20, 7 March 2020 (UTC)
 * I can't think of any - in fact, I only see advantages. However, when I implemented this a couple of years ago for some of the Catalog Lookup Links to address some other issues reported by users, this was rigorously changed to direct links by Headbomb. I tried hard but never found his argumentation conclusive (he didn't like the hatnote and said it would be inconsistent with other links (even though some of the CLLs I created were consistently using this scheme right from the start)), but at some point I gave up unconvinced.
 * In either case, we can easily address the consistency issue if we'd switch to use this scheme for all such identifiers in the citation template framework and the CLL templates (and since Mediawiki's ISBN magic token functionality is no longer around, this could not interfere with yet another link style as well). Once rippled through the pages I think this scheme would make reverse lookup much easier to use again on the affected pages. Probably unavoidable at first, some users might wonder why not linking to the target page directly, but if we explain the reasons for this in the template documentation I think everyone will eventually see the advantages and be happy with it in the end. To play it extra-safe we could even create a new R from identifier rcat for these kinds of redirects.
 * --Matthiaspaul (talk) 14:56, 7 March 2020 (UTC)
 * Rereading your post, I see that I misunderstood. If it is ok to insert a hyphen between the leading zero and an unhyphenated sbn identifier value, we can set the static portion of the internal link to include  .  I understood that you want   only when the sbn was hyphenated so 0-356-02201-3 or 0356022013; that requires a rewrite.
 * —Trappist the monk (talk) 14:28, 7 March 2020 (UTC)
 * Of course, the latter would be even better (but probably not worth the effort), but, yeah, I just meant to prefix with "0-" even in the unhyphenated case.
 * --Matthiaspaul (talk) 14:56, 7 March 2020 (UTC)
 * Rereading your post, I see that I misunderstood. If it is ok to insert a hyphen between the leading zero and an unhyphenated sbn identifier value, we can set the static portion of the internal link to include  .  I understood that you want   only when the sbn was hyphenated so 0-356-02201-3 or 0356022013; that requires a rewrite.
 * —Trappist the monk (talk) 14:28, 7 March 2020 (UTC)
 * Of course, the latter would be even better (but probably not worth the effort), but, yeah, I just meant to prefix with "0-" even in the unhyphenated case.
 * --Matthiaspaul (talk) 14:56, 7 March 2020 (UTC)
 * —Trappist the monk (talk) 14:28, 7 March 2020 (UTC)
 * Of course, the latter would be even better (but probably not worth the effort), but, yeah, I just meant to prefix with "0-" even in the unhyphenated case.
 * --Matthiaspaul (talk) 14:56, 7 March 2020 (UTC)

There is a zillion reason not to create pointless (identifier) links, chief among them is that none of these links require disambiguation, but there is no difference between a 'manual' link to International Standard Book Number and one made through a citation template, much like there is no difference between a 'manual' link to electron and one made via Particles. If you're interested in finding a list of articles that link to International Standard Book Number without a citation template, then it's a simple matter of making a  search for those, either in the search bar, or with AWB. &#32; Headbomb {t · c · p · b} 04:26, 8 March 2020 (UTC)
 * That's a strawman and beside the point - I would really appreciate a more problem- and solution-oriented approach.
 * We are trying to find an elegant (and easily doable!) solution to solve a longstanding problem (which you simply keep on ignoring). By routing through redirects, I think we have a solution which does not even create any disadvantages elsewhere. (An even better solution would be more advanced filter capabilities built into "What links here". Unfortunately, this is not under our control and therefore not doable - but even if it would be done some years in the future, the currently proposed solution would not interfere with it, so there is no reason not to go for it.)
 * Why should users have to use external tools like AWB (which is a security risk and not even possible to use in many environments) or use arcane advanced search syntaxes even most technical users won't be able to reproduce just to perform simple actions like reverse lookup for normal article development? That's absurd.
 * The underlying problem with those "manual" links is that they pollute "What links here". This is also a long-standing problem for links from navigation templates, but it is even more prevalent with links from citation templates, adding f.e. hundreds of thousands of links to the International Standard Book Number article which would otherwise have only a few hundred incoming links. However, those remaining links are what most users are interested in in normal article development and when researching the topic. On the other hand, there are users only interested in the template links. So, it is desirable for users to have a choice, get all incoming links, only all normal links, or only all template links.
 * People even thought of getting rid of the template links in "What links here" by routing them through namespace prefixes. While this avoids the pollution, it creates the problem that it makes reverse lookup of those links impossible (and thereby might cause problems for some bot tasks), breaks the network of links, invalidates statistics, etc.
 * So, what is needed is some kind of middle ground between direct links and prefixed links, a system where all incoming links still show up in "What links here" for link network integrity, but where they can be grouped into "template links" and "non-template links". Since "What links here" presents links in random order, routing citation template links through redirects is adding some meta info without removing something, so there is no change for users who don't need to distinguish between them, but those who do now have a chance to select/filter the type of links they are interested in.
 * For this grouping to work, almost any redirect name would do, be it "Citation identifier link to Standard Book Number" or "Standard Book Number (identifier)". However, a parenthetical disambiguator like "(identifier)" looks nicer, can be applied universally (because identifier names never include brackets), and helps to visually separate an identifier's name from the disambiguator in tooltips. You are right that technically the disambiguator is not absolutely necessary, but there is also nothing which would disallow to use one - we even have an rcat for this: R from unnecessary disambiguation.
 * Basically the only potential "disadvantage" of routing through redirects is that a small "Redirected from ..." hatnote will be shown on the target page. However, with a carefully chosen disambiguator like "(identifier)", this even looks sensible and almost as if it would have been designed for this very purpose, so I see this even as an advantage rather than a disadvantage. In either case, it is only cosmetics.
 * There is one difference between links from navigation templates and links from citation templates. In navigation templates we normally try to use direct links so that the selected topic is shown in boldface on the target page in order to help navigation. While this is unlikely to happen in citation templates, where it happens it is even highly undesirable. Routing the identifier links through redirects has the nice sideeffect of ensuring that this cannot happen at all any more.
 * So, based on actual use cases there are many solid arguments pro linking through redirects and pro the "(identifier)" name. For consistency, we just need to make sure to switch the citation templates and CLL templates at about the same time.
 * --Matthiaspaul (talk) 12:21, 8 March 2020 (UTC)
 * The problem here is that you assume there is a problem where there is none. Wikipedia is written for readers, and the links in citation templates should be optimized to save readers the hassle. What's an ISBN? It's the International Standard Book Number. This is what pops up when you hover the link, and this is where you are taken when you click on ISBN. There's no sense in presenting readers 'International Standard Book Number (identifier)' when hovering the ISBN links, as if there is a different kind of ISBN out there, or forcing them to go through a pointless redirect before they end up at the article they expect. If you want to know what links to International Standard Book Number without templates, you can search for that in the same way everyone search for so-called 'direct' links that aren't embedded via templates: . Or without proposing nonsensical schemes like "Citation identifier link to Standard Book Number" or "Navbox link to Standard Book Number" or "Not-a-citation, but-still an identifier link to Standard Book Number" or "See also section link to Standard Book Number". &#32; Headbomb {t · c · p · b} 12:31, 8 March 2020 (UTC)
 * Yeah, Wikipedia is for readers, and readers want to use "What links here" when they research a topic. And these direct template links make it next to impossible for them to use "What links here" efficiently. This is a real use case and problem, and if we can address it in some elegant way, it would be almost irresponsible to ignore it. Forget about the "insource" thing - I guess less than 0.1% of the contributors know that this syntax even exists and how to use it, let alone normal users.
 * Regarding tooltips for a link named ISBN in a citation, displaying "International Standard Book Number" or "International Standard Book Number (identifier)" is about the same from a user's perspective. Both look nice and as if designed for this purpose, and since Wikipedians are used to this naming scheme, the "(identifier)" appendage does not interfere with the name at all. I think, in this specific context, an ISBN identifier in a citation, the later case is even more descriptive and to the point than just "International Standard Book Number". Nobody will assume this would be a different kind of identifier which actually happens to be named "International Standard Book Number (identifier)", that is, including the brackets. In contrast to this, something like "Citation identifier link to Standard Book Number" really looks ugly, that's why I gave it as a counter-example for a redirect name when we want to avoid the "(identifier)" name appendage.
 * I can't see anything forceful in going through redirects, it's actually good practise to link through the most specific semantically and syntactically matching redirect available. This allows more specific reverse lookup and also simplifies future maintenance, not only in this particular case, but in general (also per WP:NOTBROKEN).
 * --Matthiaspaul (talk) 14:02, 8 March 2020 (UTC)
 * What if we do an experiment? What if we change the label link for the isbn rendering from International Standard Book Number to International Standard Book Number (identifier)?  If I understand what Editor Matthiaspaul is saying, then we should see a dramatic reduction in the number of listed articles at Special:WhatLinksHere/International Standard Book Number.  Do I have this right?
 * If the experiment is successful, and there is still objection to the redirect's 'identifier' dab appearing when readers hover the mouse pointer over the identifier label then we might do something like this:
 * → ISBN
 * —Trappist the monk (talk) 14:34, 8 March 2020 (UTC)
 * —Trappist the monk (talk) 14:34, 8 March 2020 (UTC)


 * Sounds good to me. I support this. --Matthiaspaul (talk) 16:15, 8 March 2020 (UTC)

There will be exactly the same amount of links to that special page because links to redirects are also listed there. And I also object to having pointless redirects in the first place. The links to ISBN from a citations are no less important than links from manual citations or from a mention in prose or in a see also section, or from a navbox. Especially when done unilaterally without a dedicated RFC asking if people want citations to link to stuff like PubMed Identifier (identifier). &#32; Headbomb {t · c · p · b} 20:37, 8 March 2020 (UTC)


 * Trappist already suggested a way how to suppress the display of "(identifier)" if this would be really necessary (I don't think it is, but it is good to know that we can if we need to). Also, we could use "PubMed ID (identifier)" or "PubMed (identifier)" or even "PMID (identifier)", or choose another scheme like "PubMed Identifier (citation link)", "PubMed Identifier (template link)" or "Citation link: PubMed Identifier". The point is to go through a redirect rather than linking directly. This will improve functionality/usability significantly, whereas the name of the redirect is cosmetics only. Technically, it could even be different for different identifiers (of course, it should not for consistency reasons, so that people can memorize this convention as a general scheme). In my opinion, "(identifier)" is nicely short, unobtrusive and not too specific - and at the same time it is self-explanatory enough for a generic name. And "PubMed Identifier" is one of very few, where a word would be repeated, something that doesn't harm because by using the parenthetical disambiguation scheme, it is immediately obvious to users that "PubMed Identifier" is the name of the identifier, and "(identifier)" is for classification inside WP. But again, there are options to avoid that as well.
 * --Matthiaspaul (talk) 10:59, 9 March 2020 (UTC)


