Help talk:Citation Style 1

From Wikipedia, the free encyclopedia
  (Redirected from Template talk:Cite book)
Jump to: navigation, search
Citation templates (conception)
Citation templates (reality)

Using names of months[edit]

i have two connected problems on Cite news on cywiki: 1. When I copy this most recent version to our older module on cywiki, it causes an error which I'm unable to correct and 2. The date format on cywiki is always dd name of month yyyy (eg 19 Gorffennaf 2017), but as you can see on cy:Alfred Russel Wallace we can only use yyy-mm-dd (eg 2017-07-19). Any help would be warmly received! Llywelyn2000 (talk) 04:40, 22 July 2017 (UTC)

@Llywelyn2000: What is "an error which I'm unable to correct"?
Does the date format problem only happen with cy:Template:Cite news, or does it affect cy:Template:Cite web as well? Is it only the |access-date= parameter, or |date= as well? What happens if you use English-language dates like 19 July 2017 - does the problem disappear? --Redrose64 🌹 (talk) 08:30, 22 July 2017 (UTC)

I'm also perplexed. You said this most recent version by which I would expect the 28 May 2017 version of the module suite. But, at, the date there looks to be 26 May 2016.

The |access-date= validation at does not work because the MediaWiki time parser does not understand non-English month names. Try this in your sandbox at

{{#time:U|5 Ebrill 2007}}

then try this:

{{#time:U|5 April 2007}}

you should get:

{{#time:U|5 April 2007}} → 1175731200

Because this problem exists in all non-English wikis, all versions of the module suite since 30 July 2016 have a fix for the problem.

What is your intent? Do you want to just fix your version of your older fork of the module suite? Do you want to the module suite to track the modules?

I can see that someone has copied en:Module:Citation/CS1/Date validation/sandbox to cy:Modiwl:Citation/CS1/Date validation/sandbox, changed the month names to Welsh and then reverted. I would not recommend ever taking the sandbox versions of the module suite; they are never guaranteed to be working. You should, instead, take the current live version into your sandboxen and make sure that whatever changes are required to suit your particular language are made first there before upgrading the live versions.

Trappist the monk (talk) 11:59, 22 July 2017 (UTC)

Many thanks to both of you. I've reproduced the problem here. I managed to translate the months, so we're partly there! @Redrose64: - I've also added your bits; the Welsh one doesn't work. Intent - to enable any form of input to produce our usual format (27 Mai 2017). Thanks again! Llywelyn2000 (talk) 05:53, 24 July 2017 (UTC)
The dates rendered in the citation at test 1 are (for me): 2017-07-19 and Adalwyd 2017-07-19. I see no 19/07/2017 in that page except where you write that the dd/mm/yyyy style is not used on Test 1 apparently shows that the module suite is working correctly; can't fix it if it ain't broke.
The reason for the failure of test 2 is described above and confirmed by the results of test 3.
Intent - to enable any form of input to produce our usual format (27 Mai 2017). Do you mean that you want the module suite to automatically reformat dates in mdy and ymd formats to dmy format?
Trappist the monk (talk) 10:17, 24 July 2017 (UTC)
I'm not making myself clear! Sorry, i'll try again:
Test 1 - should produce '19 Gorffennaf 2017'; as you say, the module should reformat to our accepted format
Test 2 - should not have that error at the end (Check date values in: |access-date= (help))
Test 3 - should allow the input to be in Welsh (5 Ebrill 2007); at the monment it creates an error 'Gwall: Amser annilys'.
Thanks! Llywelyn2000 (talk) 16:10, 28 July 2017 (UTC)
At test 1 you write this: If we input the date as 2017-07-19... what we get is 19/07/2017. I don't see that. Here is a copy paste of the rendered citation I see:
"Gwerthu llythyrau’r Cymro oedd yng nghysgod Charles Darwin". Golwg 360. 2017-07-19. Adalwyd 2017-07-19.
19/07/2017 is not part of that rendering. I don't know how you can be seeing a dd/mm/yyyy date format; that is not something that the modules produce.
Test 3 does not test the cs1|2 modules; it tests the MediaWiki {{#time}} parser function. It is not possible for us to fix that here. See phab:T21412.
Test 2 fails because of the problem illustrated in test 3. I have explained that newer versions of the cs1|2 module suite have overcome that problem. My recommendation to you is to do these things:
  1. import the current version of the live modules (all of them) from into their appropriate sandboxen at
  2. modify the sandboxen as appropriate to suit the Welsh language (most will not require modifications I think)
  3. test to make sure that the sandbox suite works
  4. overwrite the current live version with the sandbox version
Automatic date reformatting is not something that the cs1|2 modules have ever done. At |df= allows us to do date reformatting on a citation-by-citation basis. And that works at, but not correctly (it renders English month names). I believe that I know a solution to that which would enable automatic date reformatting and, as a side benefit, enable automatic translation of English month names into the local language month names. I'm interested in exploring that but am not interested in retrofitting it into older versions of the module suite. would be a good test-bed but it must be upgraded first.
Trappist the monk (talk) 10:41, 29 July 2017 (UTC)
The rendered reference for test 1 is wholly italicized. It should not be. This is because user agents (browsers) apply default italic styling to the <cite>...</cite> tag that wraps the citation. Wherever it is that handles site-wide styling, you can add this:
/* Reset italic styling set by user agent */
cite, dfn {
    font-style: inherit;
(may require some consensus). Alternately, we can edit cy:Module:Citation/CS1/Configuration to locally override the default. A site-wide solution would be preferred I suspect. I don't know where keeps its site-wide css; if you cannot find it, I suspect that there are editors at WP:VPT who can help.
Trappist the monk (talk) 11:19, 29 July 2017 (UTC)

@Trappist the monk: - Many thanks for your work on this! I have a lot of work to do! I'll get back to you in a millennium or two! Llywelyn2000 (talk) 08:38, 14 August 2017 (UTC)

Identifier order messed up.[edit]

Why is bibcode displaying before arxiv in?

Identifiers should be listed in alphabetical order. Headbomb {t · c · p · b} 13:49, 10 June 2017 (UTC)

The identifier labels are sorted with a case sensitive sort. 'B' has an ascii numerical value of 66 (0x42) and 'a' has an ascii numerical value of 97 (0x61). Proof for that is here, where I've added |eissn=1365-2966 and |issn=0035-8711 from the journal's wikipedia article:
Corbelli, E. & Salucci, P. (2000). "The extended rotation curve and the dark matter halo of M33". Monthly Notices of the Royal Astronomical Society. 311 (2): 441–447. Bibcode:2000MNRAS.311..441C. ISSN 0035-8711. arXiv:astro-ph/9909252Freely accessible. doi:10.1046/j.1365-8711.2000.03075.x. eISSN 1365-2966. 
Trappist the monk (talk) 14:42, 10 June 2017 (UTC)
Well, that ought to be fixed then, either with case-insensitive sorting, or by putting the sortkey in a {{lc:IDENTIFIERNAME}} type of thing. Because it wasn't like that before. Headbomb {t · c · p · b} 15:18, 10 June 2017 (UTC)
There have been no changes to the identifier sorting since at least this version (April 2013) of Module:Citation/CS1.
Trappist the monk (talk) 15:59, 10 June 2017 (UTC)
I distinctly remember those to be sorted correctly as late as this spring. But even if my memory somehow fails me, those should be sorted alphabetically, regardless of casing. Headbomb {t · c · p · b} 17:02, 10 June 2017 (UTC)
De-archived as unresolved and still in need of a fix. Headbomb {t · c · p · b} 13:32, 22 July 2017 (UTC)
De-archived again. @Trappist the monk and Jonesey95:. Headbomb {t · c · p · b} 13:58, 13 August 2017 (UTC)

Any update on doi-broken-date?[edit]

If anything, the doi should at the very least still link. Other improvements can wait/get more discussion, but the linking part should be easy to fix. Headbomb {t · c · p · b} 14:25, 11 April 2017 (UTC)

@Trappist the monk: Any way we can get this bundled in the weekend's update? Headbomb {t · c · p · b} 05:29, 26 April 2017 (UTC)
The purpose of this interstitial period is to have a last chance to find and fix bugs; to create or modify supporting documentation, categories, templates, etc. – housekeeping preparatory to the update. It is not the time for new development or new features.
Trappist the monk (talk) 11:28, 26 April 2017 (UTC)
Yeah well this has been requested a long while ago, is an easy fix, and we have over half a week left. WP:BURO applies here. Headbomb {t · c · p · b} 11:57, 26 April 2017 (UTC)

Can we now implement this? Headbomb {t · c · p · b} 00:27, 2 May 2017 (UTC)

De-archived because discussion is ongoing/unresolved. @Trappist the monk:. Headbomb {t · c · p · b} 17:20, 23 May 2017 (UTC)
@Trappist the monk and Jonesey95: pinging. Headbomb {t · c · p · b} 19:13, 12 June 2017 (UTC)
It makes sense to me to have allegedly broken DOIs linked, since the doi-broken-date is checked by a bot and (a) could have been wrongly applied or (b) could have been a temporary problem or (c) both. There are plenty of links that don't work and are not flagged as such. That's just the state of the web, and always has been. – Jonesey95 (talk) 02:00, 13 June 2017 (UTC)
De-archived as unresolved and still in need of a fix. Headbomb {t · c · p · b} 13:32, 22 July 2017 (UTC)
De-archived again. @Trappist the monk and Jonesey95:. Headbomb {t · c · p · b} 13:58, 13 August 2017 (UTC)

MR error checking[edit]

MR come in two formats, old (which I can't find documentation for, but the structure seems to be \d+([a-z]|#)?:\d+ [case-sensitive]) and new (\d+). The canonical identifier is a 0-padded 7-digit string (see [1], scroll to "New format for primary item identifier" section), but the zeros are optional, and are stripped for non-subscribers [2], and will therefore be common. The template should simply automatically zero-pad to 7 digits.

