Help talk:Citation Style 1

Add topic
Active discussions
(Redirected from Template talk:Cite web)
Citation templates
... in conception
... and in reality

Appropriate cite template for a US SEC filingEdit

Is there a template that would be most appropriate to cite material in a U.S. Securities and Exchange Commission that is online that isn't {{cite web}}? - Favre1fan93 (talk) 15:26, 26 August 2022 (UTC)

Probably cite document. It would help to know what you're trying to cite though. Headbomb {t · c · p · b} 15:35, 26 August 2022 (UTC)
@Headbomb: This and this. - Favre1fan93 (talk) 15:43, 26 August 2022 (UTC)
I'm not sure those should be cited at all to be quite honest. WP:PRIMARY sources are generally iffy (especially ones marked 'confidential'). But cite document would probably be the thing if there's consensus that those are appropriate sources. Headbomb {t · c · p · b} 15:47, 26 August 2022 (UTC)
Okay thank you! - Favre1fan93 (talk) 16:05, 26 August 2022 (UTC)
(edit conflict)
I'm not convinced that {{cite document}} is the correct template. {{cite document}} is a redirect to {{cite journal}} which I think it the wrong template for this use case. Ignoring WP:PRIMARY, a valid concern, I presume that you will not leave readers to search the entire 90-odd/30-odd pages for the one sentence that supports whatever article where you plan to use these sources. So, consider {{citation}} (using |section=Item 2. Properties as an example:
{{citation |mode=cs1 |section=Item 2. Properties |title=United States Securities and Exchange Commission Form 10-K: Annual Report |page=13 |date=March 1, 2007 |publisher=Marvel Entertainment, Inc. |url= |via=Securities and Exchange Commission}}
"Item 2. Properties". United States Securities and Exchange Commission Form 10-K: Annual Report. Marvel Entertainment, Inc. March 1, 2007. p. 13 – via Securities and Exchange Commission.
Trappist the monk (talk) 16:13, 26 August 2022 (UTC)
It's a report? {{cite report}} Izno (talk) 17:09, 26 August 2022 (UTC)
Looking at Cite document didn't appear correct either. Cite report looks like it will do what is needed. - Favre1fan93 (talk) 18:13, 26 August 2022 (UTC)
I used {{Cite report}}. That looked correct and did what I think was intended. Thanks everyone. - Favre1fan93 (talk) 18:28, 26 August 2022 (UTC)

Access indicators icons (registration vs limited)Edit

I did not search for prior discussion on this. The icons for |url-access=registration and |url-access=limited seem identical (Chrome, Edge, Firefox, Safari latest versions Mac/Windows, & Tor (Ubuntu). Can there be some differentiation? I believe they were rendered differently in the past.

In Help:Citation Style 1 § Access indicators for url-holding parameters (talk) 14:38, 30 August 2022 (UTC)

The registration and limited icons have always been the same ( ) except that the tooltips are different: 'Free registration required' and 'Free access subject to limited trial, subscription normally required'.
Trappist the monk (talk) 14:56, 30 August 2022 (UTC)
  • Recommend representing "registration" with solid shackle for visual clarity. Any opinions? (talk) 15:20, 30 August 2022 (UTC)

{{cite book}} errorEdit

I get the following when I try to include a 2nd editor for a book:

Unknown parameter |editor_first2= ignored (help); Unknown parameter |editor_last2= ignored (help); Unknown parameter |editor_link2= ignored (help)

Help? (I'm working on Piping_guan... I'll leave the error there...) - UtherSRG (talk) 14:51, 30 August 2022 (UTC)

Use a hyphen rather than an underscore in the name of the parameter. Jc3s5h (talk) 14:54, 30 August 2022 (UTC)
You want to use hyphens, not underscores, in the appropriate parameter names. Switching them out resolved the error. Imzadi 1979  14:55, 30 August 2022 (UTC)

Nevermind... I fixed it with cite bot.... You may all laugh at me now. :) - UtherSRG (talk) 14:56, 30 August 2022 (UTC)

LOL! You caught it before cite bot could write it out. Cheers, speedy! :) - UtherSRG (talk) 14:57, 30 August 2022 (UTC)

DOI citation generatorsEdit

I'm having difficulty with the tools to generate a Wikipedia citation from a DOI. [1] appears permanently broken. [2] works but only occasionally. Most times I use it, I get, "The web request failed. You might not have Internet connectivity right now or CrossRef might be having issues. Please try again later." (My Internet connectivity is fine.) These are/were very useful tools! Is there a new tool or something? Thanks. Bondegezou (talk) 09:01, 7 September 2022 (UTC)

Have you tried User:Citation bot? It will expand most citations that are in the form {{cite journal|doi=10.x.y.z}}. Other tools are listed on that page. – Jonesey95 (talk) 14:30, 7 September 2022 (UTC)

Question on citationsEdit

I reached out to BrownHairedGirl about this earlier, but I want to ask here, too, to see if there is consensus about possibly changing wording. Specifically, I was curious about the url-access-level parameter. If a news site only starts requiring registration after a certain amount of articles have already been read, should it be tagged as "requiring registration", or should that be reserved for sites that require registration for all articles?