 * Probably will be the same number of articles listed but after a bit of experimentation, I think that how the articles will be listed will be different. Right now, without the redirect, Special:WhatLinksHere/International Standard Book Number begins:
 * Anarchism ‎ (links | edit )
 * Autism ‎ (links | edit )
 * Albedo ‎ (links | edit )
 * A ‎ (links | edit )
 * Alabama ‎ (links | edit )
 * etc
 * With the International Standard Book Number (identifier) redirect I suspect that those same articles will be listed something like this:
 * International Standard Book Number (identifier) (redirect page) ‎ (links | edit )
 * Anarchism ‎ (links | edit )
 * Autism ‎ (links | edit )
 * Albedo ‎ (links | edit )
 * A ‎ (links | edit )
 * Alabama ‎ (links | edit )
 * etc
 * The Hide redirects filter will remove International Standard Book Number (identifier) and all of the articles that use the redirect from the Special:WhatLinksHere display. To prove this, I created User:Trappist the monk/sandbox redirect which redirects to my sandbox.  I then created User:Trappist the monk/test which links to my sandbox through the sandbox redirect.  You can see this at Special:WhatLinksHere/User:Trappist_the_monk/sandbox.  If you click the Hide redirects filter link, the sandbox redirect and the test page are removed from the display.
 * —Trappist the monk (talk) 22:51, 8 March 2020 (UTC)
 * Yes, the number of totally incoming links will not change, they will just be grouped differently. So far, there was no order, and users (or bots) for whom all incoming links were equally important (like apparently Headbomb above) had to recursively traverse through the whole list to ensure they catched all links. They can continue to treat the list as unsorted as before, and will not miss a single link. However, for most users, some classes of links are much more important than others. The walls of links from citation templates are mostly seen as clutter in normal article development and topic research, however, there are also users who are particularly interested in just these links. Lumping them all together, users had no choice and had to somehow live with the clutter or give up on using "What links here" efficiently. Going through the redirect, users can now suppress the bulk of links they aren't interested in simply by using the "Hide" function. And since it's optional, nothing get's lost for those who don't want or need this. Best of both worlds.
 * --Matthiaspaul (talk) 10:59, 9 March 2020 (UTC)
 * This discussion has been mentioned at.
 * —Trappist the monk (talk) 18:35, 11 March 2020 (UTC)

choosing identifier redirects

 * arXiv
 * arXiv (identifier)
 * arXiv
 * Amazon Standard Identification Number
 * Amazon Standard Identification Number (identifier)
 * ASIN
 * Bibcode
 * Bibcode (identifier)
 * Bibcode
 * bioRxiv
 * bioRxiv (identifier)
 * bioRxiv
 * CiteSeerX
 * CiteSeerX (identifier)
 * CiteSeerX
 * Digital object identifier
 * Digital object identifier (identifier)
 * doi
 * International_Standard_Serial_Number
 * International_Standard_Serial_Number
 * Electronic International Standard Serial Number (identifier)
 * Electronic ISSN (identifier)
 * eISSN
 * Handle System
 * Handle System (identifier)
 * hdl
 * International Standard Book Number
 * International Standard Book Number (identifier)
 * ISBN
 * International Standard Music Number
 * International Standard Music Number (identifier)
 * ISMN
 * International Standard Serial Number
 * International Standard Serial Number (identifier)
 * ISSN
 * Jahrbuch über die Fortschritte der Mathematik
 * Jahrbuch über die Fortschritte der Mathematik (identifier)
 * JFM
 * JSTOR
 * JSTOR (identifier)
 * JSTOR
 * Library of Congress Control Number
 * Library of Congress Control Number (identifier)
 * LCCN
 * Mathematical Reviews
 * Mathematical Reviews (identifier)
 * MR
 * OCLC
 * OCLC (identifier)
 * OCLC
 * Open Library
 * Open Library (identifier)
 * OL
 * Office of Scientific and Technical Information
 * Office of Scientific and Technical Information (identifier)
 * OSTI
 * PubMed Central
 * PubMed Central (identifier)
 * PMC
 * PubMed Identifier
 * PubMed Identifier (identifier)
 * PubMed (identifier)
 * PMID
 * Request for Comments
 * Request for Comments (identifier)
 * RFC
 * Standard Book Number
 * Standard Book Number (identifier)
 * SBN
 * Social Science Research Network
 * Social Science Research Network (identifier)
 * SSRN
 * Semantic Scholar
 * Semantic Scholar (identifier)
 * S2CID
 * Usenet
 * Usenet (identifier)
 * Usenet:
 * Zentralblatt MATH
 * Zentralblatt MATH (identifier)
 * Zbl

Presuming that we pursue this, what are the identifier-label links? In the list above are the current identifier-label links followed by: Feel free to add other possible redirect constructs
 * 1) the same links with the   dab
 * 2) is the identifier label as rendered by cs1|2 with the   dab and a  tag that holds the en.wiki article name

—Trappist the monk (talk) 15:28, 10 March 2020 (UTC)


 * I added three more, not necessarily because I prefer them, but simply for completeness.
 * I like both of your suggested schemes, and can find pro arguments for both of them. For example:
 * Linking through the expanded name of an identifier (plus "(identifier)") ensures that the tooltip will already provide some helpful information for users who don't know what the symbolic name stands for. Otherwise, people will see the expanded form only by clicking the link (or enabling scripts).
 * Linking through the symbolic name (plus "(identifier)") has the advantage that the names are a complete no-brainer from our perspective. Also, the symbolic names will never contain the word "identifier" themselves (addressing Headbomb's argument above). Also, some identifiers might not have a proper expanded name (or the expanded name might depend on the target language), so just using the symbolic name in the redirect will ensure a higher degree of consistency in the name scheme. Finally, possible future changes in the name or spelling of a redirect's target page can be handled inside the redirect and do not require the citation template's code to be updated.
 * So, the first scheme appears to be slightly more helpful, your second scheme might be slightly easier to maintain and to keep consistent in the long term future.
 * I'm happy with both of them.
 * --Matthiaspaul (talk) 22:21, 10 March 2020 (UTC)


 * links to ISBN from […] citations are no less important than links from manual citations or from a mention in prose or in a see also section, or from a navbox
 * Ah, but heaven forbid that the biography of someone who spent his/her entire life in the United States to should ever link to United States. In fact, we have a certain cadre of users who will revert you 'til doomsday for doing that. I don't agree with their logic at all, but I would argue that a link to ISBN in cite book is of even lower practical value than links to first-world geography topics so often targeted for mass-demotion to plain text.
 * "What links here" presents links in random order
 * The order looks highly irregular but is far from random. Pages are ordered by, which is   for Anarchism and   for this talk page. This numbering is (very roughly) alphabetical for our oldest pages (the order in which they were imported from some previous system), and chronological for pages created after that time.
 * But there are still no other sorting options. True chronological and true alphabetical would be a good start. Maybe even by page size. The most difficult but most useful one would be "relevance to context" as estimated either by some kind of algorithm (something akin to whatever determines Special:Search result order).
 * But certainly, improved filtering options would help, e.g. the following, which should definitely be separate items:
 * ☑ Exclude links originating from the "body" of a template (e.g.  in cite book, or any link in a navbox).
 * ☑ Exclude links originating from the parameter of a template (e.g. American, or Cambridge)
 * Piping a link to a suffixed redirect and using a tooltip hack to obscure the true target really seems like the worst possible solution. And by that I mean worse than doing nothing. ―cobaltcigs 17:23, 11 March 2020 (UTC)
 * Thanks for the explanation about how Special:WhatLinksHere pages are ordered.
 * Chastising us for the lack of filtering and sorting options available at Special:WhatLinksHere seems to me to be counter productive. It would be better to raise those issues with the developers at MediaWiki or wherever they are since we can do nothing to fix what you perceive to be broken.
 * Umm, links through redirects, whether piped or not, always obscure the true target:
 * → Standard Book Number has a tooltip: 'Standard Book Number'; redirects to International Standard Book Number
 * → SBN has a tooltip: 'Standard Book Number'; redirects to International Standard Book Number#SBN
 * → SBN has a tooltip: 'Standard Book Number (identifier)'; redirects to International Standard Book Number#SBN
 * Using a [piped] link to a suffixed redirect and using a tooltip hack:
 * → SBN   has a tooltip: 'Standard Book Number'; redirects to International Standard Book Number#SBN
 * The 'hack', if implemented, will also obscure the true target so, to me, the point you are trying to make is somewhat 'obscured'. If it is critical to reveal the true target, the 'hack' can do that:
 * → SBN  has a tooltip: 'International Standard Book Number#SBN'; redirects to International Standard Book Number#SBN
 * —Trappist the monk (talk) 18:32, 11 March 2020 (UTC)
 * Those tooltips are very inaccurate.
 * → Tooltip: International Standard Book Number
 * → Tooltip: International Standard Book Number
 * → Tooltip: International Standard Book Number
 * If tooltips are implemented (which is a seperate issues from having pointless redirects), they should give the full description of the identifier, e.g. LCCN, JFM ,  Zbl , and so on. &#32; Headbomb {t · c · p · b} 18:51, 11 March 2020 (UTC)
 * What I mean is a link to a redirect should, in all cases, show the default tooltip of that redirect's title (being the target of that link) rather than pretending not to be a redirect. ―cobaltcigs 00:05, 12 March 2020 (UTC)
 * If tooltips are implemented (which is a seperate issues from having pointless redirects), they should give the full description of the identifier, e.g. LCCN, JFM ,  Zbl , and so on. &#32; Headbomb {t · c · p · b} 18:51, 11 March 2020 (UTC)
 * What I mean is a link to a redirect should, in all cases, show the default tooltip of that redirect's title (being the target of that link) rather than pretending not to be a redirect. ―cobaltcigs 00:05, 12 March 2020 (UTC)


 * Some comments:
 * I think I generally support linking through a redirect for these identifiers. In addition to the reasons to do so listed above, it would make searching for pages using each identifier easy to identify by using Special:Search and  (which is a fast search) as opposed to   (which is a slower search). I don't find the identified negatives to be sufficiently negative, or negative at all (perhaps not positive).
 * However, I do not support adding these spans with titles. This would significantly increase the HTML associated with each listed identifier (I am thinking of the performance in this case due to a comment elsewhere about our citation/template-heavy articles) for the minor/negligible gain in utility. Users will figure it out eventually.
 * --Izno (talk) 21:16, 11 March 2020 (UTC)

Is there a decision here? Do we attempt to use redirects for identifier label links or do we maintain the status quo? If we choose to use redirects, what form do they take?

—Trappist the monk (talk) 18:36, 17 March 2020 (UTC)
 * Trying to distill the arguments brought forward so far it appears to be that going through a redirect named after an identifier's symbol (in official capitalization, plus "(identifier)") appears to have a slight (long-term) edge over naming them after the expanded form (that is "ISBN (identifier)", "doi (identifier)", "PMID (identifier)", "arXiv (identifier) etc.).
 * I would have suggested to use your span trick to show the expanded form (if available) as tooltip, but other editors seem not to be too fond of that method - perhaps we'll leave that unless requested later on.
 * --Matthiaspaul (talk) 19:11, 17 March 2020 (UTC)

I have tweaked Module:Citation/CS1/Identifiers/sandbox and Module:Citation/CS1/Configuration/sandbox to create identifier label links in this order:
 * when  is
 * from wikidata when the local wiki has  installed and wikidata has an article name for the local language
 * a locally provided article name

I have set the various  to be '&lt;ID> (identifier)' where '&lt;ID>' is the same as  ; for arxiv the label is 'arXiv' and the redirect is 'arXiv (identifier)'.

To be done is to create the redirects. There being no objections, I shall do so.

—Trappist the monk (talk) 14:52, 20 March 2020 (UTC)

As I write this, List of members of the 19th Bundestag is listed at for this error. That error occurs because the live version of the module always attempts to get identifier article titles for the identifier label wikilink from wikidata. The call to  for the Q value specified in the ~/Configuration id handler for the various ids is expensive. Because the order of evaluation described above is redirects, wikidata, locally defined, choosing to use redirects for identifier label wikilinks can avoid that expense. I have tweaked ~/Identifiers/sandbox so that when a redirect is defined and enabled, the module does not make the call to.

—Trappist the monk (talk) 16:35, 21 March 2020 (UTC)

Proposal to enhance edition= parameter to support special numerical symbols
Hi all, I would like to refresh a proposal I made quite a long while ago at Help talk:Citation Style 1/Archive 11:

The edition parameter should be enhanced to support a number of special numerical values ("1".."99") which are not conflictive with the parameter's normal use. If one of these tokens is found, the code would replace this by "1st", "2nd", "3rd".."99th" before passing on the value.

This would help to further decouple semantics (which edition?) from presentation (f.e. "3rd ed."). It would not only make it easier to add common edition information, but also improve readability, maintainability and translatability, and it would allow to centrally change the rendering in the future, would this become necessary ("3rd ed.", "third ed.", "third edition" etc.), depending on the output device (f.e., display the abbreviated form "3rd ed." on the small display of a mobile device, but "third edition" on a desktop or printout), or target language (e.g. "third edition", "dritte Ausgabe", etc.).