The old style should be put in a maintenance category so they can be updated. Anything else should be flagged as errors (e.g. |mr=mr01234, found in [3]). Headbomb {t · c · p · b} 18:12, 28 August 2017 (UTC)

JFM error checking[edit]

JFM seems to come in only one format (\d\d\.\d\d\d\d\.\d\d). Headbomb {t · c · p · b} 23:33, 2 August 2017 (UTC)


But that's are not really JFM identifiers, those are ERAM identifiers, which happen to resolve to the same database. Not quite sure how to handle that one. Headbomb {t · c · p · b} 14:30, 4 August 2017 (UTC)

I've updated them to use |id={{ERAM|foobar}} instead. Headbomb {t · c · p · b} 20:57, 4 August 2017 (UTC)
JFM could be abused to put a Zbl identifier. If the Zbl structure is found, it should throw an error and tell users to user |zbl= instead. Headbomb {t · c · p · b} 14:54, 4 August 2017 (UTC)
Unarchived as unresolved. Headbomb {t · c · p · b} 14:38, 26 August 2017 (UTC)

Zbl error checking[edit]

ZBL come in two formats, old [temporary?] format which consists of pure digits (\d+, possibly 8 digits \d{8}) and new (\d{4}\.\d{5}). Catching errors would mean having a way of catching mistakes such at |zbl=t0303.10056, e.g. [4].Headbomb {t · c · p · b} 14:15, 4 August 2017 (UTC)

\d{3}\.\d{5} will resolve (e.g Zbl 607.73009), but the correct identifier has a \d{4}\.\d{5} structure (e.g Zbl 0607.73009). Headbomb {t · c · p · b} 15:14, 4 August 2017 (UTC)

I'll also point out that I found quite a few (pre\d+) that would resolve only once the (pre) part was striped [e.g. [5], which I've updated to the new style]. I think those are temporary Zbl identifiers. Compare Zbl pre06066616 (which doesn't resolve) to Zbl 06066616 (which resolves to Zbl 1260.11001). Headbomb {t · c · p · b} 14:21, 4 August 2017 (UTC)

Zbl will often be abused to put a JfM identifier [6]. If the JFM structure is found, it should throw an error and tell users to user |jfm= instead. Headbomb {t · c · p · b} 14:53, 4 August 2017 (UTC)
This page describes the coding for the \d{4}\.\d{5} format, but I could not find anything for the other formats or temporary IDs. If the first four digits are a volume number, it makes sense that one could remove the zero padding with no loss of info. --Mark viking (talk) 20:44, 4 August 2017 (UTC)
If the \d{8} format isn't desired, that could be shown as an error too. There's only a handful of (around 41, last I checked). Headbomb {t · c · p · b} 21:02, 4 August 2017 (UTC)
I've updated them all, except [7] Zbl 06247765 in Superpermutation. This is either the temporary assignment, or an oversight in the Zbl database, or it's just a legit but undocumented code. Headbomb {t · c · p · b} 21:37, 4 August 2017 (UTC)
I've confirmed those are temporary assignments. They should be put in a maintenance category so they can be updated to the canonical Zbl identifier once they get assigned. Headbomb {t · c · p · b} 14:43, 26 August 2017 (UTC)

Handling sites that have become malicious[edit]

The site in July was promoting fake tech support. That problem seems to have gone away, but now the https version of the site has an invalid security certificate. Whether heritage-history has been taken over by bad guys forever, or they are somehow able to resolve their problem, this point up a need to be able to cite an archive of a site without having the citation contain a link to the original site, so as to minimize the risk that readers will click the link and arrive at a malicious site.

See Military history of Italy for an example of an offending link. Jc3s5h (talk) 13:36, 4 August 2017 (UTC)

|dead-url=usurped or |dead-url=unfit
Trappist the monk (talk) 13:58, 4 August 2017 (UTC)
The parameters values |dead-url=usurped or |dead-url=unfit are not described on the help page. I don't use undocumented parameter values. Jc3s5h (talk) 14:07, 4 August 2017 (UTC)
I see partial documentation at Template:Cite web/doc#URL, but it does not make clear if the value of the url parameter should be the link to the malicious site, or should just be url= Jc3s5h (talk) 14:13, 4 August 2017 (UTC)
If a parameter is documented at any of the cs1|2 template pages, it is documented. You are free to improve the cs1|2 documentation so please do. The documentation does say "setting |dead-url=usurped or |dead-url=unfit suppresses display of the original URL (but |url= is still required).' You will discover that omitting |url= or leaving |url= blank will get the |archive-url= requires |url=. Again, you are free to make the documentation better. Please do so.
Trappist the monk (talk) 15:05, 4 August 2017 (UTC)
I don't want to test this on a live article. What if the malicious site has already been added to the spam blacklist before the attempt to add the deadurl=unfit parameter. Wouldn't it be impossible to perform the edit with the offending URL still present? Jc3s5h (talk) 15:47, 5 August 2017 (UTC)
If blacklist will prevent page-save, there is no problem, right? And even if blacklist does not prevent the page-save, undo is your friend. For those cases where blacklist prevents page-save, |url= can be set to an innocuous site, perhaps with |dead-url=unfit, and |archive-url= & |archive-date= as appropriate for the archive copy and then <!--original url blacklisted--> added at the end of the cite.
Trappist the monk (talk) 10:34, 6 August 2017 (UTC)

Proposed revision to "Usage" text[edit]

The current "Usage" subdocument currently reads, "Copy a blank version to use. All parameter names must be in lowercase. Use the "|" (pipe) character between each parameter. Delete unused parameters to avoid clutter in the edit window. Some samples may include the current date. If the date is not current, then purge the page."

I would like to change this to, "You may copy a blank version to use. All parameters are optional except the |title= parameter. Parameter names must be in lowercase or will not be recognized. A "|" (pipe) character must be placed between each parameter. You should delete unused parameters to avoid clutter in the edit window. Some samples may include the current date. If the date is not current, then purge the page."

Thoughts? Objections? KDS4444 (talk) 03:25, 13 August 2017 (UTC)

"All parameters are optional except the |title= parameter" is not true. For example, |url= is required if |access-date= is used, and |last2= is required if |last1= |last1= is required if |last2= is used. Look at the "Prerequisites" section of the {{cite journal}} documentation for more requirements, and also read the explanation of title parameter requirements at Category:Pages with citations lacking titles. – Jonesey95 (talk) 15:55, 13 August 2017 (UTC)
That would be |last1= is required if |last2= is used, right?
Perhaps that text in Category:Pages with citations lacking titles (and so Help:CS1 errors) needs revisiting. Certainly |title= is a requirement except in {{cite conference}} where |book-title= can be used. Otherwise, |chapter=, |article=, |contribution=, |entry=, and |section= cannot stand alone as titles. It was once true that they could but that has long since been changed.
Trappist the monk (talk) 17:08, 13 August 2017 (UTC)

Using OAdoi[edit]

According to , nearly 50 % of the papers people look for (usually more recent publications) are available in some form of open access, even without counting academic social networks. Perhaps we should point all DOI links to instead of ? It's clearly a better service if people get to an URL where they can actually access the resource, and (used by hundreds of libraries already) redirects to if it doesn't know a better destination URL than the publisher-provided one. Alternatively, we could keep a double link, but I think we already provide too many overlapping links. --Nemo 05:31, 13 August 2017 (UTC)

Previous discussion.
Trappist the monk (talk) 09:10, 13 August 2017 (UTC)
Thanks. The situation evolved in the latest few months though, to the point oaDOI is becoming the de facto main DOI resolver for many institutions (being also included in Web of Knowledge: [8]). If we wanted to be standard-compliant we would (only) display the "doi:" pseudo-protocol, instead we prefer to provide links which are useful (as CrossRef rightly recommends for its own
Automatically linking the best PDF available is better than requiring users to edit articles and hardcode URLs or identifiers. --Nemo 22:41, 14 August 2017 (UTC)
Hi Nemo_bis! I would be surprised to see this happen, for the same reasons as the one pointed out in the previous discussion. Resolving errors are still possible (and in fact, likely, as the service relies on screen scraping). In terms of policy, even running OAbot as a bot was controversial, because a mass addition of CiteseerX links would violate WP:ELNEVER (see the discussion here). Switching the DOI resolver to would do that at a much larger scale and is therefore likely to face the same backlash. As I wrote earlier, I think this policy is out of touch with the current practices, but it seems quite hard to change it.
But there are even practical reasons why switching the DOI resolver would be a problem. Consider the following citation:
Selinger, P. (2010). "A Survey of Graphical Languages for Monoidal Categories". In Bob Coecke (ed.). New Structures for Physics. Lecture Notes in Physics. Springer Berlin Heidelberg. pp. 289–355. ISBN 978-3-642-12820-2. arXiv:0908.3347Freely accessible. doi:10.1007/978-3-642-12821-9_4. 
If we decide that the DOI is resolved by like this, then the two links will likely point to the arXiv: that is redundant. What if the reader actually wants to access the publisher's website instead? (Sometimes the documents are slightly different, for instance.) So we would need to have two links, one for and one for, but I agree that would be quite heavy (and useless in many cases, if not most).
Pintoch (talk) 07:24, 15 August 2017 (UTC)
Note that we could still have something like |auto-url=oadoi that would provide that functionality. But first we need to get that autolinking feature implemented, so when we have |doi-access=free, we get the doi link on the title. Headbomb {t · c · p · b} 15:38, 15 August 2017 (UTC)

Ongoing <references /> discussion at WP:VPPR[edit]

The village pump discussion about modifying <references /> into columns is still ongoing. I invite you to comment there. --George Ho (talk) 09:58, 14 August 2017 (UTC)

Edit request to setup or track CS1/Arguments module documentation[edit]


Please create Module:Citation/CS1/Arguments/doc with {{Improve documentation|date=August 2017}} (or better) for maintenance category tracking. I've already created Module talk:Citation/CS1/Arguments/doc and redirected here. The documentation page to be created will appear at Module:Citation/CS1/Arguments and will categorize into Category:Templates with missing or incorrect documentation.