  • The registration tag states that it should be used "even if a limited preview, abstract or review may still be available without registration", though it's unclear whether a certain amount of articles read counts as well.
  • The limited tag, which BrownHairedGirl pointed towards, states that it could be used for certain constraints, including "cap on daily views". Would a cap on available articles read also count here?

I'm sure this probably doesn't appear super important, but it's quite common for news cites to provide a certain amount of free articles before registration. Krisgabwoosh (talk) 21:41, 9 September 2022 (UTC)

I understand from the above that the registration you are discussing requires only establishing a sign-in identity, without limits in viewership.
There is no hard rule about this. However, availability trumps convenience. A source that can be accessed by simply clicking for a maximum of x times is less available than a free-registration source. I would use "registration" for the access condition.
A source with registration that requires a one-time payment following x number of views is better represented with "limited" access. Implicit in such access is that after a non-determinate number of views a paywall applies. (talk) 00:35, 10 September 2022 (UTC)
I tend to agree with 74 that |url-access=registration/subscription is appropriate for limited-view websites. Izno (talk) 21:28, 11 September 2022 (UTC)

article numbersEdit

There was a discussion at User talk:Citation bot/Archive 33 § Pages vs Issue vs at vs ... about what to do with journal articles that are numbered. Because we don't have an article number-specific parameter, I suggested using |number= and was immediately shot down by those who apparently believe that article numbers belong in the in-source parameters |pages=, |pages=, or |at=. I believe that that is incorrect because doing so corrupts the metadata if the cite requires pagination. Sure, my |number= suggestion suffers from the same fault.

So, since COinS supports article numbers with the &rft.artnum= k/v pair which has heretofore been unused by cs1|2, I have hacked the module sandboxen to support new |article-number= in {{cite journal}} only:

{{cite journal/new |title=Title |journal=Journal |volume=XIV |article-number=56 |page=15}}
"Title". Journal. XIV 56: 15.
'"`UNIQ--templatestyles-0000002C-QINU`"'<cite class="citation journal cs1">"Title". ''Journal''. <b>XIV</b> 56: 15.</cite><span title="ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.jtitle=Journal&rft.atitle=Title&rft.volume=XIV&rft.artnum=56&rft.pages=15&" class="Z3988"></span>

If we are to keep this we should decide some things:

  • how are article numbers to be annotated or punctuated?
  • are article numbers supported when both |volume= and |issue= are set? yes
  • do article numbers require |volume=?

Beyond those things, if we are to keep this, support for journal cites using {{citation}} should be supported ({{cite journal}} with |mode=cs2 is of course already supported)

Keep? Discard?

Trappist the monk (talk) 21:51, 10 September 2022 (UTC) 14:37, 11 September 2022 (UTC)

This is useful only regarding a few online journals that don't bother with journal pagination at all and those that paginate each numbered article at p. 1. Others have both article numbers and journal-level pagination, and citing both is superfluous. |article-number= should be a global alias of |pages=. The article/column title which is the actual in-source location is the primary discovery parameter (after journal-name, author etc). Parameters such as pages, article-numbers etc. are not in-source locations. From the reader's standpoint are ancillary helpers. (talk) 00:15, 11 September 2022 (UTC)
If we are to believe our own documentation, for the purposes of cs1|2, |page=, |pages=, and |at= are in-source locators; see Template:Cite journal § In-source locations as an example, others are similar.
I have never seen article numbers used with anything but journals so it seems to me that |article-number= cannot be a global alias of |pages=. Further, as an alias of |pages=, |article-number= would necessarily render the same way that |pages= renders; the article number would fill &rft.pages= instead of &rft.artnum= in the metadata; when an editor wishes to identify two or more specific pages in the source that support text, one of |pages= or |article-number= must yield, or the editor must shoehorn both assigned values into one parameter or the other.
Your discard !vote is noted.
Trappist the monk (talk) 14:14, 11 September 2022 (UTC)

Sandbox version is currently broken to the point of unusability. An article number is an individual identifier for an article. It is not even close to the same thing as an issue number. An article can easily be part of a journal that has volumes and issues; in fact I would expect that to be the case for most journals that use this referencing format. But your hacked-up version doesn't work for that case: if there is an issue number, it doesn't show the article number, regardless of whether pages are also included.

  • {{cite journal/new |title=Title |journal=Journal |volume=3 | issue = 1 |article-number=4159}}
  • "Title". Journal. 3 (1) 4159.
  • {{cite journal/new |title=Title |journal=Journal |volume=3 | issue = 1 |article-number=4159|pages=1–24}}
  • "Title". Journal. 3 (1) 4159: 1–24.