This would be very similar to the language parameter which meanwhile accepts free-flow text like "English", "German", etc. but also a number of special symbols like "en", "de" etc.

In order to avoid conflicts, the recognized special tokens should be restricted to just the numbers "1".."99", but this would already cover the common cases.

--Matthiaspaul (talk) 19:44, 13 March 2020 (UTC)
 * I don't think that this is going to happen unless and until MediaWiki CLDR and Scribunto has support for ordinal rendering. For English, it is relatively easy to convert cardinals to ordinals.  But, the rules that apply to English do not necessarily apply to other languages where the cs1|2 module suite is used.  A hint of that complexity may be found on this Unicode CLDR page.
 * —Trappist the monk (talk) 11:11, 21 March 2020 (UTC)
 * Thanks for the pointer. But what I "envision" is much simpler than implementing CLDR. It aims only at covering the majority of cases, and for the remainder the edition can still be used for free text, so there would be no backdraw. Also, my suggestion to cover the range "1".."99" was arbitrary. I think, it could be reduced to "1".."25" or even "1".."15" and still be equally useful. After all, few publications actually go through more editions. In the default implementation, this could be implemented as a simple list of 15 (or 25) language-specific replacement strings, which would need to be adjusted to the local replacements when the citation templates are used in a foreign-language Wikipedia. In locales where the scheme would not be useful at all, the strings could just be left empty (or a dummy routine be implemented), so that no replacement would occur. (At a later stage, more complicated cases could always be covered in a locale-specific routine implementing the local rule-set, but this proposal is not about such a more complicated solution. It is about giving editors a chance to start providing the information symbolically at least in the most common cases as early as possible.)
 * With the citation templates used for millions of citations, being able to encode edition information symbolically in at least the simple cases (and thereby allowing easier adjustment of the output format and translation), would still be a considerable improvement, even if the more complicated cases would not be covered.
 * --Matthiaspaul (talk) 10:47, 22 March 2020 (UTC)

Auto-hyphenation of ISBNs in citations
Cobaltcigs has implemented an auto-hyphenation function for ISBNs based on the official ruleset for hyphenation (see Format ISBN). I think, this functionality should be incorporated into the citation template framework, so that the displayed ISBNs in citations are always properly formatted no matter if they are formatted with or without hyphens in the isbn parameter (and if the hyphens are inserted in the correct locations there). This can be adapted to SBNs and the new sbn parameter as well.

Optionally, there could be a maintenance message in edit preview showing the correct hyphenation if a given ISBN is using no or a different hyphenation in the source code, so that editors could adjust the parameter input accordingly (for cosmetically reasons only, hence this should not be an error message, and only be visible in preview).

--Matthiaspaul (talk) 20:14, 17 March 2020 (UTC)
 * It might be better to fix the hyphenation statically (e.g. with bots), rather than add to the processing done each time a citation is rendered. Kanguole 11:27, 19 March 2020 (UTC)
 * I highly doubt you'll find appetite in the community for a million edits that do little but add hyphens to ISBNs. This is something best handled by templates. &#32; Headbomb {t · c · p · b} 15:12, 19 March 2020 (UTC)
 * So add it as a minor task to CitationBot or something. The above template does a sequential search of a table of 1265 strings.  A binary search would be faster, but it would still be doing a significant amount of processing for every ISBN in every article display, and always producing the same hyphenations, i.e. pointless server load.  Kanguole 15:47, 19 March 2020 (UTC)
 * See WP:PERF. &#32; Headbomb {t · c · p · b} 16:56, 19 March 2020 (UTC)
 * I think you both have good points. The current search implementation definitely could be improved algorithmically. Possibly, this could be interwoven with the checksum validation in order to reduce the number of string scans, but this would need further investigation. I guess, eventually we'd want properly formatted ISBNs also in the article source, so either a human or a bot would have to edit the article anyway.
 * Perhaps, for a start, the properly formatted ISBNs should be displayed only in preview to give editors a chance to copy&paste them back into a citation's source code. And in normal view, the template would pass on whatever it finds given in the source code for performance reasons.
 * Or we could introduce a new parameter no which would be set by bots once they have edited an article. If set to no, this would bypass the hyphenation code, assuming that the bot stored the properly hyphenated string in the citation's source code. Alternatively, properly hyphenated ISBNs could be framed using the ((isbn)) syntax. ( If I remember correctly there also was some "invalid-isbn"-kind parameter for "valid" ISBNs with checksum errors; perhaps this could be combined in order to not introduce yet another parameter - however, right now I can't seem to find this parameter in the documentation. The parameter is ignore-isbn-error .) Or we could introduce a new entry-isbn parameter for not (yet) properly hyphenated ISBNs, invoking the template's auto-hyphenation. Bots would change this to isbn while rewriting the properly hyphenated string.
 * --Matthiaspaul (talk) 18:19, 19 March 2020 (UTC)

So I suppose I could make the linking default to  and also take out the error messages (as cite book does these things internally). This way the (or a) bot would only need to change  to , rather than maintaining its own list of rules. With those changes made, the result for an ISBN already formatted correctly (or one that has no correct formatting because it's numerically invalid) would just be a null edit. ―cobaltcigs 15:04, 22 March 2020 (UTC)
 * But oh, shit: I just remembered subst'ing will fail inside the  tag. This means the bot would also have to change any ref tags containing a   to use   themselves. Something about pre-save transform order of operations. ―cobaltcigs 15:15, 22 March 2020 (UTC)

unhide missing periodical error message
We hid the missing periodical error message as a result of this discussion. Any reason to keep it hidden?

—Trappist the monk (talk) 14:46, 18 March 2020 (UTC)
 * That discussion is too long to comprehend. Please explain why you think it is now appropriate to make this change. Please pay particular attention to whether or not cite web is to be treated as a near-synonym of cite journal and whether it is invalid to use "cite web" for a URL that is not part of a larger publication. Jc3s5h (talk)
 * Likewise for anything that isn't a cite journal/magazine, e.g. cite document, which redirects to cite journal, but which shouldn't require a journal. And cite web which shouldn't require website.&#32; Headbomb {t · c · p · b} 15:34, 18 March 2020 (UTC)
 * It looks like there are currently about 52,000 pages in . I haven't worked on fixing too many of these errors. Most of the errors I have come across have either needed journal or a different template entirely. I have added an explanation of that latter situation to the help text on the category page. In general, I support unhiding of the error message for and . Per the outrage in the discussion,  should remain unaffected. I would like to see feedback from editors who have been working on resolving these errors. – Jonesey95 (talk) 15:40, 18 March 2020 (UTC)
 * Neither nor  have contributed to  since.
 * That renders as a journal is not the fault of the cs1|2 module suite.  The module suite knows only that it has been called by  so it renders what it gets as a journal citation.
 * I attempt to fix these errors as I encounter them but with 52k articles, one editor doing it manually, might get through that category in the current life. I think that I ran Monkbot/task 14 over that category on the off-chance that it might fix some of these errors (where the journal or magazine name is in publisher).  I might attempt that again.
 * —Trappist the monk (talk) 16:15, 18 March 2020 (UTC)
 * 52,000 is not that bad. There are many editors who will fix errors one or two at a time in the articles that they care about, as long as the error messages are visible. If we continue to hide the error messages, though, the count will grow, and only the few editors who frequent this page will work on them. – Jonesey95 (talk) 16:27, 18 March 2020 (UTC)
 * That renders as a journal is not the fault of the cs1 It is, by virtue of invoking cite journal rather than its own thing. &#32; Headbomb {t · c · p · b} 19:09, 18 March 2020 (UTC)
 * Hard to imagine how the module suite could be blamed. Editors decided long ago, that  should redirect to .   was  and has remained as a redirect ever since.  On the redirect's date of creation,  did not use Module:Citation/CS1 – that would not happen for another three years.
 * I found one previous discussion that contemplates creation of a template; discussion fizzled and died:
 * —Trappist the monk (talk) 20:10, 18 March 2020 (UTC)
 * It's not about putting blame on someone or something, but based on its name cite document simply has nothing to do with periodicals, regardless of how it is coded internally.
 * I would support enabling the message for cite journal and cite magazine specifically, because for them the specification of a journal or magazine is natural - and citations lacking this information can be considered incomplete. However, templates like cite web, cite document, or other templates using cite journal internally should not be affected, as most users (including myself) do not consider them to be periodicals.
 * --Matthiaspaul (talk) 20:17, 19 March 2020 (UTC)
 * does not use . Both are independent cs1 templates. ,  are redirects to  so Module:Citation/CS1 sees them as journal cites because it cannot know how  was called.
 * —Trappist the monk (talk) 12:58, 20 March 2020 (UTC)
 * I have run Monkbot/task 14 over . The bot made 9521 edits and reduced the category size by maybe 1000 articles.  I think it is time to show the missing periodical error messages.
 * —Trappist the monk (talk) 12:58, 20 March 2020 (UTC)
 * Thanks for that bot run, and I agree that the messages should be displayed. I fixed a dozen or so, and they weren't difficult. There are a lot of cite journal citations with DOI values and titles but no journal name; those should be easy to fix. Once we get rid of the low-hanging fruit, we may find after all that there is a need for a slightly different cite document or something like it for citing standalone non-book publications, perhaps just a wrapper that calls citation with cs1 (I just made that up without any testing, so it is probably not right). That should be a separate discussion from this one. – Jonesey95 (talk) 15:08, 20 March 2020 (UTC)
 * This should not be enabled until there is a way to ensure that only a very specific whitelist of templates emit the error. Namely only cite journal and cite magazine (and journal/magazine-related redirects to those templates), and not cite document and other unrelated redirects. &#32; Headbomb {t · c · p · b} 15:14, 20 March 2020 (UTC)
 * cite document is a redirect to cite journal and has been since its creation ten years ago. I don't know of such a thing in Wikipedia as an "unrelated redirect"; do you have a link explaining what that is? Anything that happens to cite journal will happen when it is called via cite document. If you want cite document to be a different template with different rules, please start a new discussion about that new topic, as I suggested above. – Jonesey95 (talk) 17:01, 20 March 2020 (UTC)
 * A document is not a journal. Therefore the two are unrelated. This should be obvious. I can create a redirect from cite database to cite journal, but that won't magically make a database a journal, not a citation to a database a citation to a journal. &#32; Headbomb {t · c · p · b} 17:09, 20 March 2020 (UTC)
 * Please create a new discussion thread and explain to us how cite document should work as a standalone CS1 template. Please provide real examples from real articles that support your proposal to create a new CS1 template. I think that I will agree with most of what you have to say, because as I look at the articles in this error category, I see some citations that do not appear to fit any of our existing templates except for the catch-all citation. Talking about this issue in this existing thread, which is about specific error messages, will make it less likely that you (and I, and other editors who will share our concern after the error messages appear in public) get what we want. – Jonesey95 (talk) 17:17, 20 March 2020 (UTC)
 * Explain to us how cite document should work as a standalone CS1 template. Simple, exactly like it currently does, not emitting an error if journal is missing. There's no need for a separate thread for this. &#32; Headbomb {t · c · p · b} 17:20, 20 March 2020 (UTC)
 * My preference would be for it to work exactly like citation does (agnostic as to which kind of citation it is citing, book, journal, web, whatever) except cs1 not cs2 by default. —David Eppstein (talk) 18:50, 20 March 2020 (UTC)
 * That would also be an option. &#32; Headbomb {t · c · p · b} 20:18, 22 March 2020 (UTC)
 * Please create a new discussion thread and explain to us how cite document should work as a standalone CS1 template. Please provide real examples from real articles that support your proposal to create a new CS1 template. I think that I will agree with most of what you have to say, because as I look at the articles in this error category, I see some citations that do not appear to fit any of our existing templates except for the catch-all citation. Talking about this issue in this existing thread, which is about specific error messages, will make it less likely that you (and I, and other editors who will share our concern after the error messages appear in public) get what we want. – Jonesey95 (talk) 17:17, 20 March 2020 (UTC)
 * Explain to us how cite document should work as a standalone CS1 template. Simple, exactly like it currently does, not emitting an error if journal is missing. There's no need for a separate thread for this. &#32; Headbomb {t · c · p · b} 17:20, 20 March 2020 (UTC)
 * My preference would be for it to work exactly like citation does (agnostic as to which kind of citation it is citing, book, journal, web, whatever) except cs1 not cs2 by default. —David Eppstein (talk) 18:50, 20 March 2020 (UTC)
 * That would also be an option. &#32; Headbomb {t · c · p · b} 20:18, 22 March 2020 (UTC)

cite encyclopedia without |title=, part 2
This has already been discussed.