It is not semi-protected per se, but anonymous contributors can't create new pages to module namespace. 2001:2003:54FA:D2:0:0:0:1 (talk) 20:14, 16 August 2017 (UTC)

Citing podcast episodes by number[edit]

I am trying to cite a particular episode of a podcast using {{cite podcast}} but I am having trouble getting the episode number to come out:

  • {{Cite podcast|url= |series=The Bad Boy Running Podcast|volume=50 |title =SPECIAL - John Kelly on finishing Barkley Marathons | |access-date=2017-04-15}}

comes out as

The obvious solution from the documentation is to use |volume=50 but that isn't showing up in the output: where am I going wrong? TIA HAND —Phil | Talk 13:47, 18 August 2017 (UTC)

At Help:Citation Style 1#Pages there is a table that shows that {{cite podcast}} does not support |volume=, |issue=, or |page(s)=. The discussion that created that table is here. There was no discussion of {{cite podcast}}.
Were it me, I would write the title of your podcast as it actually appears on the podcast's website:
{{cite podcast |url= |title=Ep50 - SPECIAL - John Kelly on finishing Barkley Marathons |website=The Bad Boy Running Podcast |publisher= |host= |date= |time= |access-date=2017-04-15}}
"Ep50 - SPECIAL - John Kelly on finishing Barkley Marathons". The Bad Boy Running Podcast (Podcast). Retrieved 2017-04-15. 
Trappist the monk (talk) 14:39, 18 August 2017 (UTC)
I agree with Trappist. As far as I can tell, podcasts are numbered by episode only in their titles. When I look at a podcast's metadata in my feed, there is no "episode" attribute or field. – Jonesey95 (talk) 14:48, 18 August 2017 (UTC)

Suggestion - Add one sentence to the Cite web template documentation[edit]

I recently asked this question in the (very helpful) Teahouse: In a website citation, if the web page title is in ALL CAPS, should we keep it that way? The answer: No, change it to title case.

I had searched for the answer prior to asking it in the Teahouse, including reading the template documentation for {{cite web}}. The {{cite web}} template documentation, under 3. Parameters > 3.4 Description > 3.4.2 Title, explains:

title: Title of source page on website. Displays in quotes.

My suggestion is to add this sentence:

I defer to your collective wisdom if this is a worthwhile addition. It would have helped me, and I did find the answer via the Teahouse.

Thanks!   - Mark D Worthen PsyD (talk) 06:39, 20 August 2017 (UTC)

That's a good suggestion, but doesn't 3. Parameters > 3.4 Description > 3.4.2 Title already have a hatnote just above the line you quoted, directing readers to the passage you're suggesting? – Margin1522 (talk) 02:40, 23 August 2017 (UTC)

Date formats for ranges[edit]

Why does "{{cite book|title=Title|publisher=Publisher before date|date=July 1 – August 1, 2008|df=mdy}}" give "Title. Publisher before date. August 1, 2008. "?

Date should remain same.

Same happens with "{{cite book|title=Title|publisher=Publisher before date|date=July 1 – August 1, 2008|df=dmy}}" which gives "Title. Publisher before date. 1 August 2008. " even if date should be 1 July – 1 August 2008 i.e. first part of range should not vanish. --Obsuser (talk) 13:06, 24 August 2017 (UTC)

reformat_dates() does not support date ranges. Never has.
I see that you've made and unmade several changes to Module:Citation/CS1/Date validation/sandbox. Why? To what purpose?
Trappist the monk (talk) 13:16, 24 August 2017 (UTC)
Can it then be disabled for that function, so that if range is present |df= parameter is ignored?
Regarding changes in that module, complete diff is this. Purpose was first to enable three-digit years, and after that when I saw about metadata using captured year I generally gave up on this to be sure because I guess that metadata is ISO so year must be four-digit. At the end of my changes, three-digit years have been enabled for "month day, year" and "day month year" formats – because there is no reason to restrict |date= to four-digit years (this does not affect |accessdate= params), maybe not even to restrict two- or one-digit years, as something might be published e.g. June 854 or even June 54 etc. Also, in "month day – month day, year" and "day month year - day month year" there were minor inconsistencies: in first case, three-digit year was allowed in if match but not in real match for extraction; in second case, vice-versa.
Cite book compare
{{ cite book | accessdate=1 June 854 | publisher=Publisher before date | title=Title | date=1 June 854 | url= }}
Live Title. Publisher before date. 1 June 854. Retrieved 1 June 854.  Check date values in: |access-date=, |date= (help)
Sandbox Title. Publisher before date. 1 June 854. Retrieved 1 June 854.  Check date values in: |access-date=, |date= (help)
--Obsuser (talk) 15:29, 25 August 2017 (UTC)
Is it possible that you are confusing CS1 with a bibliographic system? CS1 exists only to verify claims in articles (and elsewhere) by pointing to sources. The bibliographic history of a source is only pertinent when it affects reliability and verification. Sources should be cited by the date most relevant to the reader who wants to retrieve them. Whether a work was produced in 854 by some publisher means nothing in this context. A reliable (preferably easily retrievable) edition of the work published in 2014 by a certain publisher is [an example of] what should be cited, and there is always |orig-year= if specific older information is deemed relevant. 300 years from now the 2014 edition would probably no longer be accessible; if the source is still used, a contemporary edition/reprint should be used instead. Access dates are a totally different issue and should always be more or less current [at the time the citation is originally inserted]. (talk) 13:58, 26 August 2017 (UTC)
I made the interpolations above, for clarity. (talk) 14:15, 26 August 2017 (UTC)
I do not support the change to allow three-digit years. Almost all three-digit years that I have found are typo errors, things like date= November 3, 214 or date= November 3, 204. In each case, "2014" is meant. – Jonesey95 (talk) 22:53, 27 August 2017 (UTC)
I'm not necessarily opposed to three-digit years. They have their place. I suspect that the need for day and even possibly month precision in a first millennium date is rare. We could limit three-digit years to year-only dates which will allow cs1|2 to handle almost all references and bibliographies.
Trappist the monk (talk) 10:33, 28 August 2017 (UTC)
It is not rare to cite a source published when the Julian calendar was in effect at the time and place of publication. But ISO 8601 does not allow Julian calendar dates. Converting the publication date from Julian to Gregorian would go against long-standing citation traditions outside Wikipedia, and would be hopelessly confusing. Thus citation templates should not emit metadata that asserts publication dates conform to ISO 8601. If that were to be done, an editor encountering an article that cites articles with Julian publication dates would be justified converting all the citations in the article to some other citation format, such as Chicago Manual of Style, to avoid publishing false information. Jc3s5h (talk) 23:07, 27 August 2017 (UTC)
What? Where in this discussion has anyone suggested date conversions between the Julian and Gregorian calendars? Just to be clear, |df= does not do calendar conversion. All it does is format conversion (hence its name).
Trappist the monk (talk) 10:33, 28 August 2017 (UTC)
Obsuser indicated his or her belief that the Cite Q template (and hence, the Citation template) emits ISO dates as metadata. Everyone who mentions "dates" and "ISO" in Wikipedia (with no qualification of which of the many ISO standards is meant) is advocating that all dates be stated in the Gregorian calendar, whether they realize that's what they're advocating or not. I am warning Obsuser about the probably unintended implications of his or her post. Jc3s5h (talk) 12:04, 28 August 2017 (UTC)
Again: What? No one here has mentioned {{cite Q}} except you just now. The first instance of the letter 'q' (or 'Q') occurs with your mention of {{cite Q}}. The reformatting applied by |df= applies to the rendered visual format of the original date but not to the metadata. It is not possible to have |df=ymd with |date=23 June 1254:
{{cite journal|title= Title |date=23 June 1254 |df=ymd}}
"Title". 23 June 1254. 
Dates in the COinS metadata are in a format that looks like that standard's format because that standard's mechanism for date ranges (yyyy-mm-dd/yyyy-mm-dd) is terse and to the point. For dates outside of the Gregorian calendar, only the year is made part of the metadata so that we avoid any need for calendar conversion. YMD dates in cs1|2 look like that standard's format and are Gregorian like that standard but otherwise do not conform to that standard.
In the comments that I wrote in the function make_COinS_date() in Module:Citation/CS1/Date validation I did suggest that these dates conform to that standard. Perhaps I should revise those to say that the COinS date format resembles the standard's format.
Trappist the monk (talk) 14:00, 28 August 2017 (UTC)
Sorry, I shouldn't have mentioned Cite Q in this context. I looked on the main page for metadata information. It does mention that metadata exists, but does not say what standard the metadata follows. I'm unable to understand from your comments above what standard is being followed. Could you please provide a link to the standard? Jc3s5h (talk) 18:44, 28 August 2017 (UTC)
See this discussion.
Trappist the monk (talk) 19:17, 28 August 2017 (UTC)
@Trappist the monk: Can reformatting then be disabled for that function, so that if range is present |df= parameter is ignored?
@ No, CS1 is a sort of bibliographic system. Yes, newer version is desirable but not neccessary (854 one can still be used), and I thing |orig-year= is also date i.e. part of verification system.
@Jonesey95 and Trappist the monk: I agree to limit three-digit years to year-only dates (maybe to allow also month-and-year-only dates).
@Jc3s5h: No need for warning each other; when I said I guess that metadata is ISO so year must be four-digit I meant that if 874 is captured possible ISO date would be "874-06-01" and I guess it is incorrect because it must be "yyyy-mm-dd"; so that's why I gave up on three-digit years wherever there was something like "for metadata" next to year (maybe I'm wrong about this, Trappist the monk knows better what is this metadata actually and would allowing three-digit years there corrupt it). --Obsuser (talk) 12:19, 28 August 2017 (UTC)
Yet one would think that most citable sources produced during the Julian (or any other) calendar period would have contemporary editions in the current calendar. Works that survive centuries usually do. And I don't think CS1 is anything other than what is explicitly stated: a more-or-less standardized, custom collection of guidelines for providing a measure of reliability to an anonymously produced encyclopedia. (talk) 12:27, 28 August 2017 (UTC)
The custom that has emerged in Wikipedia is to cite the date of a paper publication, even if what the editor actually viewed was a verbatim image of it on microfilm, or a verbatim digital image from a source like Google Books. On the other hand, if the editor viewed a newer edition, which was re-typeset and re-paginated, then the date of the newer edition would be cited. There are many publications from the time when the Julian calendar was used that are readily available as digital images. Whether a newer edition would be more reliable than an older edition depends on the nature of the claim being supported. Jc3s5h (talk) 13:02, 28 August 2017 (UTC)
If there is a custom of citing the print publication without crediting of the actual media consulted, such custom should be retired. It contradicts WP:SAYWHEREYOUGOTIT and may prevent readers who want to verify the sources from accessing the claimed source. There are several parameters that can be used (|type=, |via= etc.) or templates ({{cite web}} being an obvious candidate). Apart from that, it seems that dating the source needs additional clarity. There appears to be some confusion regarding work date and publication date, and their respective usage in the |date= field. (talk) 19:05, 28 August 2017 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────WP:SAYWHEREYOUGOTIT states

So long as you are confident that you read a true and accurate copy, it does not matter whether you read the book using an online service like Google Books; using preview options at a bookseller's website like Amazon; on an e-reader (except to the extent that this affects page numbering); through your library; via online paid databases of scanned publications, such as JSTOR; using reading machines; or any other method.