If this is to be implemented at all, it must work separately from issue. Your question of whether to tie it to issue makes zero sense, and indicates a fundamental misunderstanding of what article numbers even are. They are a replacement for page numbers, not for issue numbers. —David Eppstein (talk) 05:49, 11 September 2022 (UTC)

Why are you so angry? What is it that gives you the impression that what I did was anything more than skeletal groundwork to test how |article-number= might be implemented and the result rendered?
The article identified in User talk:Citation bot/Archive 33 § Pages vs Issue vs at vs ... is here. That journal apparently does not use issue numbers when it uses article numbers. Because that was my model, I hacked the module suite to reflect that. I then posted the above where I wondered whether |issue= should be supported when |article-number= is used.
Why are you so angry about this incremental process?
Trappist the monk (talk) 14:14, 11 September 2022 (UTC)
|article-number= now renders with or without |issue=.
Trappist the monk (talk) 14:37, 11 September 2022 (UTC)
How about focusing on content and not on imagined attitudes of other people?
Next question: Some publications (notably the LIPIcs series of conference proceedings) mix article numbers and page numbers, so that for instance if article number 33 of LIPIcs volume 189 happens to have 17 pages, then it is formatted (both in the official publisher metadata and in the page numbers on the actual pages of the publication) as having a page range of 33:1–33:17, with the article number entirely encoded within the pages and not as a separate parameter. Other sources list both an article number and a number of pages, without assigning compound numbers to the individual pages; the same paper is listed in MathSciNet as "Art. No. 33, 17 pp.", and if it were indexed in zbMATH it would be "Article 33, 17p." The first of these two formats could be handled by putting the compound numbers into a pages parameter, but that is problematic if you want the article number reflected in the CoInS MeTaDaTa or however it's capitalized (does anyone still use cOiNs?). The second is not allowed, because even if you tried to use |article-number=33 |pages=17pp, some bot would complain that you're using the pages parameter wrong and try to "fix" it to something else.
So do you have a suggestion for how to deal with compound pages and with article-number + number-of-pages referencing? You can argue (as we have seen here before) that numbers of pages should not be included in references because they're not used to identify the reference, but clearly they are used to describe it, and that description could be useful to some readers of the reference for instance in trying to figure out how much effort they would have to put in to read the reference. Also, whether they should be included is clearly a matter of opinion rather than a matter of fact, because some well-established sources (MathSciNet) do include it. —David Eppstein (talk) 20:57, 11 September 2022 (UTC)
Perhaps I will when the first words out of your mouth are not shouted. I think my initial post was clear enough for you to know that the code is a prototype yet you chose to shout about its unusability.
The |article-number= proposal is currently constrained to {{cite journal}}. Your paper is, I think, best cited with {{cite conference}} using the pagination as printed in the proceedings. The publisher's BibTeX citation clearly shows that the publisher thinks that the paper is one of a number of papers collected in a proceedings (note that there is no article number there). |article-number= is supported only by COinS journal objects so {{cite conference}} can never fully support |article-number= were we, at some point, to allow that.
As far as I know, cs1|2 has never supported the notion of total pages. There have been several discussions here about that. We could fully support a |tpages= parameter in {{cite book}}, {{cite thesis}}, and other non-periodical templates because COinS has the rft.tpages k/v pair for book and dissertation objects. But, that is a discussion for another time and another place. This discussion is about |article-number= in {{cite journal}}.
Trappist the monk (talk) 01:05, 12 September 2022 (UTC)
You think journal articles can have article numbers but that conference proceedings papers cannot? Your dogmatic approach to what a citation can be or cannot be is something that causes me to occasionally get frustrated here. People who publish stuff, in various ways, are often flexible in how they define those publications in a way that allows others to find them. Citation formats here must stay as flexible, in order to allow editors to use them to correctly refer to things, and to avoid forcing editors to misuse their parameters because the parameters as used correctly are inadequate to the task. This latest bizarre piece of rigidity, that conference papers are somehow so different from journal papers that they cannot even be thought of as an example of something with an article number even in situations where they obviously do have article numbers, and must be prevented from using similar formatting and parameterization, is a case in point. (Also, fyi, I type with my fingertips, and usually not at the same time as speaking in any way; my mouth is not involved in the process. But your aversion to boldface text for emphasis is noted for future reference. Is italic ok or should I just never emphasize any part of what I write when communicating with you?) —David Eppstein (talk) 01:17, 12 September 2022 (UTC)
This sentence is the only sentence I have written about article number support in conference citations:
|article-number= is supported only by COinS journal objects so {{cite conference}} can never fully support |article-number= were we, at some point, to allow that.
We can choose to support |article-number= in {{cite conference}} as a non-metadata parameter as we do for other parameters (|access-date=, |agency=, |department=, |editor=, |language=, |medium=, ...); parameters that display something but that don't contribute to the citation's metadata. Before we do that, I think that we should figure out how or whether we support |article-number= in {{cite journal}}.
Trappist the monk (talk) 16:46, 12 September 2022 (UTC)
Incidentally, for a random example of a paper in a conference that its publisher clearly identifies using article numbers, see doi:10.1145/3225058.3225061. In the previous example, the publisher used compound page numbers of the form NN:1–NN:PP, but the databases MathSciNet and zbMATH separated it out as Article number NN, PP pages. In this example, it is reversed: the publisher uses Article number NN, pages 1–PP while the database DBLP [3] uses compound page numbers of the form NN:1–NN:PP. I think we can conclude from these examples that multiple publishers and databases think that these two forms are semantically equivalent and that the choice between one or the other is a matter of style. —David Eppstein (talk) 01:39, 12 September 2022 (UTC)
On second thought, I don't think this deserves all the back-and-forth, it is such a minor issue. The article number, a sparingly used piece of information, is not necessary in order to find an article. People look for articles by title, journal name & date, and by author. Anything else (page range, article number) is ancillary. Article number maybe helpful but it doesn't have enough traction for the effort (technically, both article page range and article number are article representations, just like the article title is). It can be easily input, as noted previously, in |at= if there is overwhelming need. The fact that the metadata scheme chokes (again) on this does not concern this page. We discuss the structure & presentation of citation data, and metadata should follow what is decided here. This is a one-way relationship. By all means fix the metadata, and there's probably no need to waste time telling us about it here. I am certain any participant in this talk page who wants to follow that application knows how to find the relevant project pages. (talk) 12:57, 12 September 2022 (UTC)
Added support for |article-num= in {{citation}} when |journal= has a value:
{{citation/new |title=Title |journal=Journal |volume=XIV |article-number=56 |page=15}}
"Title", Journal, XIV 56: 15
'"`UNIQ--templatestyles-00000038-QINU`"'<cite class="citation cs2">"Title", ''Journal'', <b>XIV</b> 56: 15</cite><span title="ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.jtitle=Journal&rft.atitle=Title&rft.volume=XIV&rft.artnum=56&rft.pages=15&" class="Z3988"></span>
Trappist the monk (talk) 16:46, 22 September 2022 (UTC)