However: there is e.g. Zagrebački leksikon, an encyclopedia citation template. It is meant to be used both to cite individual articles, and to cite the entire encyclopedia in the Bibliography section and have shortened footnotes for the articles, like in Timeline of Zagreb. Both of these uses are quite legitimate. Since title param is now mandatory, I don't see a way of making the suggested "solution" work without applying some rather ugly hacks to the Zagrebački leksikon template. These hacks would also be unnecessary because in reality title is no more "mandatory" in cite encyclopedia than page is mandatory in cite book.

My suggestion would therefore be to make the title optional. GregorB (talk) 14:05, 22 March 2020 (UTC)
 * The real solution would be to provide support for none in more than just cite journal &#32; Headbomb {t · c · p · b} 14:57, 22 March 2020 (UTC)
 * applying some rather ugly hacks to the Zagrebački leksikon template, change this:
 * to this:
 * Not really so ugly and actually quite common practice in wrapper templates.
 * Alternately, you can do this:
 * —Trappist the monk (talk) 15:29, 22 March 2020 (UTC)
 * Thenone solution is not bad, as a way of specifying that the title was intentionally omitted.
 * The above is precisely the "ugly hack" I was referring to. Its ugliness is not in its complexity, it's in the semantic gymnastics with the params, and the fact it's quite unnecessary, which is the thrust of my argument here: it makes no sense to work around something being mandatory when that something should not be mandatory in the first place. GregorB (talk) 23:20, 22 March 2020 (UTC)
 * —Trappist the monk (talk) 15:29, 22 March 2020 (UTC)
 * Thenone solution is not bad, as a way of specifying that the title was intentionally omitted.
 * The above is precisely the "ugly hack" I was referring to. Its ugliness is not in its complexity, it's in the semantic gymnastics with the params, and the fact it's quite unnecessary, which is the thrust of my argument here: it makes no sense to work around something being mandatory when that something should not be mandatory in the first place. GregorB (talk) 23:20, 22 March 2020 (UTC)

SSRN limit now exceeded


The limit should be increased to at least 4000000. &#32; Headbomb {t · c · p · b} 18:55, 10 March 2020 (UTC)


 * It seems pointless to update it once every few months, breaking new citations for weeks while everyone ignores the error. Why have an upper limit at all? -- Tim Starling (talk) 10:03, 24 March 2020 (UTC)
 * The purpose for the limit is to catch simple typographic errors. Yeah, it's a weak test but since there isn't a check-digit or some other formatting cue, limiting the value has been our only way of discovering erroneous values.  If you have a better solution, please tell us what that solution is.  Simply raising the limit to a huge number merely hides invalid identifier numbers that have been used in a citation but have not yet been issued.  I will revert your edit at Module:Citation/CS1/Configuration/sandbox.
 * —Trappist the monk (talk) 11:53, 24 March 2020 (UTC)
 * I only raised the sandbox limit to 10M, sufficient to catch an error in the number of digits, I didn't raise it to a "huge number". Your change reduces it to 4M, which is only sufficient to catch typos in the first digit if the typo happens to change the digit to a number greater than 3, and so small that it will probably start throwing errors again within a year. The problem we have is that every citation of an SSRN published later than about February is showing a "check SSRN" error. The problem has been known since early March. The main template hasn't been synced since January, and if I test the current sandbox, it fails with a Lua error, so it can't be raised at all. -- Tim Starling (talk) 22:34, 24 March 2020 (UTC)
 * While the module should have been updated eons ago, the SSRN corpus does not increase at a rate of 500,000 submissions per year. And it doesn't only catch errors on the first digit, if you have extra digits, it will catch that too. &#32; Headbomb {t · c · p · b} 22:47, 24 March 2020 (UTC)
 * ID 3500012 was 29 December, and ID 3545710 was 25 March, which is a rate of about 525 papers per day, 192,000 per year. So we can expect it to exceed 4M in August 2022, if it continues to grow at the same rate. A little later than my guesstimate but I still think the cure is worse than the disease. Presumably most "check SSRN" errors to date have been caused by inappropriate limits, not by typos. The trouble is that nobody is updating these limits. It's understandable: I probably have spent about an hour on this already, and I probably still have an hour of work left if I want to do the update. If we change the limit to 10M, then we will next have to deal with this problem in about 33 years, which sounds good to me since that's about when I will next have time for this. -- Tim Starling (talk) 03:54, 25 March 2020 (UTC)

Problem in cite journal
I had trouble adding article-url and article-url-access to a cite in the Cuban Missile Crisis article as follows:



I left it with url=https://www.usni.org/magazines/proceedings/1986/april-supplement/fighting-unwinnable-wars instead. Wtmitchell (talk) (earlier Boracay Bill) 08:39, 25 March 2020 (UTC)
 * article-url is an alias of chapter-url. As such it is for use with cs1|2 templates that accept chapter.  What you want, I think is:
 * —Trappist the monk (talk) 11:47, 25 March 2020 (UTC)
 * —Trappist the monk (talk) 11:47, 25 March 2020 (UTC)
 * —Trappist the monk (talk) 11:47, 25 March 2020 (UTC)

Position of date
Consider:



Why does the position of the date in the last example change according to whether or not author-mask1 is present? Andy Mabbett ( Pigsonthewing ); Talk to Andy; Andy's edits 14:08, 22 March 2020 (UTC)
 * This has been discussed before. I can't find the discussion. I think there was more-or-less an agreement to use a similar date position regardless of whether the lowest-level work has an author or not. (The lowest-level work could be an article, dictionary entry, television episode, chapter, book, etc. depending on the nature of what is being cited.)
 * However the templates were never edited to reflect the results of the discussion. Jc3s5h (talk) 14:47, 22 March 2020 (UTC)
 * Date position is toward the head of the citation when there is a displayed name (author, editor). In your last example, you set 0 so there is no name to display.  As far as the rendering is concerned, there is no author.  At sometime in the distant past, editors developing the original cs1|2 templates determined that publication dates were at the front when an author/editor name is displayed and towards the end when no names are displayed – presumably so that date isn't the first element displayed in a rendered citation.
 * The discussion that Editor Jc3s5h mentioned is.
 * —Trappist the monk (talk) 15:05, 22 March 2020 (UTC)
 * The discussion that Editor Jc3s5h mentioned is.
 * —Trappist the monk (talk) 15:05, 22 March 2020 (UTC)


 * I believe the original impetus for the placement of dates was consistency with parenthetical referencing, which uses Author-Date or Author-Title. So that when following a short reference, readers would immediately recognize the full reference as its expansion. In contrast, in short-refing uncredited material in larger works, the norm has been to include the enclosing source in the place of the missing author, as in JournalTitle-Date (generally, Source-Date). However in the relevant full citation the leading element is Title, which does not appear in the short. So it was thought a good practice that somewhere in the full citation Source and Date appear close together. 98.0.246.242 (talk) 02:27, 24 March 2020 (UTC)
 * The issue described by 98.0.246.242 could be resolved by following the advice of printed stye guides: when there is no author, use a shortened title in the footnote. Although a machine will not recognize the long and short titles as referring to the same article/work/etc., a human will. For example,


 * Jc3s5h (talk) 12:31, 25 March 2020 (UTC)

tag "website"
Hi, everybody, i find it very confusing that the tag is named "website", but you cannot give a website-url there. If i quote with cite web, i do have a tag "url=" where i do enter the url of the article and then there is "website", where i normally would enter the main website, for instance: the url is "www.blablabla.com/specialtextonthisauthor.html - then i would enter this in the "url"-place and under "website" would say "www.blablabla.com - but this doesn't work. Shouldn't one name the tag "website" then "title of the website" instead? Kind regards and stay safe, --Gyanda (talk) 20:06, 29 March 2020 (UTC)
 * It is for the name or title of the overall site, not the address and not the title of the individual web page. If you find it confusing, you can always use the synonym work for the name. —David Eppstein (talk) 21:21, 29 March 2020 (UTC)
 * It took me quite a while to understand that, David. Thanks for the explanation. I will try "work". Kind regards, --Gyanda (talk) 23:16, 29 March 2020 (UTC)

How to create a footnote with both date and year as a ref=harv citation
How can I create a footnote with both date and year as a ref=harv citation? I mainly trying to do this for a bunch of newspaper articles from the same month and don't want to keep using 1921a, 1921b, 1921c, so on. I want the footnote to be NEWSPAPER 19 Apr 1921, NEWSPAPER 9 Apr 1921 and so on in Harvard citation format. I've asked this on another desk before but forgot where it is located. KAVEBEAR (talk) 21:49, 28 March 2020 (UTC)


 * I would have thought this would work, but it doesn't generate an ID. Surprisingly it also doesn't display an error for the date (at least the COINS isn't corrupted).
 * --Izno (talk) 22:00, 28 March 2020 (UTC)
 * Anchor ID is not created because there are no contributer, no author, and no editor names. cs1|2 requires names; just a date is pretty meaningless.  You can do:
 * —Trappist the monk (talk) 22:19, 28 March 2020 (UTC)
 * Ah, I don't use this ever. Would it not be sensible when harv to emit an error when at least one of the required pieces is not present? --Izno (talk) 22:59, 28 March 2020 (UTC)
 * Probably not, mostly because many just put harv by reflex, without intended to use it for generating anchors. It would make a lot of sense as a preview message though. &#32; Headbomb {t · c · p · b} 23:40, 28 March 2020 (UTC)
 * Perhaps this:
 * with:
 * —Trappist the monk (talk) 22:19, 28 March 2020 (UTC)
 * —Trappist the monk (talk) 22:19, 28 March 2020 (UTC)
 * Ah, I don't use this ever. Would it not be sensible when harv to emit an error when at least one of the required pieces is not present? --Izno (talk) 22:59, 28 March 2020 (UTC)
 * Probably not, mostly because many just put harv by reflex, without intended to use it for generating anchors. It would make a lot of sense as a preview message though. &#32; Headbomb {t · c · p · b} 23:40, 28 March 2020 (UTC)
 * Perhaps this:
 * with:
 * —Trappist the monk (talk) 22:19, 28 March 2020 (UTC)
 * with:
 * —Trappist the monk (talk) 22:19, 28 March 2020 (UTC)
 * with:
 * —Trappist the monk (talk) 22:19, 28 March 2020 (UTC)
 * —Trappist the monk (talk) 22:19, 28 March 2020 (UTC)
 * —Trappist the monk (talk) 22:19, 28 March 2020 (UTC)
 * —Trappist the monk (talk) 22:19, 28 March 2020 (UTC)
 * —Trappist the monk (talk) 22:19, 28 March 2020 (UTC)