So the custom of citing the date of the print publication rather than the date of the publication that faithfully reproduces the images of the print publication is in absolute conformity to WP:SAYWHERYOUGOTIT. Jc3s5h (talk) 19:11, 28 August 2017 (UTC)

If I'm consulting a scanned copy of a source, say a book scanned into Google Books, I probably won't know when the scanned copy was created to credit a different date of "publication". Since it is a faithful reproduction of the original, I cite it as the original print book that was scanned, and add the Google Books like with |via=Google Books. The "say where you got it" is satisfied by the URL and |via= attribution. Even if I'm consulting a document that was scanned and uploaded onto Wikimedia Commons (like commons:File:AASHTO USRN 1985-06-26.pdf) where I do know when the document was scanned, that date doesn't change the publication details of the source I'm consulting. Instead, I'd argue that giving a 2014 date for a 1985 document would confuse matters in the citation. Imzadi 1979  21:32, 28 August 2017 (UTC)
There are two different aspects here. The first is attribution: a copy of a source should be declared as such. What is cited is the copy, not the original (this applies to later print editions of books as well). If this is how the editor cites the source, then that is how the citation can be verified by a reader. There are parameters in CS1 that make such declarations plain, as discussed above. The second aspect is the fact that work date and publication date have different meanings and serve different purposes. Work date in the present context serves to identify a source whereas publication date, in addition to discovery of the source, offers a means to retrieve it. However either can be used in the |date= field, and this may lead readers to believe that an editor consulted a source from say, 1492. This is improbable, and perhaps not what the average Wikipedia reader (who likely has very little knowledge of the technicalities of citation systems) would expect as proof of an article claim. (talk) 23:00, 28 August 2017 (UTC)

code tweaks[edit]

I have tweaked Module:Citation/CS1/Date validation/sandbox so that date-holding parameters that hold date ranges are skipped.

Cite book compare
{{ cite book | date=July 1 – August 1, 2008 | archive-date=2017-08-29 | publisher=Publisher before date | title=Title | df=dmy-all | archive-url=// | url=// }}
Live Title. Publisher before date. 1 August 2008. Archived from the original on 29 August 2017. 
Sandbox Title. Publisher before date. July 1 – August 1, 2008. Archived from the original on 29 August 2017. 

I have also restored the date checking to 4-digit years for dates with month or with month and day. Simple year, year ranges, and circa year dates continue to support 3- and 4-digit years.

Trappist the monk (talk) 10:45, 29 August 2017 (UTC)

Change for Help:Citation Style 1#Access date[edit]

I suggest changing "Note that access-date is the date that the URL was checked to not just be working..." to "Note that access-date is the date that the URL given in |url= (or |archive-url=, if real URL is dead) was checked to not just be working...". This is because several months ago, I asked someone whether access date can refer to the archive-url or no, and the answer was (as I remember) no – but this is weird because case when there is no access date parameter at all and someone updates article or its references by adding access dates and checking URLs to fix dead ones is possible (so added accessdate can apply for accessing archived version of original URL, which [archived one] can be considered completely valid and generally same [relative to the original one], especially if from Proof that there are such cases is that access date is chronologically after archive date (i.e. date URL became dead) in many cases already; even tracking cat for such cases might be introduced but it wouldn't be for URLs accessed after they became dead (because we cannot know this) but only to know for cases where access dates is after archive date. --Obsuser (talk) 15:40, 25 August 2017 (UTC)

Access date clearly conveys now and should continue to clearly convey that it is associated with the main URL of the citation, not any other URL. Links can and are archived before they become dead. --Izno (talk) 15:53, 25 August 2017 (UTC)
@Izno: what exactly do you mean by the main URL? Citations to chapters, encyclopedia contributions, etc. often only contain |chapter-url=, since the URL to the work as a whole is not useful to readers; indeed quite often there isn't a URL for the whole work: authors of chapters/contributions regularly upload their part when the work as a whole isn't online. Peter coxhead (talk) 17:40, 25 August 2017 (UTC)
The url specified in |url=, as the documentation currently states. --Izno (talk) 19:52, 25 August 2017 (UTC)
I think the documentation at Help:Citation Style 1#Access date is incorrect or out-of-date: it doesn't reflect the way that that the citation templates now behave. |accessdate= without |url= is ignored:
  • {{citation |chapter=Some chapter |title=Some title |accessdate=2017-08-25}} → "Some chapter", Some title 
whereas adding |chapter-url= correctly produces:
  • {{citation |chapter=Some chapter |title=Some title |chapter-url= |accessdate=2017-08-25}} → "Some chapter", Some title, retrieved 2017-08-25 
Peter coxhead (talk) 20:42, 25 August 2017 (UTC)
@Izno: Why do you want to say that it must refer only to the |url=? That is the point of this discussion: proposal to clarify that it can refer to the |archive-url=, |chapter-url= etc. also (and it is not true that it clearly conveys now only to |url=, there are already many cases – as I said – where access date refers to archive url [it is after URL became dead], or chapter url). If I'm wrong, than it would be useful to track cases where accessdate refers to other URL (if possible).
@Peter coxhead: I agree. --Obsuser (talk) 12:25, 28 August 2017 (UTC)

template:cite encyclopedia[edit]

The parameter url-access=subscription isn't displaying properly in {{cite encyclopedia}}. I've had to switch to {{cite web}} when citing limited access sources. The template talk page redirects here for some reason. Pariah24 (talk) 22:44, 26 August 2017 (UTC)

Most of the talk pages for the templates like {{cite book}}, {{cite encyclopedia}}, {{cite journal}}, {{cite web}} etc. redirect here, because they share a lot of code, and problems in one almost always mean problems in some or all of the others. So to avoid pointless duplicated change requests, we have centralised them here. --Redrose64 🌹 (talk) 08:51, 27 August 2017 (UTC)
|url-access= doesn't work with {{cite episode}}, {{cite encyclopedia}}, {{cite conference}}, or {{cite book}} because there is no |chapter-url-access= (and aliases). {{cite encyclopedia}} and the others, except {{cite book}}, internally promote |title= and |url= to |chapter= and |chapter-url= so that the title may be rendered in quotes. Because there is no |chapter-url-access=, |url-access= cannot be similarly promoted.
I have hacked the module sandbox to add |chapter-url-access= so that |url-access= is also promoted.
{{cite encyclopedia/new |title=Ångström (unit), n. |encyclopedia=OED Online |date=June 2017 |publisher=Oxford University Press |access-date=23 August 2017 |url= |url-access=subscription}}
"Ångström (unit), n."Paid subscription required. OED Online. Oxford University Press. June 2017. Retrieved 23 August 2017. 
|chapter-url-access= has the same rules as |url-access=.
The new |chapter-url-access= allows works with {{cite book}}:
{{cite book/new |title=Title |chapter=Chapter |chapter-url=// |chapter-url-access=subscription}}
"Chapter"Paid subscription required. Title. 
Trappist the monk (talk) 10:28, 27 August 2017 (UTC)

Bug when using script-title ISO 639-1 prefix and url-access together?[edit]

Hey I think I found a bug that I figured I should point out. Right now the code:

{{cite journal|last1=Wang|first1=Tianqi|last2=Liang|first2=Geqiu|title=Zhōngguó tángláng mù xīn jìlù shǔ jí yī xīn zhǒng jìshù|journal=Acta Scientiarum Naturalium Universitatis Sunyatseni|date=1995|volume=34|issue=2|pages=84–86|script-title=zh:中国螳螂目新记录属及一新种记述|trans-title=New Record of ''Choeradodis'' and One New Species of Mantodea from China|url=|url-access=subscription|via=[[CNKI]]}}


Wang, Tianqi; Liang, Geqiu (1995). "Zhōngguó tángláng mù xīn jìlù shǔ jí yī xīn zhǒng jìshù" <bdi lang="zh" >中国螳螂目新记录属及一新种记述Paid subscription required [New Record of Choeradodis and One New Species of Mantodea from China]. Acta Scientiarum Naturalium Universitatis Sunyatseni. 34 (2): 84–86 – via CNKI. 

Specifically the link reads:

["Zhōngguó tángláng mù xīn jìlù shǔ jí yī xīn zhǒng jìshù" <bdi lang="zh" >中国螳螂目新记录属及一新种记述]

You'll see the <bdi lang="zh" > hanging out in the middle there.

If you remove the url-access or the script-title ISO 639-1 prefix it works, but they don't seem to like each other.

I imagine this is an easy fix so I'd bring it to people's attention. Thanks :)

Umimmak (talk) 11:53, 28 August 2017 (UTC)

Don't have time to look at the code right now, but looks like the culprit could be a missing closing html tag. (talk) 12:39, 28 August 2017 (UTC)

I think it's not an easy fix in the sense that a simple tweak will fix it. The problem lies in the creation of the external link. The code is at Module:Citation/CS1/sandbox (which fixes other problems with the live module's handling of kerning). Editors whined and complained when the access signal wrapped to another line so we tried adding a non-breaking thin space between the end of the link and the access icon. The results of that experiment were disappointing; it did not work. So we opted for adding <span class="nowrap">...</span> around the last word and the icon. The last word is separated from the other words in the label by a space character. If you look at the whole rendering (simplified from the original) you can see that the code found the last space character in the <bdi lang="zh" > tag and inserted the <span class="nowrap"> tag there:

<cite class="citation journal"><span class="plainlinks">[ "Zhōngguó tángláng mù xīn jìlù shǔ jí yī xīn zhǒng jìshù" <bdi lang="zh" <span class="nowrap">>中国螳螂目新记录属及一新种记述</bdi><span style="padding-left:0.15em">[[File:Lock-red-alt.svg|9px|link=|alt=Paid subscription required|Paid subscription required]]</span></span>]</span>. ''Acta Scientiarum Naturalium Universitatis Sunyatseni''.</cite><span title="ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.jtitle=Acta+Scientiarum+Naturalium+Universitatis+Sunyatseni&rft.atitle=Zh%C5%8Dnggu%C3%B3+t%C3%A1ngl%C3%A1ng+m%C3%B9+x%C4%ABn+j%C3%ACl%C3%B9+sh%C7%94+j%C3%AD+y%C4%AB+x%C4%ABn+zh%C7%92ng+j%C3%ACsh%C3%B9+%E4%B8%AD%E5%9B%BD%E8%9E%B3%E8%9E%82%E7%9B%AE%E6%96%B0%E8%AE%B0%E5%BD%95%E5%B1%9E%E5%8F%8A%E4%B8%80%E6%96%B0%E7%A7%8D%E8%AE%B0%E8%BF%B0&" class="Z3988"><span style="display:none;">&nbsp;</span></span>

Closing that space doesn't fix the problem because the code will simply find the required space between 'bdi' and 'class'.