Allowing the url-access parameter for citations with no url but chapter-urlEdit

Inspiring myself from user Prototyperspective's intervention above. Today when trying to update the source list of the article Charles Hamilton, 5th Earl of Abercorn with:

{{Cite book|last=Cressy |first=David |author-link=David Cressy |editor-first=DeLloyd J. |editor-last=Guth |editor2-first=John W. |editor2-last=McKenna |date=1982 |title=Tudor Rule and Revolution: Essays for G R Elton from His American Friends |chapter=Binding the nation: the Bonds of Association, 1584 and 1596 |publisher=Cambridge University Press |location=Cambridge |pages=217–234 |isbn=978-0-521-09127-5 |chapter-url= |url-access=registration}}

I got the error:

{{cite book}}: |url-access= requires |url= (help).

I could of course add a |url= setting it to the URL (without the page) given in |chapter-url=, but would it not make more sense to accept |chapter-url= as equivalent to |url=? With thanks and best regards, Johannes Schade (talk) 17:49, 15 September 2022 (UTC)

Cressy, David (1982). "Binding the nation: the Bonds of Association, 1584 and 1596". In Guth, DeLloyd J.; McKenna, John W. (eds.). Tudor Rule and Revolution: Essays for G R Elton from His American Friends. Cambridge: Cambridge University Press. pp. 217–234. ISBN 978-0-521-09127-5.
Trappist the monk (talk) 17:52, 15 September 2022 (UTC)

Edit requestEdit

the ISBN link is a redirect – it should be changed from ISBN (identifier) to its target, ISBN. theleekycauldron (talkcontribs) (she/her) 05:09, 16 September 2022 (UTC)

@Theleekycauldron: Exactly which link are you asking to be changed? A similar request was made at Help_talk:Citation_Style_1/Archive_84#Protected_edit_request_on_1_August_2022, so you might like to read the discussion there. -- John of Reading (talk) 06:49, 16 September 2022 (UTC)
ah, I see. thanks :) theleekycauldron (talkcontribs) (she/her) 07:06, 16 September 2022 (UTC)

URGENT | Getting Lua error in ମଡ୍ୟୁଲ:Citation/CS1 at line 3862: attempt to get length of field 'message_tail' (a nil value). under "References" in Odia WikipediaEdit

Redirected here from: ⚓ T318008 Getting Lua error in ମଡ୍ୟୁଲ:Citation/CS1 at line 3862: attempt to get length of field 'message_tail' (a nil value). under "References" in Odia Wikipedia (

Steps to replicate the issue (include links if applicable):

  • Check any page with reference/citation for getting the Lua error. It's happening in almost across all pages.
  • For example visit: An Odia Wikipedia page

What happens?:

  • We see errors in references
Lua error in ମଡ୍ୟୁଲ:Citation/CS1 at line 3862: attempt to get length of field 'message_tail' (a nil value).

What should have happened instead?:

  • There should not be any errors.