 * I wouldn't. Short references should be that. The subnotation  is confusing. One could assume that there is another citation from the same source with the same date. Using the serial subnotation with the proper timeframe reference (  etc. or   for example) lets the reader know that there are several similar citations in the referenced period. 98.0.246.242 (talk) 23:43, 28 March 2020 (UTC)
 * Trappist's way is what I have done in the past when the auto-generated CITEREF (aka harvid) is not appropriate or not working. – Jonesey95 (talk) 00:33, 29 March 2020 (UTC)
 * Are there any existing articles that uses something like this. Jonesey95, you mentioned using it in the past. In what article? KAVEBEAR (talk) 16:44, 29 March 2020 (UTC)
 * There is nothing wrong with using } or . But the norm in  references has been to use Author-Date (meaning year or rarely month year) or Author-Title. This has to do with the way the full reference is presented, and also indexing comes into play. Most databases main-index by author, subindexing by date and/or title. And they additionally index by source, again subindexing by title and/or date. Granted that this is not universal and it is becoming less so with new techniques such as on-demand indexing etc. But it is fair to assume that one will find a work faster by searching these fields. It is (or used to be until yesterday) harder to find a work by date/title combination. Short references reflect all that. Emphasis on . 98.0.246.242 (talk) 23:36, 29 March 2020 (UTC)
 * I make a lot of little edits, so I'd really have to dig to find one that matches this description, but here is one where I used harvid to differentiate two sources with the same author and year. It's not perfect, but I was only trying to solve the immediate no-link problem, not try to perfect the article. – Jonesey95 (talk) 17:27, 29 March 2020 (UTC)
 * Benjamin Hope uses this custom format quite a bit. – Jonesey95 (talk) 17:10, 30 March 2020 (UTC)

I would try to stay close to what printed style guides like Chicago Manual of Style do so as not to surprise the reader or other editors. Lets start with what the reader will see in the bibliography. KAVEBEAR didn't mention an author for either story, but newspaper stories almost always have a headline. So the first element in the citation, as rendered, will be the article title, and that is what will determine placement in the bibliography. In the case of a duplicate title, the tie will be broken by the next element, the name of the newspaper. But this is the same. So the title will be broken by the next element, the volume of the newspaper. This logic is abominable. So I would do one of two things: Jc3s5h (talk) 13:36, 29 March 2020 (UTC)
 * 1) Let the article rot until a consistent order of citation elements to be implemented as requested at
 * 2) Rip out all the citation templates and use some other citation style that is more appropriate for the article.
 * Can you sort by titles instead? “Obituary...“ 1918 / “Obituary of John Johnson” 1918 or ”Funeral...” 1918 / “Funeral of John Johnson” 1918. KAVEBEAR (talk) 16:40, 29 March 2020 (UTC)
 * As you probably know, bibliographies are sorted manually. I would sort by the first rendered element. For most publications, this is/are the author(s). If there is no author, then the first element is the title of the lowest-level work, which would be the name of the article for a newspaper or journal. But what if the same newspaper re-uses the same headline several days in a row, and doesn't give an author? The next different element in the citation would be the issue number, which is often omitted from a newspaper story. That makes it a mess. It would be much better to render the date immediately after the article title, and use it to break ties. Jc3s5h (talk) 18:12, 29 March 2020 (UTC)

Error in date of cite newgroup example
Wait! I fixed it myself :) See: https://en.wikipedia.org/w/index.php?title=Template%3ACite_newsgroup%2Fdoc&type=revision&diff=948490292&oldid=945888694

Removed spurious comma from:

to:

- Here's what I was going to ask :)

In Cite newsgroup#vertical format the parameter access-date has an error. I believe the cite newgroup documentation uses an automatically generated date. Unfortunately there is a spurious comma(',') contained. Perhaps this from a previous change from MMMMM DD, YYYY format to DD MMMMM YYYY format.

Template:Cite newsgroup The vertical format example

instead of:

The horizontal format example above it uses the alternate, but correct, date.

Thanks! Lent (talk) 09:31, 1 April 2020 (UTC)

Where can I find the range of value |pmc
Hello, I have a question. Where can I find the range of value |pmc. I want to know because my Vietnamese wiki has a little issue about this cs1. I want to enlarge the range of value: from 1 - 6000000 to 1 - 8000000. I'm looking forward to hearing from you. Đư&#39;c (talk) 16:12, 1 April 2020 (UTC)
 * vi:Mô_đun:Citation/CS1/Identifiers
 * —Trappist the monk (talk) 16:27, 1 April 2020 (UTC)

HDL parameter escaping
The citation templates encode the ? characters in Handles which breaks them. It is better than the HDL template that simply displays nothing.

HDL template:

HDL parameter:

URL parameter:

AManWithNoPlan (talk) 17:06, 3 April 2020 (UTC)
 * I don't think that cs1|2 should display proxy server query parameter that are part of the hdl. I think that we should probably intercept the query string, encode the hdl url and then reattach the query string and apply that url to the handle portion of the hdl.  There are several query string parameters that we will need to intercept.  See hdl doc.  I'll think about how to do this.
 * —Trappist the monk (talk) 18:01, 3 April 2020 (UTC)
 * You are correct, not only does the ? get escapes and break the link, but the display is ugly. AManWithNoPlan (talk) 19:32, 3 April 2020 (UTC)

—Trappist the monk (talk) 22:38, 3 April 2020 (UTC)

Who is the publisher?
The article I came across with Bare URLS for citations was Charles Henry Elliott-Smith. Lots of sources came from https://www.thegazette.co.uk/London/issue/x/supplement/x/data.pdf, and I used cite news for them since automatic citations wouldn't work. Some of them never stated the publisher, another stated it was published by HIS MAJESTY'S STATIONERY OFFICE, and I think the publisher is The London Gazette. So who is the publisher? &#123;&#123;replyto&#125;&#125; Can I Log In's (talk) page 00:14, 4 April 2020 (UTC)
 * The London Gazette is published by The Stationery Office (it says in the linked articles). Or Her Majesty's Stationery Office, before 1996. —David Eppstein (talk) 00:17, 4 April 2020 (UTC)
 * There is a template for that: :
 * Because it's a journal of record, no publisher is necessary.
 * —Trappist the monk (talk) 00:24, 4 April 2020 (UTC)
 * Because it's a journal of record, no publisher is necessary.
 * —Trappist the monk (talk) 00:24, 4 April 2020 (UTC)
 * —Trappist the monk (talk) 00:24, 4 April 2020 (UTC)

Tracking for bad ref
Do we have tracking for problems like this one? In most cases, this would appear as a missing template, so could possibly string match the first part of the entry to find them. For example,  returns 1, but   returns 0. Thanks! Plastikspork ―Œ (talk) 16:46, 31 March 2020 (UTC)
 * No. In this citation:
 * cs1|2 gets  from ref.  Because that value is not , cs1|2 uses it as is.  Before the value becomes the   attribute of the citation's wrapping  tag, it is anchor encoded which changes it to.
 * cs1|2 presumes that whatever is in ref is a correct value.
 * —Trappist the monk (talk) 18:32, 31 March 2020 (UTC)
 * Granted that an editor who uses ref is likely not a beginner and should be careful not to nest or enter template notation in this field unless they know exactly what they are doing... so is it prudent (and easily doable) to limit entries to text,  and  ? Or is this an undue restriction? 98.0.246.242 (talk) 23:56, 31 March 2020 (UTC)
 * I would think we could at least track unusual cases (not  or   or plain text).
 * I believe we have tracking for "archived copy" in the title, but what about ACTUAL ARTICLE TITLE BELONGS HERE. Some script happy editor is blindly changing article titles to this rather than leaving it blank to make a visible error.  See, for example, this search which returns articles like Brian Newman.  Plastikspork ―Œ (talk)  01:57, 6 April 2020 (UTC)
 * This one does not time out and accordingly catches more. I left a note on Tony1's talk page to have that behavior stopped. --Izno (talk) 02:39, 6 April 2020 (UTC)
 * I would think we could at least track unusual cases (not  or   or plain text).
 * I believe we have tracking for "archived copy" in the title, but what about ACTUAL ARTICLE TITLE BELONGS HERE. Some script happy editor is blindly changing article titles to this rather than leaving it blank to make a visible error.  See, for example, this search which returns articles like Brian Newman.  Plastikspork ―Œ (talk)  01:57, 6 April 2020 (UTC)
 * This one does not time out and accordingly catches more. I left a note on Tony1's talk page to have that behavior stopped. --Izno (talk) 02:39, 6 April 2020 (UTC)

Should this template always require a title?
If you are using sfns to multiple articles in the same encyclopedia are you supposed to have a separate citation to each article title? The title= error message is annoying. —DIYeditor (talk) 07:36, 5 April 2020 (UTC)
 * You are referring to this source? Schaus is the editor.  The entries in that work are individually authored so  may be useful:
 * This:  →  links to the  which links to Schaus:
 * In, change encyclopedia to title.
 * —Trappist the monk (talk) 11:34, 5 April 2020 (UTC)
 * Thank you, yes that is the one. Really appreciate you looking into it. —DIYeditor (talk) 09:18, 6 April 2020 (UTC)
 * —Trappist the monk (talk) 11:34, 5 April 2020 (UTC)
 * Thank you, yes that is the one. Really appreciate you looking into it. —DIYeditor (talk) 09:18, 6 April 2020 (UTC)

Issue numbers with a space after the comma (e.g., "9, 557")
cite news appears to place an extra space after the comma in issue numbers that are four or more digits long. For example, the following citations taken from East Knoyle War Memorial and Herbert Maryon all have this problem:



Does anyone have an idea of how to fix this? I see that the issue has previously been noted, but less productively. --Usernameunique (talk) 15:13, 6 April 2020 (UTC)


 * The previously noted discussion contains a link to a found this discussion wherein lies the answer:
 * See.
 * I'm pretty sure that url shorteners are discouraged because readers can't know where that link is really taking them and that is a security risk. This one says   but the reader ends up at a Columbia University password protected barrier.  Find a better source?
 * —Trappist the monk (talk) 19:23, 6 April 2020 (UTC)
 * I'm pretty sure that url shorteners are discouraged because readers can't know where that link is really taking them and that is a security risk. This one says   but the reader ends up at a Columbia University password protected barrier.  Find a better source?
 * —Trappist the monk (talk) 19:23, 6 April 2020 (UTC)
 * —Trappist the monk (talk) 19:23, 6 April 2020 (UTC)


 * Thanks for extracting that information, . It would be nice to have a more elegant solution, but at least it works for now. And yes, Gale's links are unfortunate; even their unshortened links are unwieldy, little more informative, and just as redirect prone. They're also (as far as I can tell) always tied to an institution, so you can't link to a log-in-with-whatever-your-institution's-credentials-are page. But assuming that 1920 newspaper is out of copyright, and there's somewhere else I could download and place it, I'd be happy to do so. --Usernameunique (talk) 03:41, 7 April 2020 (UTC)

bad_paramlink error message is wrong when it occurs to author parameter
In short, this markup produces a weird error message. This most likely comes from line 1406.--ネイ (talk) 08:20, 30 March 2020 (UTC)
 * Markup:
 * Error message:
 * Just to add - I mean "author-last1" is wrong. The input markup itself is certainly incorrect.--ネイ (talk) 08:22, 30 March 2020 (UTC)

Fixed, I think, in the sandbox:

—Trappist the monk (talk) 10:58, 30 March 2020 (UTC)

And, because I found this at Clement Attlee, check firstn for conflicting wikilinks:

—Trappist the monk (talk) 14:39, 4 April 2020 (UTC)
 * Thank you very much for the fix.--ネイ (talk) 05:37, 7 April 2020 (UTC)

bad url error message fix
—Trappist the monk (talk) 23:22, 8 April 2020 (UTC)

preprint parameter validation tweaks
For the preprint templates, I have adopted the mechanism that supports templates with unique parameters. This allows some streamlining of the validation code. These are all using that mechanism (first four should show no errors: To these four, I added Publisher, a parameter that is not supported in the limited parameter set: —Trappist the monk (talk) 23:53, 8 April 2020 (UTC)