The solution to this particular problem is not easy for another reason: interleaving html tags is not permitted. What the code is trying to do is this:

<span class="plainlinks">[ "Transcribed Latin text title" <bdi lang="zh">Original language script <span class="nowrap">title</bdi><span style="padding-left:0.15em">[[File:Lock-red-alt.svg|...]]</span></span>]</span>

MediaWiki will rewrite that and put the closing </bdi> some probably-inappropriate place (especially if |script-title= holds a right-to-left script – Arabic, Hebrew, etc).

Were the language something other than Chinese where there were spaces between words we might do this:

<span class="plainlinks">[ "Transcribed Latin text title" <bdi lang="zh">Original language script</bdi> <span class="nowrap"><bdi lang="zh">title</bdi><span style="padding-left:0.15em">[[File:Lock-red-alt.svg|...]]</span></span>]</span>

Yeah, not so simple and not merely a matter of an omitted tag.

Trappist the monk (talk) 16:21, 28 August 2017 (UTC)

Why are we not inserting our own closing tags into the proper places, instead of letting Tidy (or whatever) have a guess at it? --Redrose64 🌹 (talk) 16:32, 28 August 2017 (UTC)
We do insert our own closing tags. Nothing that I have written here suggests that we aren't writing complete markup. But, I have said that the markup that we are writing is malformed. It is malformed because of the way the code is written. At the particular place where we assemble the title, the script title, the url and the access signal, the code does not know about <bdi>...</bdi> tags. Because of that for most scripts, English will do here for an example, it places the <span class="nowrap"> between <bdi lang="en"> and </bdi>:
{{cite journal/new |title=Transcription title |journal=Journal |script-title=en:A title in some other script |url= |url-access=subscription}}
"Transcription title" A title in some other scriptPaid subscription required. Journal. 
The output for that looks like this (coloring added and metadata removed for clarity):
<cite class="citation journal"><span class="plainlinks">[ "Transcription title" <bdi lang="en" >A title in some other <span class="nowrap">script</bdi><span style="padding-left:0.15em">[[File:Lock-red-alt.svg|9px|link=|alt=Paid subscription required|Paid subscription required]]</span></span>]</span>. Journal.</cite>
You can see in the above that the tags are interleaved as I described and that closing tags are present.
From this page's source for the example template we get this (coloring added for clarity):
<span class="plainlinks"><a rel="nofollow" class="external text" href="">"Transcription title" <bdi lang="en">A title in some other <span class="nowrap">script</span><span style="padding-left:0.15em"><img alt="Paid subscription required" src="//" title="Paid subscription required" width="9" height="14" srcset="// 1.5x, // 2x" data-file-width="512" data-file-height="813" /></span></bdi></a></span>
In the above, mediawiki has closed the <span class="nowrap"> tag prematurely and the <bdi>...</bdi> tags enclose not only the script title but also the lock image markup. This latter might have detrimental effects. Or not; but the markup is still wrong in part because we gave it malformed markup in the first place even though that markup had all of its closing tags.
Trappist the monk (talk) 17:33, 28 August 2017 (UTC)
Since this seems to require careful code re-write, shouldn't editors be discouraged from using the ISO codes in |script-title= until a solution emerges? There is |language= as an interim fix. (Unless that too presents a problem). (talk) 18:51, 28 August 2017 (UTC)
I don't think so. I suspect that this particular problem is relatively rare. |language= is not a 'fix' because all that it does is categorize the source as a non-English language source and render the language in the final citation.
Trappist the monk (talk) 09:57, 29 August 2017 (UTC)
That's not it at all. |language= and all other parameters are there to give information to readers about the cited source. In this case, to identify a strange-looking script and provide an important detail about the original source. It is a "fix" only in that sense. The ISO option in |script-title= is a technicality concerning browser rendering. If it breaks the display of the citation for humans. as it does in this case, it has no business there. The focus of CS1 seems hopelessly off. (talk) 14:16, 29 August 2017 (UTC)

Parameter for Wikidata ID, redux[edit]

User:Headbomb and I suggested a parameter for a work's Wikidata ID; there was support, but discussion has been archived. What's happening about this, please? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:50, 28 August 2017 (UTC)

I still support this. But how should the link be presented? WikidataQ21706380? WDQ21706380? WDQID21706380? Q-ID21706380? Something else? Headbomb {t · c · p · b} 17:57, 28 August 2017 (UTC)
One of the first two (if the second, with {{abbr}} or similar), or with a tiny icon. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:49, 28 August 2017 (UTC)
Pigsonthewing (talk · contribs) So I take it we need to declare this via |wikidata=Q21706380. Headbomb {t · c · p · b} 22:09, 28 August 2017 (UTC)
@Headbomb: Yes. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:26, 29 August 2017 (UTC)
For my part, I don't see overwhelming support for this. Some support, yes, but also some opposition. When asked how this new parameter would be useful, Editor Pigsonthewing replied, in part: Furthermore, that identifier can in turn be used to fetch identifiers and other metadata for the publication, the author, publisher et al. which seems contrary to the opinion expressed by Editor Headbomb and seconded by Editor J. Johnson: We should most definitely not draw citation data from wikidata.
Trappist the monk (talk) 11:08, 29 August 2017 (UTC)
I didn't say it would be used to fetch the metadata by the template; but any reader can do so once they know the Wikidata ID; either manually, or by using a tool of their own preference (e.g. this page on Scholia). I don't think either Headbomb (who proposed and supports the addition of the parameter) or J. Johnson objected to that. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:24, 29 August 2017 (UTC)
(EC) Nothing is proposed about drawing information from Wikidata, this would be treated no differently than |doi= or |mr=. Could it be used to fetch stuff from Wikidata? It could, in theory. We might even decide this is desirable in the future. But for now it's simply about given a link to Wikidata, and there was no objection on that. Headbomb {t · c · p · b} 17:26, 29 August 2017 (UTC)
We did have some objections on the simple addition of a new identifier, and I subscribe to Jc3s5h's comments on that. I do not think all unique identifiers are worth displaying to our readers in citations : we should only include well-established bibliographical identifiers that readers will find useful. I suspect many readers would be annoyed to see yet another unique id they do not care about popping up in citations. Just to be clear, I personally love Qids, but I am just not sure this is the right place for them. As a random reader, what does it bring to me? I can click on that identifier, and see a page with the metadata of that citation on a Wikidata. Fine, but I already had the metadata in the citation. Not everybody is a Linked Open Data enthusiast who will experience a warm and fuzzy feeling only at the sight of a Qid… − Pintoch (talk) 19:12, 29 August 2017 (UTC)
The relevant question is "is this useful in relation to the purpose of a citation?" Citations are cluttered already with IDs and potential IDs (review Help:Citation Style 1#Identifiers). I haven't yet seen a strong case in relation to the purpose of a citation, as opposed to information that might be useful to someone. Peter coxhead (talk) 20:59, 29 August 2017 (UTC)
A Wikidata ID might constitute an opportunity to move all of those IDs elsewhere and either a) leave users to investigate themselves or b) pull the identifiers from Wikidata automatically with each invocation (even if you don't want to pull an entire citation from Wikidata). --Izno (talk) 22:08, 29 August 2017 (UTC)
Using a Wikidata ID to replace many of the others might be useful, I agree, although there would need to be discussion on which ones. Peter coxhead (talk) 22:18, 29 August 2017 (UTC)
You ask "As a random reader, what does it bring to me?"; just above your question, I wrote "any reader can do so [fetch identifiers and other metadata for the publication, the author, publisher et al] once they know the Wikidata ID; either manually, or by using a tool of their own preference (e.g. this page on Scholia)". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:03, 29 August 2017 (UTC)
fetch identifiers: CS1/2 already supports loads, the author, publisher et al: that's also easy to include in the citation template, either manually, or by using a tool of their own preference: CS1/2 templates are already inter-operable with many tools thanks to COinS… Again I am playing the devil's advocate here, but I think people are just very likely to reject this change. We should be very careful not to foster the skepticism that already exists around Wikidata among some Wikipedia editors. Changes bringing more Wikidata integration to Wikipedia should bring real value to the community (e.g. better integration in infoboxes) instead of splashing our Wikidata ids all over the place for no apparent benefit. − Pintoch (talk) 06:13, 30 August 2017 (UTC)
CS1/2 already supports loads, the author, publisher et al... What? cs1|2 does not support 'loading' any data from anywhere any other than the template's wikisource. What is it that you really mean?
The opportunity, it seems to me, for integration of cs1|2 and wikidata is best begun by making {{cite Q}} a sterling exemplar of correct use of that resource. Alas, I fear that the opportunity is slipping away. {{cite Q}} could be written to enforce best practices to ensure that the underlying data at wikidata are properly curated. Unfortunately, data deficiencies are being 'fixed' by tweaks to the template code rather than the correct fix to the data source. If {{cite Q}} becomes recognized as a quality tool, then perhaps there is a future for wikidata in cs1|2. But, if slipshod craftsmanship of {{cite Q}} is allowed to continue, I don't hold out much hope for wikidata in cs1|2.
Trappist the monk (talk) 11:31, 30 August 2017 (UTC)
By already supports loads, I meant "CS1/2 already supports loads of identifiers" (it has support for a lot of identifiers). Sorry about that! − Pintoch (talk) 12:35, 30 August 2017 (UTC)
True, you didn't say it would be used to fetch the metadata by the template but you did not say that the identifier would be a link only; you did not say that the identifier would not be used by the templates to fetch metadata from wikidata. Because this discussion is about modifying cs1|2 to support a wikidata identifier, don't you know that editors might understand your statement to mean: "once implemented, the templates will be able to fetch metadata from wikidata"? Without a statement to the contrary, why shouldn't they draw that conclusion?
Trappist the monk (talk) 11:08, 30 August 2017 (UTC)
What could the purpose of such a parameter be but to set up future fetches from wikidata? The link itself is not something that would be of much use to readers. —David Eppstein (talk) 12:54, 30 August 2017 (UTC)
Wikidata contains a lot more metadata than what we include in our citations. ORCIDs for authors for instance. We also typically leave out ISSNs, publishers, etc... when citing journal articles. There are lots of benefits beyond data fetching. Headbomb {t · c · p · b} 13:29, 30 August 2017 (UTC)
Sure − so Qids could be useful in COinS, for instance (or even just in wikicode, maybe). But is it really worth displaying that to human readers? − Pintoch (talk) 13:43, 30 August 2017 (UTC)
Yes. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:17, 30 August 2017 (UTC)
(edit conflict) Now there's a novel idea: emit all of the identifiers as metadata in COinS, but potentially leave (some of) them out of the displayed version of the citation. Imzadi 1979  15:16, 30 August 2017 (UTC)
No, hiding data isn't a novel idea, it's often been suggested, and rejected as harmful, because errors are hidden. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:29, 31 August 2017 (UTC)
For the third time: "any reader can do so [fetch identifiers and other metadata for the publication, the author, publisher et al] once they know the Wikidata ID; either manually, or by using a tool of their own preference (e.g. this page on Scholia)" Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:17, 30 August 2017 (UTC)
Andy, I think we got your point, repeating it verbatim will not make it more convincing. Do you want me to repeat why I think the use cases you are talking about are not useful to the average reader? I can rephrase if that was unclear. But let's avoid being too assertive and have a constructive discussion together! − Pintoch (talk) 16:48, 30 August 2017 (UTC)
Well, if people could stop acting as though no such argument had been presented... Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:08, 30 August 2017 (UTC)