Please help. Soumendrak (talk) 18:57, 16 September 2022 (UTC)

message_tail is no longer used. That table used to be defined in or:ମଡ୍ୟୁଲ:Citation/CS1/Utilities – if you look there you will see that there is no definition for message_tail. That suggests that one or more modules in the suite are out-of-sync; my guess is that the main module or:ମଡ୍ୟୁଲ:Citation/CS1 needs to be re-imported from Module:Citation/CS1.
Trappist the monk (talk) 19:16, 16 September 2022 (UTC)
I just looked at the histories of several of the modules in the suite. Most appear to August 2021 so if you are intending to update to the current version, all of the module suite must be imported (best to import to the sandbox versions, test that the sandbox works and then update the live version). or:ମଡ୍ୟୁଲ:Citation/CS1/Utilities looks like it was updated 2022-09-16 by you. Because I could, I reverted the most recent edits to ~/Utilities and previewed or:ଜ୍ୟୋତିବା ମନ୍ଦିର. No glaring red error messages. I did not save; I'll leave that to you.
Trappist the monk (talk) 19:34, 16 September 2022 (UTC)
Thanks for your inputs, I think it's better to sync all the modules to the latest version.
I am afraid, I do not know how to import to sandbox version Is there any documentation out there you can guide me?
Thank you. Soumendrak (talk) 03:02, 17 September 2022 (UTC)
After reimporting the main module or:ମଡ୍ୟୁଲ:Citation/CS1 from Module:Citation/CS1 getting a new error
Lua error in ମଡ୍ୟୁଲ:Citation/CS1 at line 4188: attempt to index field 'url_skip' (a nil value).
- Soumendrak (talk) 03:19, 17 September 2022 (UTC)
imported ମଡ୍ୟୁଲ:Citation/CS1/Configuration from en:Module:Citation/CS1/Configuration and it seems to be fixed the issue.
Thanks a lot for quick inputs.
- Soumendrak (talk) 03:36, 17 September 2022 (UTC)
@Soumendrak: Update all of the modules in the suite; you cannot expect that cs1|2 at will work as it should if you do piecemeal updates like you are doing.
Trappist the monk (talk) 13:19, 17 September 2022 (UTC)
Thanks for the input. I have imported all the relevant CS1 modules mentioned on Module:Citation/CS1 page.
- Soumendrak (talk) 17:42, 18 September 2022 (UTC)
@Soumendrak: Ok, but you should know that sandbox modules are for development. We do not recommend their use for anything. What I wanted to get across to you is that you should import our live modules to your sandbox modules. Make the necessary changes for localization to your sandbox modules. When your sandbox modules are working as they should, then, and only then, update your live modules from your sandbox modules. Do not expect our sandbox modules to be working, stable, or reliable, ever.
Trappist the monk (talk) 18:00, 18 September 2022 (UTC)
ok, got it.
- Import from live modules to sandbox modules
- Test there, check for errors
- Then move from sandbox to live.
Thank you. Soumendrak (talk) 18:14, 18 September 2022 (UTC)

Update SSRN base urlEdit

  • Aolain, Ni; D, Fionnuala (2010-05-19). "Women, Vulnerability and Humanitarian Emergencies". SSRN 1611818.

points to

which doesn't work anymore. It should instead point to

Headbomb {t · c · p · b} 21:37, 18 September 2022 (UTC)

Displaying "n.d." is obscure for readers, "undated" is betterEdit

For citations without a date, the abbreviation "n.d." appearing somewhere in the citation is a bit baffling for readers. I would suggest that the displayed citation should show "undated"; the template should allow "n.d" to be entered as at present for compatibility, or "undated". "No date" is other possible clear wording. Best wishes, Pol098 (talk) 12:59, 19 September 2022 (UTC)

Citations are baffling for readers. The reason being that Wikipedia citations were originally, and still, based on citation systems targeting the small minority of readers familiar with the concept and its applications. Because it was the easier thing to do, and was safely within the comfort zone of their field of expertise. For readers, the present system looks quite similar and equally (if not more) baffling. In such a system, "n.d." fits perfectly, and is consistent. It also happens to be common, established cataloguing markup for undated materials. Since the present system just like its ancestors, disregards the needs of the general readership, piecemeal reform will only add confusion. The OP's request should not be applied. (talk) 16:20, 19 September 2022 (UTC)
Most general and book citations using {{citation}} and variants are fairly clear and obvious, as a random look through articles will confirm - they are not terribly baffling. "Archived from the original" becomes obvious the first time it is clicked. While journal citations are more obscure, they're likely to be understood by those who will use them, and the volume, issue, DOI, etc. do not tell the reader anything useful anyway. But the date, in particular, can be important and relevant, and n.d., if used, is the only really obscure part of a citation. A couple of random examples (with date) that are typical, and shouldn't cause anyone any difficulty:

"Maori leaders at odds over flash mob haka". 3 News NZ. 20 September 2011. Archived from the original on 27 December 2011.

Hartigan, Ryan (2011). "Embarrassing Time, Performing Disunity: Rugby, the haka, and Aotearoa – New Zealand in the United Kingdom". Performance Research. Taylor & Francis. 16 (2): 37–43. doi:10.1080/13528165.2011.578728. S2CID 194059694.

So I maintain my suggestion (for discussion, not a request). Best wishes, Pol098 (talk) 20:29, 19 September 2022 (UTC)
I agree. "Undated" or "No date" would be an improvement. Improvement is good, unless the only thing that's good enough is a revolution where we describe each source of information in full sentences. How's this sound?

I got this information at 3 News NZ at the web address with the title "Maori leaders at odds over flash mob haka" and the date 20 September 2011. That address doesn't work any more, but now you can get it at on the Internet Archive where it was copied on 27 December 2011.