 * It would be good if was also implemented while we're at it. &#32; Headbomb {t · c · p · b} 00:01, 9 April 2020 (UTC)

language_parameter function ignores parameter-separator configuration
This is regarding Module:Citation/CS1, lines 1545–1550. In this function, the display of multiple language names honors parameter-pair-separator and parameter-final-separator but ignores parameter-separator (and always uses &lt;comma>&lt;space>). I am trying to fix it in jawiki (where we set all three to "、") but not sure whether replacing &lt;comma>&lt;space> directly with parameter-separator would cause any problems in the gsub call immediately after. It would be great if this could be fixed here, so we can directly import into jawiki.--ネイ (talk) 05:47, 7 April 2020 (UTC)
 * To provide a sample: de, fr, pt. In jawiki this is displayed as "ドイツ語フランス語、ポルトガル語" and we want it to be "ドイツ語フランス語、ポルトガル語".--ネイ (talk) 05:52, 7 April 2020 (UTC)
 * Try this:
 * Report back.
 * —Trappist the monk (talk) 13:45, 7 April 2020 (UTC)
 * Thank you very much. This works perfectly, and I do not think there are any legitimate cases for  to appear in the language parameter input. If this could be added to the sandbox and synced later on, it would be great. (I do not prefer copying from a talk page or a sandbox page to our production module in another language)--ネイ (talk) 06:37, 8 April 2020 (UTC)
 * Better way that doesn't rely on the content of the items in the table:
 * This, as was the previous attempt, is in the en.wiki sandbox.
 * —Trappist the monk (talk) 11:13, 8 April 2020 (UTC)
 * Thank you - this also works and is cleaner than the previous fix. We will introduce this fix in jawiki once the change is synced from the sandbox to the module.--ネイ (talk) 14:05, 9 April 2020 (UTC)
 * —Trappist the monk (talk) 11:13, 8 April 2020 (UTC)
 * Thank you - this also works and is cleaner than the previous fix. We will introduce this fix in jawiki once the change is synced from the sandbox to the module.--ネイ (talk) 14:05, 9 April 2020 (UTC)

biorxiv appearance
In the past, the bioRxiv DOI was concatenated from to  with the understanding that 123456 was the identifier (and indeed was used in many URLs). However, the recent update to the biorxiv DOIs would mean this changes from to  which really isn't clear, further obfuscates what is actually used by people, and loses the 'identity' of the biorxiv string as an identifier/pseudoidentifier. So I propose we show what should be clear to everyone


 * bioRxiv:10.1101/123456
 * bioRxiv:10.1101/2020.02.07.937862

A simple AWB run (or run) should be more than sufficient to make the updates from 123456 to 10.1101/123456 in a timely manner. &#32; Headbomb {t · c · p · b} 18:40, 17 March 2020 (UTC)

—Trappist the monk (talk) 15:54, 9 April 2020 (UTC)

Check author names are not numeric for Cite web
Some websites provide incorrect information, and converting links to references with Citoid can result in invalid last or first names which are not currently flagged as errors by EXTRACT NAMES. Eg last2 = 2012 which is a year, or first = 11 January (a date part). I am unsure if it is ever valid for a last oe first name to contain digits, or have only digits. I suggest checking the content of author names and using the maintenance category to flag invalid ones. Amousey (talk) 21:14, 6 April 2020 (UTC)
 * I suspect that there are legitimate names that contain digits. However, dates, telephone numbers, time, email addresses, urls, page numbers, and the like are not.  Unfortunately, these things are often mixed with valid names; citoid seems to make a habit of including birth and/or death dates for authors in citations that it writes.
 * I have an alternate proposal. When an author / editor / contributor / translator / others (and any of their aliases) are assigned a value that is composed solely of digits or digits with punctuation, that constitutes an error; when there are letters in the parameter value, that is probably an error but shall be treated as a maint issue.  This latter case covers the legitimate name-that-includes-digits.  Parsing apart the legitimate from the absurd is not what cs1|2 should be doing.
 * Have you, or someone else, raised this issue at phabricator? If not, please do.  Creating error and maint categories that citoid will just fill and that human editor will have to struggle to keep empty is a pointless waste of editors' time, energy, and resources.
 * —Trappist the monk (talk) 22:15, 6 April 2020 (UTC)
 * I would prefer to see this filed at Phabricator as a feature request for Citoid. – Jonesey95 (talk) 23:56, 6 April 2020 (UTC)
 * Not sure that I understand what you mean by that. Surely citoid won't be fixing stuff that is already broken and already exists.  I don't see citoid stripping birth/death dates from author/editor names before placing them in a citation template as a feature request.  Instead, that is a rather important bug fix.
 * —Trappist the monk (talk) 00:11, 7 April 2020 (UTC)
 * Sorry to be unclear. I was trying to agree with you that Citoid should not add years to author parameter values, and that submitting a request via Phabricator to fix that problem is the best way to fix this problem long-term. – Jonesey95 (talk) 15:10, 7 April 2020 (UTC)
 * For web sources, Citoid is built to some degree on replicating an understanding of the page you are citing. If the source website has changed, then that replication may need to be changed as well. In this case, you will need to identify the website and configuration at issue and correct those; that starts probably with Phabricator.
 * For some sources (particularly OCLC), that's an artifact of the export between OCLC and Citoid. This source requires changes in Citoid I think. (And this concern reminds me strongly of other issues that are already documented with OCLC citations.)
 * Either way, I think Phabricator is the correct first stop. There might still be some value in a maintenance category looking for current issues. --Izno (talk) 00:56, 7 April 2020 (UTC)
 * For some sources (particularly OCLC), that's an artifact of the export between OCLC and Citoid. This source requires changes in Citoid I think. (And this concern reminds me strongly of other issues that are already documented with OCLC citations.)
 * Either way, I think Phabricator is the correct first stop. There might still be some value in a maintenance category looking for current issues. --Izno (talk) 00:56, 7 April 2020 (UTC)

Tweaks to the sandbox removed some nonsense and added maint cats when a name-holding parameter does not hold any letters:

Alphanumeric names not caught:

This change will create 5 new maint cats.

—Trappist the monk (talk) 22:20, 9 April 2020 (UTC)

translating cite web
Hi! Could I get some help on translating cite web to and from other languages in WP:CXT. When translating into Scots, for example, the cite web references come up as grayed out. The "issues" gives:

"Missing reference: A reference could not be transferred to the translation since it uses a template with a different structure. Please, edit the reference in the translation to fill the missing information." It then gives a link to Content translations/Templates. However, the template is identical for both languages.

How do I go about making these translate properly through the translation tool? Any ideas? I previously asked at WT:CXT. Best Wishes,  Lee Vilenski (talk • contribs) 07:59, 10 April 2020 (UTC)
 * I don't know anything about WP:CXT; never used it, probably never will. If the template is identical for both languages, is there really an issue?  Wouldn't a translator simply copy the template rather than attempt to translate it?  Are you trying to translate cs1|2 templates that are inside  tags?  If so, then see Content translation/Templates.
 * —Trappist the monk (talk) 11:44, 10 April 2020 (UTC)

module suite update 18–19 April 2020
I propose to update cs1|2 module suite over the weekend 18–19 April 2020. Here are the changes:

Module:Citation/CS1:
 * i18n; definitions moved to ~/Configuration; discussion
 * use css to mute bold self-links; discussion
 * allow ietf tags with 3-char lang codes; discussion
 * support for  character entity in page number lists; discussion
 * patch wikisource icon bug; discussion
 * fix author & authorlink bad param error messages; discussion
 * all cs1|2 templates make anchor IDs when there are names; discussion
 * i18n: tweak language parameter list rendering; discussion
 * refine handling of parameters unique to specific templates; discussion
 * bad url error message fix; discussion
 * tweak preprint template parameter validation; discussion
 * maint cats for name-holding parameters with only digits and punctuation as values; discussion
 * i18n add  keyword to ~/Configuration; discussion

Module:Citation/CS1/Configuration:
 * i18n; definitions moved into ~/Configuration
 * moved limit values for identifiers that have them from ~/Identifiers
 * add detection of nbsp invisible character; discussion
 * add support for |sbn=; discussion
 * all cs1|2 templates make anchor IDs when there are names
 * standardize identifier param case order; discussion
 * refine handling of parameters unique to specific templates
 * i18n add  keyword

Module:Citation/CS1/Whitelist:
 * add support for |s2cid=; discussion
 * add support for |sbn=
 * refine handling of parameters unique to specific templates
 * tweak preprint template parameter validation

Module:Citation/CS1/Date validation:
 * Change mdy check from " *" to " +" after day number in order to catch bad dates like "26November 2019"; discussion

Module:Citation/CS1/Identifiers:
 * i18n; definitions moved to ~/Configuration
 * support new biorxiv id format; discussion
 * add support for |s2cid=
 * add support for |sbn=
 * suppress display of query strings in |hdl=; discussion
 * require complete biorxiv identifier; discussion

Module:Citation/CS1/styles.css: —Trappist the monk (talk) 19:36, 12 April 2020 (UTC)
 * use css to mute bold self-links
 * support high-resolution icons; discussion
 * Not specifically relevant here, but the proposition to apply the updates on workdays for English-speaking countries has some merit imo (it was proposed following the update debacle of some months ago). It should be considered futher. 50.74.165.202 (talk) 14:19, 13 April 2020 (UTC)
 * I remember that comment, but cannot find it. Can you?
 * —Trappist the monk (talk) 16:29, 13 April 2020 (UTC)
 * Guess one can slog through the archives, will take some time for it. It was a comment by an (admin?) editor who pointed out among other things that there is a better chance to catch errors, incompatibilities, mistakes, etc. during weekdays because that is when most admins would be online and presumably on wiki. 98.0.246.242 (talk) 01:34, 14 April 2020 (UTC)
 * It was the 2020-02-29 6.14 comment by sysadmin cdanis. Nemo 08:53, 14 April 2020 (UTC)
 * Thanks. 98.0.246.242 (talk) 17:50, 14 April 2020 (UTC)

i18n: keyword none
An oversight on my part,  is a keyword used in postscript to suppress terminal punctuation, in ref to suppress creation of an anchor ID, in title ( and  when journal has a value) to suppress  errors, in type to suppress automatic type annotation







—Trappist the monk (talk) 18:07, 10 April 2020 (UTC)
 * I suppose I should say what it is that I did. Added   to the   table in Module:Citation/CS1/Configuration/sandbox.  Module:Citation/CS1/sandbox uses the value in the table for the listed parameters.  This allows other wikis to specify 'none' in their language.
 * —Trappist the monk (talk) 18:11, 10 April 2020 (UTC)
 * Perhaps 'none' should also be added as a possible value for some of the author* parameters in order to indicate that a source does not specify an author. So far, the suggestion is to add a hidden HTML comment like, but since the exact wording is difficult to remember, this is not easily machine-readable. It would be better if the template would allow this condition to be specified explicitly through some symbolic value like 'none' so that we could issue a specific (centrally defined and maintained) placeholder text would we decide that this is more useful than not displaying anything at all in the future (or there could even be some alternative output format using CSS to make the display of such messages user-configurable).
 * I guess, there are no real authors named 'none', so this would not be in conflict with normal use.
 * --Matthiaspaul (talk) 19:33, 14 April 2020 (UTC)

reserved character '#' in enumerated parameter names
Because of this discussion, I've been having a go at rewriting Template:Citation Style documentation/author using the cs1 doc support code that I wrote to extract canonical and alias names from Module:Citation/CS1/Configuration. I have already done that for Template:Citation Style documentation/language but ~/author is a bit more difficult.

The doc support function function  returns  with a '#' enumerator place-holder character. I don't intend to change that because it's the number sign. Editors will copy parameter lists from template docs for use in articles. I have seen &lt;name> in articles. Those uses were caught because the module sees that parameter name as invalid. Not so for last#.

The module looks at incoming enumerated parameters and replaces the enumerator digits with the place-holder character:  →. Function  then looks in the appropriate table in ~/Whitelist for a match to   and returns the assigned value if found, nil else.

But, when the incoming parameter is &lt;name>, the parameter appears to be one where the digits have already been replaced so  finds   and returns the assigned value. Because  is not properly enumerated, its presence in a template is only detected by its absence in the final rendering. When  is the only name or is the last name of a list of names, it is not rendered so is only detected by the observant editor. When one of several names, but not the last name, then detected by the gap it causes in the listed names.

I have fixed this in the sandbox. It will be part of the imminent update.