Another reason to include Wikidata IDs is that bots can compare what's in the templates to what's on Wikidata, and alert humans to discrepancies. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:29, 31 August 2017 (UTC)

Hmm, again I'm not sure that is very useful: many citations already have some id, so we can compare the metadata in the template and the data associated to that id. In the vast majority of cases, the data that is on Wikidata was created from one of these sources by a tool (such as fatameh), so it's not like we are getting access to a new data source. Granted, in some cases, a human editor might have added some information (such as disambiguating an author), but that seems to be very rare for now. There would also be the possibility to transfer authorlinks from Wikipedia citations to Wikidata items (to disambiguate authors there), but again that is something we can already do based on the existing identifiers. In any case, as you point out, these use cases would be for bots, so I do not see the point of displaying the id to human readers.
Another idea: if Scholia is the tool you want to access from Wikipedia, what about putting directly a link to that tool? Instead of adding something like "WD: Q38197781", it would add "ScholiaQ38197781"?
Anyway, I think there is a simple way to trial your idea and show that the community is not going to reject it. Just create an id template, say {{Scholia}}, along the lines of {{doi}} or {{arXiv}}, and add it to citations in the |id= field:
{{Cite journal| doi = 10.1016/j.genhosppsych.2013.11.007| issn = 0163-8343| volume = 36| issue = 3| pages = 310–317| last1 = Dobscha| first1 = Steven K.| last2 = Denneson| first2 = Lauren M.| last3 = Kovas| first3 = Anne E.| last4 = Corson| first4 = Kathryn| last5 = Helmer| first5 = Drew A.| last6 = Bair| first6 = Matthew J.| title = Primary care clinician responses to positive suicidal ideation risk assessments in veterans of Iraq and Afghanistan| journal = General Hospital Psychiatry| accessdate = 2017-09-01| date = 2014-05-01| url = | id={{Scholia|Q38197781}} }}
Dobscha, Steven K.; Denneson, Lauren M.; Kovas, Anne E.; Corson, Kathryn; Helmer, Drew A.; Bair, Matthew J. (2014-05-01). "Primary care clinician responses to positive suicidal ideation risk assessments in veterans of Iraq and Afghanistan". General Hospital Psychiatry. 36 (3): 310–317. ISSN 0163-8343. doi:10.1016/j.genhosppsych.2013.11.007.  ScholiaQ38197781
If there is wide adoption for this, it will be easy to create the id in CS1/2 and migrate the ids to native parameters. This is what has happened to {{CiteSeerX}}, for instance. CiteSeerX was already used a lot in |id= before it became natively supported, and I migrated the |id={{citeseerx}} to |citeseerx= with a simple regular expression. − Pintoch (talk) 08:11, 1 September 2017 (UTC)
OK, I've knocked something up as {{Scholia}}. It lacks error trapping, which will be needed before widespread use, and I'd probably replace the text "Wikidata" with a tiny icon (and maybe do the same for "Scholia", once an icon is avilable). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:12, 11 September 2017 (UTC)

Please don't display any IDs from unreliable sites (whether Wikidata, Scholia, Quora, Findagrave, or whatever else you can come up with). Cite should be used to link to reliable sources and repositories, not user-generated or otherwise unreliable ones. Fram (talk) 11:51, 11 September 2017 (UTC)

Not in favor per Fram. Ealdgyth - Talk 14:08, 11 September 2017 (UTC)
All citations are user-generated and no more reliable or unreliable than the person who types them in, whether that be on Wikipedia or elsewhere. Making use of a central repository for sources (such as Wikidata) helps keep citations consistent and reduces typos and transcription errors. We should be very much in favour of efforts to stop typing in a hundred versions of the same cite when one is sufficient. --RexxS (talk) 17:20, 13 September 2017 (UTC)

Truthful publication dates[edit]

In Help_talk:Citation_Style_1/Archive_10#Date_metadata Trappist the monk indicates that metadata is emitted for publication dates, and the metadata is meant to be in the COinS format, and a series of references leads to this date format description. That document in turn refers to ISO 8601, which only allows the use of the Gregorian calendar and Proleptic Gregorian calendar. I voiced the concern in the original discussion that any date before Thurday 1 March 1923 should not be emitted as metadata, because that is the date that the last country (Greece) switched from the Julian to the Gregorian calendar.

Citation Style 1 adopts the date formats from Wikipedia:Manual of Style/Dates and numbers. That manual clearly allows Julian calendar dates: "A date can be given in any appropriate calendar, as long as it is (at the minimum) given in the Julian calendar or the Gregorian calendar or both, as described below." Quite a few articles about events between 1582 and 1923 contain footnotes stating which calendar is used in the article; automatically putting (without date conversion) dates into the metadata is likely to contradict these footnotes.

All the documentation that any reasonably diligent editor will read indicates Julian calendar publication dates between 1583 and 1923 may be, and in some case should be, in the Julian calendar. The templates at present silently change the meaning of what the editor has written.

Thus I call for the citation templates to not emit metadata for publication dates before 1 March 1923 of the Gregorian calendar, or at the very least, only give the year. Jc3s5h (talk) 20:11, 28 August 2017 (UTC) Fixed first link 07:12 29 August 2017 (UT)

I think that the real link that Editor Jc3s5h meant to provide is this one: Help_talk:Citation_Style_1/Archive_10#Date_metadata.
Trappist the monk (talk) 09:52, 29 August 2017 (UTC)

Access-level parameters[edit]

I'm wondering if a non-breaking space should be placed before the access-level icon. It seems odd for the icon to be separated from its link, especially if the icon winds up on a line of its own.

Additionally, is there any reason that registration and limited are not acceptable values for the doi-access and jstor-access parameters (and perhaps bibcode-access, hdl-access, ol-access, and osti-access, though I'm not particularly familiar with them)? Of course subscription isn't accepted since it's assumed that content that has a DOI or that is on JSTOR requires a subscription unless otherwise indicated, but what about, for example, the occaisional paper that is accessible for free on JSTOR provided one registers for a free account?

Any thoughts? (talk) 20:55, 30 August 2017 (UTC)

  1. There should be a non-breaking space yes. Trappist the monk (talk · contribs) has been working on this, but I don't know what came out of it.
  2. Those should be supported per the RFC we held, but they haven't been rolled out yet. The main reason to not support them was that doi/jstor are usually closed access, and thus "flagging what is normal" was seen as undesirable. The RFC did conclude that they should at least be allowed, and then editors can chose to flag this or not.
  3. |bibcode-access=free/|doi-access=free/|osti-access=free/|jstor-access=free etc all work already
Headbomb {t · c · p · b} 21:08, 30 August 2017 (UTC)
I presume that you are referring to this template:
{{cite journal |last1=Hill |first1=I. D. |last2=Wichmann |first2=B. A. |last3=Woodall |first3=D. R. |year=1987 |title=Algorithm 123: Single Transferable Vote by Meek's Method |journal=The Computer Journal |volume=30 |issue=3 |pages=277–281 |issn=1460-2067 |doi=10.1093/comjnl/30.3.277 |doi-access=free |ref=harv }}
If you look at the code produced from that you will see this bit:
<span class="plainlinks">[// 10.1093/comjnl/30.3.277]&#8239;[[File:Lock-green.svg|9px|link=|alt=Freely accessible|Freely accessible]]</span>
In the midst of that you will find this:
which is a narrow no-break space. As you have discovered, it does not work. I suspect that it doesn't work because the access signal is an image and not a letter or number.
Trappist the monk (talk) 21:56, 30 August 2017 (UTC)
That's precisely the citation I had in mind. Do you know if there is any way of getting it to work with the image? (Despite the fact I was citing The Computer Journal, I'm afraid that my background is definitely not in computer science.) (talk) 22:22, 30 August 2017 (UTC)
@Headbomb: I can see how "flagging what is normal" might be considered undesirable, but the case I'm discussing isn't "flagging what is normal". JSTOR, for instance, requires a subscription for most content (that I've seen, at least), so I'm just hoping to be able to tag those less common cases where registration or limited are appropriate. And nothing I was referring to should affect the free value. (talk) 22:22, 30 August 2017 (UTC)
Non-breaking spaces - and the white-space:nowrap; declaration - are intended for use when there are two pieces of text which should not be divided. Access icons are images, and images are not text. Some browsers may prevent wrapping when text is "joined" to an image in this manner, but this behaviour is not documented and must not be relied upon. --Redrose64 🌹 (talk) 08:31, 31 August 2017 (UTC)
Well, although "registration" means "a free registration is required to access the source" I'm not sure it gives the right impression for the JSTOR links you refer to, since, since you both need to make an account and you can only read three articles within a 14 day period and you can't download the article like you'd be able to do with the truly free ones / if you paid for it, just see one page at a time. It's a lot more jumping through hoops than something where you make an account and you access it for free. To me "registration" means you have normal unlimited access, you just need to make an account once. Umimmak (talk) 15:33, 31 August 2017 (UTC)