Yes I intend that proposal as gentle sarcasm. SchreiberBike | ⌨  21:03, 19 September 2022 (UTC)
The intention is irrelevant. In fact, the verbose explanation above is far more understandable by practically anyone who can read unadorned English, compared to the current shorthand that citations use. Assuming that any random user of the millions of unique users that visit Wikipedia daily ignores the annoyance of the strange in-text (fiootnote) notation and is curious enough to follow it, they will eventually land on a paragraph with additional strange formatting, resembling the writings of someone who cannot form a sentence. Complete with strange terms and numbers, and various forms of names and dates seemingly thrown in haphazardly. Also assuming they have an idea of the basic Wikipedia policy about verification, they will probably wonder why Wikipedia makes it so hard for any reader to do so. With the confusion of footnoting and referencing systems, with all their peculiar characteristics, special cases and exceptions. And they cannot turn to the virtually non-existent documentation for help.
The current system is for those in the know. That category knows exactly what "n.d." means. More than that, they probably expect it, as the professional norm. By adding beginners' formatting to a more or less expert system you do a disservice to both. (talk) 22:59, 19 September 2022 (UTC)
We have used n.d. for an undated publication since its institution along with date checking in CS1 in 2013. A discussion not long after indicated that this is what Chicago, APA, and MLA do. Indeed as 71 implies this is the professional standard for citations (with what was then some consensus on the point; standards have advanced a decade in that time and those citation styles do change, so I haven't looked to see either for new advice from those orgs or contradictory from others).
I don't see a strong reason to differ from those in this regard. The point is that in almost all cases that these will be set off by round brackets, and if a reader hasn't figured out that round brackets indicate a date, I don't know how to help them. That's even if they don't already have their own experience with citations, which I personally gained in high school English.
A few other discussions which I have not perused. Izno (talk) 23:27, 19 September 2022 (UTC)
The strong reason is WP:TECHNICAL. Those style guides are for academic audiences in a context where print space is at a premium. Our style guide is for a general audience under WP:NOTPAPER. "Undated" is much easier to understand than "n.d." —David Eppstein (talk) 00:05, 20 September 2022 (UTC)
@David Eppstein: puts it very well. Our readers (and indeed many of our writers) are not well-versed in the jargon of APA or MLA. "undated" is clear and unambiguous. DuncanHill (talk) 00:08, 20 September 2022 (UTC)
Within the contours of the current system, the n.d. shorthand, which is editor markup predating any citation system, fits nicely among all the other elements. If one is able to read & understand CS1\2 citation shorthand, "n.d." is no mystery.
CS1\2 is not accommodating the general reader on any level (presentation, documentation, ease-of-understanding, verification applicability etc). Half-hearted attempts to help the (presumably inexpert) reader with minor or cosmetic elements are imo placebos that distract and evade.
After all this time, I no longer believe that CS1\2 can ever be a general-purpose citation system understandable by inexpert readers, i.e. fitting an encyclopedia of similar declared goals. It will plod along as a semi-expert system with a confused design philosophy. Within its narrow scope, it can be fractionally, serially improved. Ok, cool. But "Undated" vs or in addition to "n.d." is not an improvement, it's another sideways move. (talk) 01:47, 20 September 2022 (UTC)
This question of "how terse is too terse" comes up every once in a while. I think this is on the line of pp. that outputs with pages in most citations and (1) or 3 that outputs with issue and volume in {{cite journal}}, and even the fact cite journal does not output any indicator of the page number (ok, : 123). If we want to go down the road of arguing that users can't figure out n.d. given the other context clues, then we should look at every other excessive abbreviation, and there are far worse offenders than this one, and I know the discussion here has not been favorable to redoing all those as well. (Especially given the fact that other citation styles do support n.d..) Izno (talk) 02:29, 20 September 2022 (UTC)
Agreed. We've started with other citation guides as a basis, and given that there are some pretty standard abbreviation conventions out there, I don't think we should be ditching many of them, including n.d. It's all find and dandy to spell everything out, but readers are still going to encounter these conventions when they read the source material for their editing here.
That said, I wholeheartedly agree that the terse journal notation style is not as helpful when we also use standard abbreviations "Vol.", "no." and "pp." in other templates. My impression is that it was added as a way to embrace subject-matter experts who are accustomed to such notation, even as we have, or had, a guideline saying that journal names need to be spelled out in full for our generalist audience. I would rather focus on that most terse notation so that our citations have a similar level of abbreviation first before then deciding if other fairly common abbreviations should be changed to full words. Imzadi 1979  04:49, 20 September 2022 (UTC)
Opening it up to "undated", "no date", etc., I feel makes harv/sfn-style templates complicated; is there going to be support for (Last n.d.) as well as (Last & Undated), (Last & No date) (already see how these two differ), etc.?
How much evidence is there that readers are even confused by seeing n.d. anyway? Umimmak (talk) 01:30, 20 September 2022 (UTC)
That's hard to tell since relatively few citations on here are specified as having no date. The sample size is too small. Glades12 (talk) 09:40, 20 September 2022 (UTC)
I think anyone familiar with the idea of citations can usually figure out the meaning from looking at them; the formatting Wikipedia has chosen helps. I didn't know what "n.d." meant the first time I saw it but searching for n.d. leads quickly to a page where the last item explains it (probably in violation of the rules for disambiguation pages). Citations are absolutely necessary so we have to have them in some form. I haven't seen a reasonable alternative and writing them out in sentence form seems silly. If we were to eliminate the less obvious things like bold to indicate volume number and ":" for page number and "n.d. meaning no date, that would help. We should help where we can. SchreiberBike | ⌨  13:32, 20 September 2022 (UTC)
I repeat the typical examples, chosen randomly from the haka page, I gave before, which may help to form an opinion. They are totally clear to a new reader, I think, except that the academic journal details - which are absolutely irrelevant when reading an article or linking to a reference - such as volume, issue, DOI, etc. are obscure.