—Trappist the monk (talk) 17:38, 15 April 2020 (UTC)

Discussion at Village Pump
Village_pump_(technical) -- Green  C  14:18, 18 March 2020 (UTC)
 * Village pump (technical)/Archive 179 --Matthiaspaul (talk) 19:22, 16 April 2020 (UTC)

cite web and mirror urls?
Any chance of cite web having the ability to specify a mirror-url (as distinct from and separate from a true archive-url)? For example, in, the single current source is no longer available from Intel directly (a PDF reference file) at the url listed. However, I found an alternative mirrored version, but it doesn't feel "correct" to take over the primary url field as it's possible it is still available directly from Intel at a different url. Presumably if url-status was "dead" then the displayed url would go mirror-url (if non-null), then archive-url. Thoughts? —Locke Cole • t • c 06:25, 7 April 2020 (UTC)
 * I would insert the mirror url as in mirror_url and also use Mirror_Website_Title. 98.0.246.242 (talk) 02:32, 8 April 2020 (UTC)
 * My only issue with doing that is then someone has to go spelunking through the article history to find the true original url (though there are tools to make this process simpler). Plus for 3rd party consumption I think it's valuable to have the original url for researchers/students that want to dig deeper into deadlinks. —Locke Cole • t • c 04:08, 8 April 2020 (UTC)
 * That goes way beyond what a citation system is supposed to provide. Especially the present one. It is not a research platform, it is a policy tool (WP:V). 98.0.246.242 (talk) 01:44, 9 April 2020 (UTC)
 * Hence this discussion, as I don't think it's necessarily a bad thing to take advantage of something that's fairly common on the web (mirroring). —Locke Cole • t • c 05:03, 10 April 2020 (UTC)

When links go to different web sites they are different citations, even though the site content might be the same. If you want to keep the original Intel link for whatever reason invoke the template two times. The archive-url field is only meant for web archive sites like archive.org and archive.today -- Green  C  04:23, 8 April 2020 (UTC)
 * If the archive-url is working, just leave in the original dead URL, put in dead, and the document will be available from the archive. No problem. – Jonesey95 (talk) 04:29, 8 April 2020 (UTC)
 * See JMP (x86 instruction). It is not an archive URL in the archive-url field. They are trying to make a primary URL into an archive URL. Both those URLs require a work. -- Green  C  04:39, 8 April 2020 (UTC)
 * That was a strange error. I have fixed it. The archive-url leads to the PDF in question, so all is well. – Jonesey95 (talk) 13:40, 8 April 2020 (UTC)
 * Thank you. It's actually pretty common for users to do this. They find another copy elsewhere and drop it into the archiveurl argument. There is a fairly large percentage of archiveurl's that do not correspond to any known web archiving service. Another problem is an archive URL that is a completely different from the primary URL. Both of these probably need tracking. -- Green  C  15:36, 9 April 2020 (UTC)
 * Which further bolsters the argument for having a mirror-url IMO. —Locke Cole • t • c 05:03, 10 April 2020 (UTC)
 * Green, sometimes there are valid reasons why the archive-url looks quite different from the original url and is still a valid archive-url, see: Village pump (technical)/Archive 179
 * --Matthiaspaul (talk) 20:40, 16 April 2020 (UTC)
 * RE: "error"; as I indicated, as mirrors are a thing for web content (as distinct from archive's like the WayBack Machine, etc), is there an objection to adding a mirror-url field as described in my original comment? —Locke Cole • t • c 05:03, 10 April 2020 (UTC)
 * I don't see a need to support a mirror-url. Pick one of the existing options: Switch the url or use archive-url as intended. You are not "usurping" someone's original URL if it no longer works. --Izno (talk) 14:29, 10 April 2020 (UTC)
 * I don't think I was clear: while my current use was to maintain the original url but provide a mirrored url (since, at the time, it was believed a proper archive wasn't available), there is still a use going forward to provide mirror-url's before the original url dies (in addition to, not instead of, the usual archive-url). —Locke Cole • t • c 19:27, 10 April 2020 (UTC)
 * It breaks the design of the citation template. Each URL has a corresponding work (or publisher), date, title, archive-url, IDs etc.. the template is designed to cite a single website or source. You are trying to combine two websites into a single citation and this breaks the model. What happens if the mirror URL is at a different website (different work field), published on a different date, has a different title, has a different access-date and then the mirror dies and the archive bots try to archive it where do they put the archive URL? Do you suggest we also create parallel mirror arguments for all these fields? If you want to cite two websites invoke two copies of the template. -- Green  C  20:34, 10 April 2020 (UTC)
 * It breaks the design of the citation template. [...] ... are you familiar with what a mirror site is? TL;dr: it's basically an archive, in fact it's actually better than some archives because there's usually no banner or indicator you're viewing something other than the original work. So all of those additional fields (work, publisher, title, etc.) are identical to the original content. Literally the only additional fields for this change would be  and possibly  . Functionally, if url-status were set to dead and mirror-url was non-null (and, if implemented, mirror-url was also not dead) then mirror-url would be exposed to readers; if both fields are dead, only then would the archive be invoked. In all cases, the various additional fields you noted remain the same because they all point to the same things.. —Locke Cole • t • c 03:24, 11 April 2020 (UTC)
 * Your proposing all these extra fields requiring significant code changes to deal with the interactions with other fields; creating unresolvable link rot problems (what happens when the mirror URL dies); mismatched work and possibly other metadata getting out of sync (mirrors are not always exactly the same metadata); new maintenance and tracking category issues; bots have to be rewritten to deal with the new fields; no objective understanding for the community of what a mirror site is (eg. a Reuters story carried at both WaPo and NYT some people will see as a mirror of the content). Or, simply invoke two copies of the cite template. Or, simply replace the primary URL with the new URL at the mirror. BTW please do not put mirror URLs in the archive-url field as this creates link rot that bots can not fix/resolve, the archive-url is for one of these: List of web archives on Wikipedia   --  Green  C  05:11, 11 April 2020 (UTC)
 * Your proposing all these extra fields [...]... two fields, actually. [...] requiring significant code changes to deal with the interactions with other fields; As someone who used to edit templates rather regularly, I feel very confident in saying it's not actually that significant. creating unresolvable link rot problems (what happens when the mirror URL dies) And this is in any way worse than the present situation? For web content that has a mirror site, this actually presents less issues, not more... mismatched work and possibly other metadata getting out of sync (mirrors are not always exactly the same metadata); This is not a problem, as I already explained. None of those other fields "[get] out of sync" because they are identical to their original source.. they are a mirror site.. bots have to be rewritten to deal with the new fields; Other than template syntax bots (if such things exist), I don't see a problem that is too burdensome to outweigh the advantages. no objective understanding for the community of what a mirror site is (eg. a Reuters story carried at both WaPo and NYT some people will see as a mirror of the content) They, like you, could start by reading mirror site. It's at mirror site. Just click here if you want to learn what a mirror site is. Mirror site. Or, simply invoke two copies of the cite template. And then watch as someone helpfully comes along later, not understanding the significance of the second identical source, and removes it. Brilliant! Or, simply replace the primary URL with the new URL at the mirror. As I said above: I don't think I was clear: while my current use was to maintain the original url but provide a mirrored url (since, at the time, it was believed a proper archive wasn't available), there is still a use going forward to provide mirror-url's before the original url dies (in addition to, not instead of, the usual archive-url). Not sure how else to explain that... BTW please do not put mirror URLs in the archive-url field as this creates link rot that bots can not fix/resolve, the archive-url is for one of these: List of web archives on Wikipedia Golly, thanks. —Locke Cole • t • c 05:54, 11 April 2020 (UTC)
 * You don't need to explain to me what a mirror is, though I doubt most people understand the distinction between a mirror and merely a copy of the content at another website. Even assuming they make that distinction, why would we have both a mirror and a web archive? It's an unnecessary complication that doesn't solve any problem that isn't already solved with web archive URLs (or multiple copies of the template). It creates problems as noted above, including editors will confuse a copy of the content found somewhere else as being a mirror. Telling them to read mirror site works as well as telling people to only use web archive links in the archive-url field ("Golly, thanks"), assuming they even read the docs. --  Green  C  17:28, 11 April 2020 (UTC)
 * Well, at least you know how a diff works. Now if only there'd been a  field I could have used instead, oh... right. —Locke Cole • t • c 18:09, 11 April 2020 (UTC)

Citing native scripts of the author's name and/or title of a work in Citation Style 1
In Zhongguo wenxue shi there is an article from Chen Guoqiu that I wish to include in the further reading section as an additional resource about a book. It is important, in regards to Chinese and Japanese works, to include the name in the native script of the author and of the title alongside romanized forms of both and/or the translation of the title - Giovanni Vitello does that in his citation in his journal article for a reason. However I am not sure in this citation template how to include those in separate parameters. If there is an "author-nativescript" or "title-nativescript" and/or "title-romanization" I would like to have those fields. These would also help with "editor", "chapter", "work", and "newspaper" parameters too.

Thanks, WhisperToMe (talk) 22:07, 15 April 2020 (UTC)
 * Not for any of the name parameters. But, script-title, script-chapter, script-work, script-newspaper, and others; these are documented at the various templates.
 * I don't that there has been much call for  variants of the name-holding parameters.
 * ——Trappist the monk (talk) 22:21, 15 April 2020 (UTC)
 * Thank you! I got script-title and script-journal to work! It would be nice if script-author was implemented too. It's very specific to Asian Studies (China and Japan) I believe, which may explain why there hadn't been much demand overall. I can bring this up at the respective WikiProjects. WhisperToMe (talk) 22:39, 15 April 2020 (UTC)
 * I notified the China, Hong Kong, Japan, Macau, Malaysia, Singapore, and Taiwan projects. Korea might find this useful as people are known under Hanja names, though it may have to be very old books that use the hanja? WhisperToMe (talk) 22:48, 15 April 2020 (UTC)
 * Came here from being notified at Singapore project. The  parameters for names would be used if we are using Chinese sources, which I see more common for BLPs for our local artists as we have invariably more artists in the Chinese medium. I would also see this happening if more of us Singapore-based editors start to dive into the history of Singapore, pulling out articles from non English-mediums like Chinese newspapers from the archives at https://eresources.nlb.gov.sg/newspapers/. robertsky (talk) 03:10, 16 April 2020 (UTC)
 * Having  variants of the name-holding parameters would be very useful in Japanese sourcing as there are often (almost always) multiple ways to write a name. For a given name example, see Hiroyuki. While not as common for surnames, there are still often 2-3 ways to write many of them. ··· 日本穣  ·  投稿  · Talk to Nihonjoe ·  Join WP Japan ! 08:56, 16 April 2020 (UTC)
 * Here are a few older requests for script-author/editor* and script-publisher variants:
 * Help_talk:Citation_Style_1/Archive_12
 * Help_talk:Citation_Style_1/Archive_2
 * And here is one for script-series:
 * Help_talk:Citation_Style_1/Archive_30
 * As I have run into situations where I could have made good use of all of them several times already (although not all of them at the same time), I support the addition of them.
 * --Matthiaspaul (talk) 07:23, 18 April 2020 (UTC)
 * I would like to add my support for a way to list author names in the original script alongside their Romanization. The reason for this (as I learned in writing a recent article on a book whose reviews complained about the author not doing this) is that for Japanese and Chinese names, transliteration is a one-way transformation: you can use the Romanization to find out what the original name sounds like, but not how it is spelled, because there are too many different kanji that could be used in names but are homophones. So in scholarship that involves names from east Asia, having the original name is essential. And it is not possible to use only the original name because MOS:ROMANIZATION forbids it. So we need side-by-side Romanization and original script. —David Eppstein (talk) 07:42, 18 April 2020 (UTC)
 * I would like to add my support for a way to list author names in the original script alongside their Romanization. The reason for this (as I learned in writing a recent article on a book whose reviews complained about the author not doing this) is that for Japanese and Chinese names, transliteration is a one-way transformation: you can use the Romanization to find out what the original name sounds like, but not how it is spelled, because there are too many different kanji that could be used in names but are homophones. So in scholarship that involves names from east Asia, having the original name is essential. And it is not possible to use only the original name because MOS:ROMANIZATION forbids it. So we need side-by-side Romanization and original script. —David Eppstein (talk) 07:42, 18 April 2020 (UTC)