Should the template remove redundant periods?[edit]

E.g. "Torvalds, Linus (May 9, 1996). "Re: Linux Logo prototype.". Archived".. I believe the MOS allows the latter removed, not the former), is there an exception for refs? I had no luck with postscript=none (strangly did nothing?!) and mode is for other. Any alternative? Best would be if the template did it in agreement with the MOS(?). comp.arch (talk) 11:24, 1 September 2017 (UTC)

Please provide a sample template. I believe that some periods are already ignored. – Jonesey95 (talk) 13:12, 1 September 2017 (UTC)
It's this one, I think.
There is a function in Module:Citation/CS1 called safe_join() that is supposed to remove duplicate punctuation when that punctuation matches the template's separator character (in this case a dot). But, safe_join() apparently doesn't have a case for ."]. So I added one:
Cite web compare
{{ cite web | last=Torvalds | deadurl=yes | date=May 9, 1996 | first=Linus | df=mdy-all | title=Re: Linux Logo prototype. | archivedate=May 30, 2012 | url= | archiveurl= }}
Live Torvalds, Linus (May 9, 1996). "Re: Linux Logo prototype.". Archived from the original on May 30, 2012. 
Sandbox Torvalds, Linus (May 9, 1996). "Re: Linux Logo prototype". Archived from the original on May 30, 2012. 
But, that doesn't look quite right to me. Compare the above to a simple example without the separator character inside the quotes (this is the norm):
"Title". Example. 
So, perhaps when |title= has a trailing separator character, we should remove the title's trailing character rather than remove the template's separator character.
Trappist the monk (talk) 13:42, 1 September 2017 (UTC)

Article on multiple pages, continues under a different title[edit]

Hi, guys! I would like to cite the material on page 10 of an article that appears on page 1 as "Girl Kisses Boy", but is continued on page 10 as "Boy Cries After Being Kissed". Normally I would use the first title and note "pp. 1, 10" to indicate that the article covers multiple pages. Should I be using the second title and just "p. 10"? Thanks! -Location (talk) 02:16, 8 September 2017 (UTC)

Or treat the 2nd as a subtitle ? —TheDJ (talkcontribs) 07:43, 8 September 2017 (UTC)
Do you mean |title=Girl Kisses Boy; Boy Cries After Being Kissed with |page=10? -Location (talk) 15:04, 8 September 2017 (UTC)
|title= should be the title of the work as it appears in the contents/title page. That is the way most works are indexed and that is the way they can be found. You can use |at="§ Boy Cries After Being Kissed". p. 10. (talk) 13:58, 10 September 2017 (UTC)
If we're talking about a newspaper article, they typically use a shorter or different title for the continuation. At the bottom of the section on page 1, it would say "See 'Boy Cries After Being Kissed', p. 10" or similar and then at the top of the section on page 10, it would say "'Boy Cries After Being Kissed' from p. 10" or similar. Personally, I cite it under the main title on page one and cite the two pages together. Unless the article is especially long, the information being cited should be easily located whether it is on either page. Imzadi 1979  16:52, 10 September 2017 (UTC)
Thanks for the replies. On multiple occasions I have had cited material reverted because someone didn't pay attention to the fact that the cited material within the linked newspaper article was continued on a second page. I guess that not really my fault, but I was hoping to cite in a way to prevent that from happening. Thanks again! -Location (talk) 02:45, 11 September 2017 (UTC)

Language parameter[edit]

Hello! There's an apparent oversight with the "language" parameter. There are many journals, proceeding, collective works, etc. that have content in various languages, and the language cited should refer to a specific entry. Currently, when one uses {{cite journal}} or {{cite book}} with the "chapter" parameter, the "language" parameter appears next to the journal or book title, respectively, whereas it should appear next to the article title or chapter title, because that is where it is relevant. Constantine 14:20, 8 September 2017 (UTC)

|language= is not attached to any particular title-holding parameter; readers are, I guess, expected to infer that it applies to the most specific title. For example, if |type= or |series= is included in a {{cite journal}} template:
{{cite journal |title=Teitl |journal=Journal |type=Type |series=Series |language=cy}}
"Teitl". Journal (Type). Series (in Welsh). 
There are a bazillion cs1|2 parameters. Figuring out how to render each parameter in its optimal position according to which of the other bazillion parameters are selected for the particular citation is an onerous task; one that will likely never be accomplished.
Trappist the monk (talk) 14:46, 8 September 2017 (UTC)
Granted that there are many parameters, but presumably the order of appearance is related to the order in which the parameter is defined. So why not define it right after title and/or chapter? Alternatively, or in addition, why introduce a "chapter-languange" parameter, analogous to "chapter-url"? Constantine 18:02, 8 September 2017 (UTC)
Order of appearance and order of definition are wholly unrelated. Parameters are rendered in an order that somewhat resembles the order established by external style guides like Chicago, ALA, etc.
Trappist the monk (talk) 09:45, 9 September 2017 (UTC)

Date error not flagged[edit]

There appears to be no flagging of a date error when {{cite web}} is called from {{Kilde www}}

As an example {{Kilde www|dato=26.-28. April 2002|publisher=Second MIT Conference on Technology, Archaeology, and the Deep Sea, Bosten, MA Presented at MIT, Cambridge, Mass. 26–28 April 2002}} renders as "[no title]". Second MIT Conference on Technology, Archaeology, and the Deep Sea, Bosten, MA Presented at MIT, Cambridge, Mass. 26–28 April 2002 (26.-28. April 2002).

This should give a date error for the dots in the date and another for the dash. Keith D (talk) 17:15, 8 September 2017 (UTC)

If one is to believe the 'Pages transcluded onto the current version of this page' list at the bottom of this page, the cs1|2 module suite is not used by that template. This is confirmed by looking at your example this way:
"[no title]". Second MIT Conference on Technology, Archaeology, and the Deep Sea, Bosten, MA Presented at MIT, Cambridge, Mass. 26–28 April 2002&#32;(26.-28. April 2002).
and comparing that to this rendering which does use {{cite web}}:
<cite class="citation web">. Second MIT Conference on Technology, Archaeology, and the Deep Sea, Bosten, MA Presented at MIT, Cambridge, Mass. 26–28 April 2002. 26.-28. April 2002.</cite><span title="ctx_ver=Z39.88-2004&" class="Z3988"><span style="display:none;">&nbsp;</span></span> <span style="font-size:100%" class="error citation-comment">Check date values in: <code style="color:inherit; border:inherit; padding:inherit;">&#124;date=</code> ([[Help:CS1 errors#bad_date|help]]); </span><span style="font-size:100%" class="error citation-comment">Missing or empty <code style="color:inherit; border:inherit; padding:inherit;">&#124;title=</code> ([[Help:CS1 errors#citation_missing_title|help]]); </span><span style="display:none;font-size:100%" class="error citation-comment">Missing or empty <code style="color:inherit; border:inherit; padding:inherit;">&#124;url=</code> ([[Help:CS1 errors#cite_web_url|help]])</span>
. Second MIT Conference on Technology, Archaeology, and the Deep Sea, Bosten, MA Presented at MIT, Cambridge, Mass. 26–28 April 2002. 26.-28. April 2002.  Check date values in: |date= (help); Missing or empty |title= (help);
If one looks into the source for {{kilde www}}, there are two mentions of cite web one of which is in a comment but the other is not. That code is sufficiently complex that, without I take time that I haven't got right now, I cannot explain. I would have suggested that a conversation with Editor Jimp (who authored the cite web section of that code would be in order but that editor has not edited since March).
There are relatively few instances of this template so perhaps the right thing to do is to upgrade those instance to use the correct cs1|2 template (the example here might want to be {{cite conference}}).
Trappist the monk (talk) 18:27, 8 September 2017 (UTC)
(edit conflict) It appears that cite web is involved only when the template is substed. That template, and other foreign-language citation-translation templates, should be set up to auto-substitute, like {{Internetquelle}}, but the date substitution is broken. I have posted on the template's talk page to see if anyone is willing to fix the date substitution bugs. – Jonesey95 (talk) 18:31, 8 September 2017 (UTC)
Thanks, could not work out the code. Keith D (talk) 23:56, 8 September 2017 (UTC)

Dragon magazine mess[edit]

Dragon has a somewhat nonstandard scheme, whereas it has an issue number, a volume number, a number number, and then pages.


  • Dobson, Michael (March 1986). "The Forum" Dragon. #107 Vol. 10 no. 10. p. 6.

Now in the wild, this is often cited as

  • Dobson, Michael (March 1986). "The Forum". Dragon #107. Vol. 10 no. 10. p. 6. 
  • Dobson, Michael (March 1986). "The Forum". Dragon #107. 10 (10): 6. 
  • Dobson, Michael (March 1986), "The Forum", Dragon #107, 10 (10): 6 

which has the annoying tendency to put the issue number in the journal field. How do we fix this?