"Maori leaders at odds over flash mob haka". 3 News NZ. 20 September 2011. Archived from the original on 27 December 2011.

Hartigan, Ryan (2011). "Embarrassing Time, Performing Disunity: Rugby, the haka, and Aotearoa – New Zealand in the United Kingdom". Performance Research. 16 (2): 37–43. doi:10.1080/13528165.2011.578728. S2CID 194059694.

Compare with:
"Maori leaders at odds over flash mob haka". 3 News NZ. n.d.

Smith, John (n.d.). "Maori leaders at odds over flash mob haka". 3 News NZ.

Remember also that, depending upon Wikipedia setup, details of references are popped up when hovering over the number in the article text.

I'm concerned only with reader experience, not complying with standards, or editor convenience. Best wishes, Pol098 (talk) 15:06, 20 September 2022 (UTC)
At best, a little coding to auto-format it as n.d. in the output wouldn't be inappropriate if it's truly felt that the abbreviation is that obscure, much as {{circa}} outputs c. for readers. I don't see a need to remove the abbreviation though given that it's common in other citation guides. Imzadi 1979  15:20, 20 September 2022 (UTC)
I probably misunderstand, but nobody is suggesting that citations should be done away with. Only that they should be made understandable, so that 1. their relationship to the text (in body or footnote) is unquestionably obvious 2. their path to the verification process a logical, easy conclusion. The OP proposal is schizophrenic, which is no fault of the OP, but reflects the confused design of the Wikipedia citation systems. Simply put, they don't know if they are geared to experts or to the general public. [A]nyone familiar with the idea of citations likely doesn't need to be told what "n.d." means. The majority of readers though are very likely unfamiliar with either the concept or the particulars of citations. For them, adding/changing to "undated" is a meaningless detail in a sea of strange terms parading in an unfathomable sequence. (talk) 16:21, 20 September 2022 (UTC)
I'm not sure whom I'm responding to since a series of IP editors have contributed here, but do you (or anyone else) have a suggestion of a better way to present citations? SchreiberBike | ⌨  22:35, 21 September 2022 (UTC)
I suppose you are responding to the argument, irrespective of the nominal participant. That covers it. Indeed there are suggestions, but this discussion is rather narrow for them to be added here. The main one is systemwide and would involve a "general" and an "expert" mode of presentation. So each would be internally consistent. Another topic. (talk) 12:45, 22 September 2022 (UTC)


Why doesn't

  • {{cite book|chapter=Óró Sé do Bheatha 'Bhaile|title=Sean-Nós Nua|first=Sinead|last=O'Connor|year=2002|language=ga}}
  • O'Connor, Sinead (2002). "Óró Sé do Bheatha 'Bhaile". Sean-Nós Nua (in Ga).

work? I see "in Ga" instead of a more readable description of the Irish language. (I know this is not actually a book; it's just an example.) —David Eppstein (talk) 21:05, 20 September 2022 (UTC)

The Ga language exists. Template:Citation Style documentation/language/doc#Language names lists "ga" and says " When these names are used in |language=, cs1|2 will attempt to validate them but such attempts are not likely to succeed." Imzadi 1979  21:23, 20 September 2022 (UTC)
Help:Citation Style 1 says to use two-letter codes. Maybe that advice should be modified in cases where the two-letter code does not work? Alternatively maybe the codes can be considered to be case-sensitive? —David Eppstein (talk) 21:30, 20 September 2022 (UTC)
Where, exactly, does Help:Citation Style 1 [say] to use two-letter codes? cs1|2 supports two- and three-character language tags as well as most of the IETF and IETF-like language tags supported by MediaWiki.
Trappist the monk (talk) 22:30, 20 September 2022 (UTC)
"Because cs1|2 templates are often copied from to other wikis, the use of language codes is preferred so that language names render in the correct language and form: espagnol at a French-language wiki instead of the English word "Spanish"." —David Eppstein (talk) 00:08, 21 September 2022 (UTC)
That sentence only says that use of a language tag is preferred over the use of a language name; the number of characters in the language tag is not mentioned.
Trappist the monk (talk) 00:33, 21 September 2022 (UTC)
The list of language tags does not provide a language tag for Irish of any length other than the two-letter one. —David Eppstein (talk) 04:25, 21 September 2022 (UTC)
Because MediaWiki does not support the ISO 639-2, -3 tag gle:
{{#language:gle|en}} → gle
Trappist the monk (talk) 14:30, 21 September 2022 (UTC)
When evaluating the value assigned to |language=, cs1|2 looks first at the list of MediaWiki-supported language name. The evaluation is case-insensitive because editors will write |language=french. So, for |language=ga, cs1|2 finds Ga because that is a language name that MediaWiki supports. Of all of the MediaWiki-supported languages, there are only seven where the language tag is the same as language name; ga (Irish) and Ga (gaa) are the only two where the tag is spelled the same as an unrelated language name. The others are:
{{#language:fon|en}} → Fon
{{#language:isu|en}} → Isu
{{#language:luo|en}} → Luo
{{#language:tiv|en}} → Tiv
{{#language:vai|en}} → Vai
{{#language:yao|en}} → Yao
These, of course, are not a problem.
Trappist the monk (talk) 22:30, 20 September 2022 (UTC)