citing different pages of the same book
Hi guys, I'm using the cite book template, and sometimes, within an article, I need to cite different pages of the same book, in different parts of the article. One way to do this is to have each page as a different reference. But that would entail creating several different references in the same article, all with the same information about the same book repeated many times, and only the page number parameter would be different in each instance. I imagine there must be a smarter way to do this. Could anyone please let me know? Thanks! Dr. Vogel (talk) 20:38, 11 April 2020 (UTC)
 * I take a page out of the CMOS play book myself. They use full footnotes and shortened footnotes for repeats. In my writing, I'll frequently cite a reference book multiple times. For the first footnote, I run the citation in full and add harv to the template. For the subsequent references, I use harvp. So it looks something like:
 *  Imzadi 1979  →   21:06, 11 April 2020 (UTC)
 * Yes, I was going to suggest exactly the same style. I think it's a good fit for articles whose referencing styles put full citations in Citation Style 1 for other sources without this issue. My only addition would be to use sfnp for subsequent citations instead of &lt;ref&gt;harvp&lt;/ref&gt;. —David Eppstein (talk) 21:17, 11 April 2020 (UTC)
 * For examples of using multiple pages from multiple sources, poke through long biographical articles like Herman Melville. Looking at the wikitext of articles like that can give you a sense of how to construct short citations that link to full citations. – Jonesey95 (talk) 21:18, 11 April 2020 (UTC)
 * ✅ Thank you all for helping, and for replying so quickly. Dr. Vogel (talk) 21:37, 11 April 2020 (UTC)
 * I personally dislike sfnp and its brethren for a simple reason. Because the coding for the footnotes it generates is not in , those references aren't included with the others when isolating the references via the various scripts for that purpose. Others' mileage may vary on that issue though.  Imzadi 1979  →   04:19, 12 April 2020 (UTC)
 * Just because I'm curious, what scripts are those? Why is it that you think  is not in <ref ></ref> tags?  Here is the output of your   rewritten as  :
 * —Trappist the monk (talk) 17:39, 12 April 2020 (UTC)
 * I use two scripts. The first one I found is at User:Dr_pda/editrefs.js, and it "adds a link to the toolbox which, when clicked, searches the article for <ref ></ref> tags and presents them in textboxes for ease of editing". The second at User:PleaseStand/segregate-refs.js is the one I use more though because it creates a second edit window below the main one with all of the footnotes listed sequentially and applies temporary names to all of the unnamed references in the body text, allowing me to work with the main text and the footnotes separately but at the same time. (The first replaces the main editing box with the individual text boxes for each footnote.) Both of these scripts ignore sfnp and its brethren because sfnp is not enclosed in <ref ></ref> within the wikicode. (They also ignore templates like inflation-fn, but because that is such a single-purpose it doesn't hamper editing the same way.)  Imzadi 1979  →   15:35, 13 April 2020 (UTC)
 * Thank you.
 * —Trappist the monk (talk) 16:14, 13 April 2020 (UTC)
 * The easiest (and probably the most common) way to cite multiple pages of a work is to use pages instead of page and provide a comma-separated list of pages or page ranges there:
 * Thank you.
 * —Trappist the monk (talk) 16:14, 13 April 2020 (UTC)
 * The easiest (and probably the most common) way to cite multiple pages of a work is to use pages instead of page and provide a comma-separated list of pages or page ranges there:


 * --Matthiaspaul (talk) 16:00, 14 April 2020 (UTC)
 * Although there is nothing wrong with this, for the sake of clarity, context, and text-source integrity I would avoid it. Imo a short ref with the specific location, linking to the bibliographic full reference is best, but this is just a preference. 98.0.246.242 (talk) 17:48, 14 April 2020 (UTC)
 * For references used to support many different statements in an article, providing specific locations might make sense, but in most cases, only a handful of locations will be specified in an article, so that the introduction of an extra level of indirection appears to be overkill (unnecessary clutter) as it is easy enough for the reader to find out the corresponding pages. Also, there are no backlinks from the full citation to the short one, and there is no means to jump from one location in an article to another citing the same reference (although perhaps a different page). However, to address a desire for specific pages with backlinks, perhaps something like the following could be incorporated into CS1 style citations in the future:


 * --Matthiaspaul (talk) 18:59, 14 April 2020 (UTC)
 * This comes close: https://meta.wikimedia.org/wiki/WMDE_Technical_Wishes/Book_referencing
 * --Matthiaspaul (talk) 08:00, 18 April 2020 (UTC)

Subtitles
Has there been discussion on how to handle a Subtitle of a book? I just want to make sure I'm doing it right. I can't find anything in the MOS or Template:Cite book that provides direction. I'd love it if there was an actual parameter for it in the template, but I'd appreciate even some guidelines on how to display the title vs subtitle. Thanks. Canute (talk) 16:55, 5 April 2020 (UTC)


 * WP:SUBTITLES has: "The standard separator for the title and the subtitle (that is, in cases where both taken together don't constitute a continuing phrase) in the page name is a colon followed by a space, as in [Orlando: A Biography]" – in the cite template(s) there is afaik no separate parameter for a book's subtitle: when a book has one, I've used the WP:NCB format in the "title" parameter if I wanted to include the subtitle in the cite template.
 * I don't think a separate parameter is needed/desirable for cite templates (more complexity; more confusion; no tangible advantage). --Francis Schonken (talk) 18:27, 5 April 2020 (UTC)


 * I think, if the source provides a specific separation style between main- and sub-title (for example in the front matter), we should follow it. In cases, were this isn't clear from the source, we should preferably use either ": " or " – ", but I have also seen editors using " - ", ", ", "; " ". ", " | " and " > ".
 * This topic has been discussed before. I do think, a sub-title parameter would be useful (as is done in citation templates in some other Wikipedias).
 * It would allow us to provide a default separator and thereby free editors from having to invent their own style. This could be updated centrally if a different style would be preferred in the future. This would lead to more consistency and easier maintenance.
 * Following the separator, the sub-title could just be appended to the main title for meta-data purposes. This would be backwards-compatible with the current use.
 * The display of sub-titles could be made user-configurable utilizing CSS, so that users, who don't want to read the full titles, could suppress the sub-titles. Also, on small or narrow displays (like mobile phones), we might want to display subtitles differently (perhaps smaller) or not at all by default. Having a split into main and subtitle would allow for such kind of target-specific renderings, whereas this is impossible if the sub-title gets lumped into the title parameter.
 * Of course, we would need some way to override the default separator through some parameter like title-separator. Since valid separators can be more than one character (including leading and/or trailing spaces), the way to specify different separators would need some more thought (but is doable) to remain intuitive and easy to use.
 * --Matthiaspaul (talk) 20:29, 16 April 2020 (UTC)
 * I think an easy way to handle this would be as follows. By default, the template will assume " – " as separator. If, however, title-separator is used, its argument will be used as separator with a space appended, so : would result in ": ",, in ", ", and string in "string ".
 * Alternatively, if the default separator would be ": ", title-separator could be used to override the separator as before, but with one special case: – or - would result in " – " rather than "– " (so that users would not have to resort to &amp;nbsp; for this common case).
 * --Matthiaspaul (talk) 12:01, 18 April 2020 (UTC)
 * Oppose, totally unnecessary complication (what would be the benefit of it?). --Francis Schonken (talk) 12:10, 18 April 2020 (UTC)
 * Already given above. --Matthiaspaul (talk) 12:50, 18 April 2020 (UTC)
 * Nah, I don't see it. I see clear disadvantages (as explained: additional complexity, to name only one). "Some other Wikipedias do it" is also not necessarily either an advantage or disadvantage (too much of a WP:OTHERLANGS reasoning). "We could do this or we could do that" also doesn't explain "why" doing this or that would be an advantage, at the cost of adding several layers of additional complexity to a set of templates that are already so complex that template documentation can't keep up, leave alone be clear on how to use the templates uniformly. --Francis Schonken (talk) 13:27, 18 April 2020 (UTC)

Lua error in Module:Citation/CS1 at line 845: attempt to index field 'defaults' (a nil value).
This error is popping up today, as seen on the Warren Street tube station article:


 * Lua error in Module:Citation/CS1 at line 845: attempt to index field 'defaults' (a nil value).
 * Backtrace:
 * Module:Citation/CS1:845: ?
 * Module:Citation/CS1:2223: in function "citation0"
 * Module:Citation/CS1:3858: in function "chunk"
 * mw.lua:518: ?
 * [C]: ?
 * mw.lua:518: ?
 * [C]: ?

Thanks! Slambo <small style="color: black;">(Speak) 13:01, 18 April 2020 (UTC)


 * Maybe I was a little hasty in reporting... Refreshing the page, the error is not shown. Slambo <small style="color: black;">(Speak)  13:04, 18 April 2020 (UTC)
 * Errors like that occur during a module suite update because the update cannot apply to all of the modules simultaneously and instantaneously. During the brief interval that begins with the saving of the first updated module until the completion of the saving of the last updated module, the whole suite is out of sync and therefor broken.
 * —Trappist the monk (talk) 13:34, 18 April 2020 (UTC)

biorxiv error


Resolves to the wrong link, https://doi.org/ 10.1101/ 10.1101/F355933 instead of the correct https://doi.org/10.1101/355933. &#32; Headbomb {t · c · p · b} 17:05, 18 April 2020 (UTC)
 * This also wrongly emits a Category:CS1 errors: bioRxiv &#32; Headbomb {t · c · p · b} 17:10, 18 April 2020 (UTC)
 * It looks fine to me as of this time stamp. Maybe the page in question just needed a purge after the module update, which happened a few hours ago. If it is still happening on a page after a null edit, link to that page here. Edited to add: Trappist fixed it.– Jonesey95 (talk) 18:19, 18 April 2020 (UTC)
 * Two problems; prefix was one, the other was that the error cat was prematurely set; not now.
 * —Trappist the monk (talk) 18:23, 18 April 2020 (UTC)

Cite thesis issues
The first "cite thesis" entry on African humid period is producing "thesis thesis" in the displayed text. Jo-Jo Eumerus (talk) 20:36, 19 April 2020 (UTC)
 * You have told the template that the source is a 'thesis degree thesis' instead of a 'masters or phd or whatever degree thesis'. That source should tell you the degree; it times out for me so perhaps you'll have to dredge up an archive or a different source.
 * The template is working correctly.
 * —Trappist the monk (talk) 20:44, 19 April 2020 (UTC)


 * I have corrected the four references in the external links to add the degree.  Hawkeye7   (discuss)  20:51, 19 April 2020 (UTC)
 * Thanks for this. Jo-Jo Eumerus (talk) 12:26, 20 April 2020 (UTC)

Updating refToolbar
I've opened a discussion on how to make the RefToolbars work with the new Harv mode of the templates, here, in case any of the template maintainers has opinions. Jo-Jo Eumerus (talk) 12:35, 20 April 2020 (UTC)

Thanks
I know, not the normal use of a help talk page, but since this one is essentially the CS1/CS2 maintenance noticeboard and presumably watchlisted by all the people who maintain them: Thanks for changing the templates so that I no longer have to type out harv when using the RefToolbar & sfn. Jo-Jo Eumerus (talk) 20:47, 21 April 2020 (UTC)
 * I think you've come to the wrong page! This is the page for complaining when we make things more consistent, not thanking us! How dare you!?!? (purely joking, in case it is unclear).
 * But seriously, thanks for taking the time to visit and thank us. Let us know if you run into any trouble with the new configuration. It is possible that some tools may need updates. – Jonesey95 (talk) 21:29, 21 April 2020 (UTC)
 * But seriously, thanks for taking the time to visit and thank us. Let us know if you run into any trouble with the new configuration. It is possible that some tools may need updates. – Jonesey95 (talk) 21:29, 21 April 2020 (UTC)