Three options exist, IMO

  • Add a hack, so that for |magazine=/|work=/|journal= = The Dragon / Dragon / Dragon Magazine, that we allow both |issue= and |number=
  • Add |num-issue=yes, letting the template know that those are distinct fields.
  • Shove both issue/number in |issue= or |number=, i.e. |number=3, #111 to create
    • Dobson, Michael (March 1986). "The Forum". Dragon. Vol. 10 no. 10 #107. p. 6. 
    • Dobson, Michael (March 1986). "The Forum". Dragon. 10 (10 #107): 6. 
    • Dobson, Michael (March 1986), "The Forum", Dragon, 10 (10 #107): 6 

What should be done? Headbomb {t · c · p · b} 20:31, 10 September 2017 (UTC)

I'm not a template programmer and don't have a strong opinion on the API. But for Dragon and Dragon+ magazines, looking at bibliographic resources such as DragonDex suggests that the issue number and page are what readers really use when locating articles; volume and number aren't even mentioned. So putting the issue number first, as in your 'in the wild" examples, seems the way to go. As a reader, I'm not bothered by the issue number being in the journal field, but understand for database purposes, it is better to have the issue number as a separate field. --Mark viking (talk) 21:01, 10 September 2017 (UTC)
I agree not one [really] cares about vol/num for Dragon magazine. I suppose we could just purge volume/number from those citation, and do something like
  • Dobson, Michael (March 1986). "The Forum". Dragon. No. #107. p. 6. 
  • Dobson, Michael (March 1986). "The Forum". Dragon (#107): 6. 
  • Dobson, Michael (March 1986), "The Forum", Dragon (#107): 6 
But the output is a bit misleading on some of them (if using {{cite magazine}}). Headbomb {t · c · p · b} 21:21, 10 September 2017 (UTC)
I support the idea of permitting both issue and number as separate parameters when both appear (and with prescribed uses for each). This isn't difficult, and I've encountered many circumstances where this would be useful, even going back to periodicals from the late Victorian era. This would also solve the common problem of issues having designations like "Winter" and "Beatles Commemorative Edition" and whatnot. These could go in |issue=, with |number= used for the number within the |volume=, when both |number= and |issue= are used, but |issue= otherwise being treated as an alias of |number=.

Failing that, I guess one can overload the current number/issue parameter: |volume=XI|issue=3 (Summer). Or, to use the first Dragon example: |volume=10|issue=10 (#107); it just seems a little messy and potentially confusing, even at the source level.
 — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  23:14, 10 September 2017 (UTC)

Generalize Template:Cite report[edit]

Propose merging {{Cite techreport}} and {{Cite press release}} to {{Cite report}}, and adding a |type= parameter that sets the descriptive wording, e.g. to "report", "technical report", "press release", "whitepaper", "standard", "specification", "form", etc. This would provide for easier and more precise citation of a wide range of governmental and NGO output, without polluting the metadata output of |title=, |work=, or wherever people are trying to randomly insert such things. Also, these descriptive terms should not be capitalized per MOS:CAPS; they're being overcapitalized in the [at least] three redundant templates, with output of "(Technical report)", "(Press release)", and "(Report)", respectively. And the title needs to be either in italics or quotation marks; I would suggest italics (at least as a default), since this is for stand-alone items, though perhaps there could be a switching parameter; some citation styles may demand quotation marks for works under a certain length.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  23:04, 10 September 2017 (UTC)

I disagree with merging press releases in; those are short-form works that shouldn't be in italics, while reports are typically longer-form works that would have italicized titles. (Personally, I gave up on using cite report and just use {{cite book}} with |type=Report.) As for the descriptive terms, those should remain capitalized. It's quite common for those to be capitalized in the citation styles. Imzadi 1979  23:47, 10 September 2017 (UTC)

Possible sandbox bug with cite interview[edit]

I stumbled across this Lua error on Template:Cite interview/testcases:

Cite interview compare
{{ cite interview | date=August 23, 2005 | publisher=Україна Молода | subject=Milla Jovovich's mother | title=Галина Логінова: Сьогодні Київ, наче весела дiвчина, гарно вбрана i нафарбована | url= | language=Ukrainian | accessdate=May 15, 2013 }}
Live Milla Jovovich's mother (August 23, 2005). "Галина Логінова: Сьогодні Київ, наче весела дiвчина, гарно вбрана i нафарбована" (Interview) (in Ukrainian). Україна Молода. Retrieved May 15, 2013. 
Sandbox Milla Jovovich's mother (August 23, 2005). "Галина Логінова: Сьогодні Київ, наче весела дiвчина, гарно вбрана i нафарбована" (Interview) (in Ukrainian). Україна Молода. Retrieved May 15, 2013. 
in Module:Citation/CS1 |language= follows title

I don't know how long it has been there or what the error means. Here's a direct call to cite interview/new, showing that the problem is in cite interview, not in cite compare:

And here is the citation with only |title=:

Cite interview compare
{{ cite interview | title=Галина Логінова: Сьогодні Київ, наче весела дiвчина, гарно вбрана i нафарбована }}
Live "Галина Логінова: Сьогодні Київ, наче весела дiвчина, гарно вбрана i нафарбована" (Interview). 
Sandbox "Галина Логінова: Сьогодні Київ, наче весела дiвчина, гарно вбрана i нафарбована" (Interview). 
just |title=

Any ideas? – Jonesey95 (talk) 18:33, 11 September 2017 (UTC)

This occurs in cite web as well:

Cite web compare
{{ cite web | date=August 23, 2005 | author=Milla Jovovich's mother | publisher=Україна Молода | title=Галина Логінова: Сьогодні Київ, наче весела дiвчина, гарно вбрана i нафарбована | url= | language=Ukrainian | accessdate=May 15, 2013 }}
Live Milla Jovovich's mother (August 23, 2005). "Галина Логінова: Сьогодні Київ, наче весела дiвчина, гарно вбрана i нафарбована" (in Ukrainian). Україна Молода. Retrieved May 15, 2013. 
Sandbox Milla Jovovich's mother (August 23, 2005). "Галина Логінова: Сьогодні Київ, наче весела дiвчина, гарно вбрана i нафарбована" (in Ukrainian). Україна Молода. Retrieved May 15, 2013. 
in Module:Citation/CS1 |language= follows title

--Izno (talk) 20:01, 11 September 2017 (UTC)

TfD notification about Template:Cite Q[edit]

There is a discussion here which may be of interest to editors of this page. Nikkimaria (talk) 00:10, 16 September 2017 (UTC)

Just to note that this is about Template:Cite Q. Thanks. Mike Peel (talk) 01:03, 16 September 2017 (UTC)

Legislation template[edit]

reference info for Railways in Melbourne
unnamed refs 37
named refs 4
self closed 2
bare url refs 1
bare ext link refs 1
cs1 refs 21
cs1 templates 21
cleanup templates 1
dead link templates 1

Hi, I originally posted this at the Help Desk but realised here is probably a better venue. I was taking a look at Railways in Melbourne with the intent of cleaning up its citations, and given that there's a fair few CS1 templates I thought that might as well be the eventual style. However, this article cites quite a few pieces of Victorian legislation, and I wasn't sure about the best way to provide a CS1 citation for these sources. There's a template, {{Cite Legislation AU}}, but I'm not sure if it's appropriate. Happy to do it manually too but I'll need some guidance on the correct style. Triptothecottage (talk) 21:31, 20 September 2017 (UTC)

I agree that {{Cite Legislation AU}} is not a good fit because it is a source-specific template that is locked to the Austlii website. cs1|2 are not really good at legislative citations because they are general purpose tools that are pretty good at rendering citations for the most commonly cited stuff: books, magazines, newspapers, journals. It can be done, though. Rewriting this one:
{{cite web |url=$FILE/96-79a050.pdf |title=Rail Management Act 1996 |website=Victorian Legislation and Parliamentary Documents |publisher=Victoria State Government |version=0.50 |date=1 July 2010}}
"Rail Management Act 1996" (PDF). Victorian Legislation and Parliamentary Documents. 0.50. Victoria State Government. 1 July 2010. 
compared to:
{{Cite Legislation AU|Vic|act|rma1996140|Rail Management Act 1996}}
Rail Management Act 1996 (Vic)
Note that these two citations, while purporting to cite the same thing, do not. The former cites the version of the legislation dated 1 July 2010 while the latter cites the presumably current version dated 12 April 2017. I suspect that {{Cite Legislation AU}} will always do that so important bits of a source might get legislated away in future making the source useless for Wikipedia's purposes.
Similar to the above, ref 33, ref 34, ref 36, ref 38, ref 39, and ref 41 can all use {{cite web}} as I did above. Ref 35 is referencing a Wikipedia article which it should not do so you might want to find a better source.
Trappist the monk (talk) 22:35, 20 September 2017 (UTC)
@Trappist the monk: Thanks for all that. I think that's a good point about the dynamic nature of the Austlii citation. (I hadn't even noticed the internal link in 35!) I'll get around to using CW for all of them. Triptothecottage (talk) 03:17, 21 September 2017 (UTC)

high risk[edit]

So there's this template at the top of several of the various cs1|2 template documentation pages. Except for {{cite web}}, they look more or less like this from {{cite book}}:

{{high-risk| 846000+ }}

Recently Editor Mr. Guye changed the template at {{cite web}} to look like this:

{{High-risk|2622800+ pages, which is ≈ {{#expr:(262280100/{{NUMBEROFPAGES:R}}) round 0}}% of all}}

I reverted because the percentage-use is not really necessary to convey the fact that {{cite web}} is widely used but also because the calculation used to arrive at that percentage is misleading so the result is mostly meaningless. Between then and now Editor Mr. Guye has reverted me with this edit summary: "This is what they do on many "High-risk" templates. It's not just something I decided to invent."

To set the record straight, I have not accused Editor Mr. Guye of inventing anything. I do not know who 'they' are so cannot speak to that part of the argument.

While it is possible, I suppose, to use {{cite web}} in all namespaces, counting all pages in all namespaces misleads editors by artificially reducing the apparent usage in the few namespaces where it matters and foremost among them is article space.

Were Editor Mr. Guye to have chosen a more appropriate base for the calculation I might not have reverted (even though I still think that the statistic is not necessary). For example:

{{NUMBEROFARTICLES}} → 5,480,929

instead of

{{NUMBEROFPAGES}} → 43,142,488

Rewriting the calculation to use {{NUMBEROFARTICLES:R}} gives a better representation of {{cite web}} use where it matters most:

{{#expr:(262280100/{{NUMBEROFARTICLES:R}}) round 0}}% → 48%

I've written all of this because WP:BRD

Trappist the monk (talk) 18:05, 21 September 2017 (UTC)

Pretty sure there was a discussion last year about the {{high risk}} template, where it was decided that two significant figures was quite sufficient; so 2600000 is no more precise than necessary. --Redrose64 🌹 (talk) 18:29, 21 September 2017 (UTC)