place, when publication-place is redundant with workEdit

Consider this citation:

Adam, Karla (15 September 2022). "The British love queues. The queen's death brought one for the ages". The Washington Post. Retrieved 15 September 2022.

The cited article bears a dateline of "London". Per the documentation of {{Cite news}}, it would be appropriate to set |place=London (or |location=London). However, a dilemma is then reached (ignore CS1 maint tags; they're from the styling added for emphasis):

Is there some way to indicate, like, |publication-place=redundant or something? It seems strange to have a system where I can indicate dateline for, say, The Wall Street Journal or The Mercury News, but not for The Washington Post, the Los Angeles Times, The New York Times, etc. -- Tamzin[cetacean needed] (she|they|xe) 18:54, 23 September 2022 (UTC)

I would omit it. It does not help the reader to verify the claim in the article in any way to know that the writer of the article was in London. – Jonesey95 (talk) 19:36, 23 September 2022 (UTC)
Location is where the publication was published, not where a particular item was written - such details don't help to locate the reference and merely confuse the reader.Nigel Ish (talk) 19:38, 23 September 2022 (UTC)
The current documentation in Template:Citation Style documentation/publisher says place: For news stories with a dateline, the location where the story was written. Should that be removed? Or some note added that would limit it to some subset of cases where it's particularly relevant? -- Tamzin[cetacean needed] (she|they|xe) 19:44, 23 September 2022 (UTC)
We've discussed this before:
A quick scan of those discussions suggests that we should, at the least, make all of |publication-place=, |place=, and |location= into exact-equivalent aliases. Doing that gets rid of the 'written at' dateline stuff because as noted above, where a source was written does nothing to help a reader locate the source. Am I mistaken in my reading of those discussions?
Currently, Category:CS1 location test has 866 articles. I can dust off my awb script and let it remove articles where |publication-place= has the same value as |location= or |place=.
Trappist the monk (talk) 19:52, 23 September 2022 (UTC)
That's how I'd read the consensus across those three discussions, yeah. And I tend to agree. I suppose I could see some scenario, less in the context of news articles and more in the context of letters or poems, where location of writing could have some disambiguatory function? But there will almost always be other ways to disambiguate that. And if we were to have a parameter for that, better a specific |written-at=, with no ambiguity in naming. -- Tamzin[cetacean needed] (she|they|xe) 20:04, 23 September 2022 (UTC)

Fails to throw a DOI errorEdit

Holm, Cyril; Lind, Hans; Vogel, Jonas Anund (December 2019). "Incentivising innovation in the construction sector: the role of consulting contracts". Construction Economics and Building. 19 (2): 181–196. doi: Retrieved 19 September 2022.

Should throw an error. is not a valid prefix. Headbomb {t · c · p · b} 20:59, 23 September 2022 (UTC)

DOI registrants can be any sequence of digits and dots (see cs1|2 looks for patterns of digits and dots that aren't yet in use. The live module already catches one and two digit registrants without subcode:
  • {{cite journal |journal=Journal |title=Title |doi=10.12/somat}}
    "Title". Journal. doi:10.12/somat. {{cite journal}}: Check |doi= value (help) – two-digit registrant without subcode
  • {{cite journal |journal=Journal |title=Title |doi=10.1/somat}}
    "Title". Journal. doi:10.1/somat. {{cite journal}}: Check |doi= value (help) – one-digit registrant without subcode
I have added a test for (presumably) unused one and two digit registrants with subcode:
  • {{cite journal/new |journal=Journal |title=Title |doi=10.12.1/somat}}
    "Title". Journal. doi:10.12.1/somat. {{cite journal}}: Check |doi= value (help) – two-digit registrant with subcode
  • {{cite journal/new |journal=Journal |title=Title |doi=10.1.1/somat}}
    "Title". Journal. doi:10.1.1/somat. {{cite journal}}: Check |doi= value (help) – one-digit registrant with cubcode
Trappist the monk (talk) 22:20, 23 September 2022 (UTC)
That they can be is not very important relative to the fact that they aren't. When we start having DOIs that don't start with 10.#### or 10.##### then we can remove that check. Headbomb {t · c · p · b} 22:39, 23 September 2022 (UTC)