Wikipedia talk:Citing sources/Archive 55

From Wikipedia, the free encyclopedia

Archive 50Archive 53Archive 54Archive 55Archive 56Archive 57Archive 60

Citing page numbers in PDF documents?

@Daniel Case in Valhalla train crash there's a large number of citations to a NTSB report as a PDF document (ref 13 in Special:Permalink/1202959297) which refer to page numbers (i.e. {{rp|58–61}}) using the PDF page numbers instead of the numbers printed on each page. What do folks think of this? I'm inclined to say it's not a good idea because somebody who has access to the report in printed form won't be able to map the citations to the paper copy. But I'd like to hear what other people feel about this. RoySmith (talk) 22:55, 4 February 2024 (UTC)

Maybe if it's a very large PDF with multiple documents with their own pagination it might be a useful, additional convenience to also include the page of the PDF, but I think in most cases it makes more sense to use the pages printed on the page when possible. Past discussion: Wikipedia talk:Citing sources/Archive 44#Which page number to use when citing PDFs? and Wikipedia talk:Citing sources/Archive 32#PDF Page Number. Umimmak (talk) 23:39, 4 February 2024 (UTC)
There was just recently a thread about this at WT:CS1. The gist is that it should cite the printed page numbers, since people may access the document in multiple formats, including paper, but links can go directly to the PDF page number, e.g.: |pages=[https://example.com/foo/bar.pdf#page=65 62–63].

PS: {{rp}} should no longer be used; it's a form of parenthetical referencing (injecting citation details into the article body instead of keeping them in the citation), and all of that was deprecated by the community in 2022. See User:SMcCandlish/How to use the sfnp family of templates for a crash course on one replacement system. It supports the parameter example I just gave, as do CS1/CS2 templates directly; in sfnp, harvp, and related templates, the shorthand |p= or |pp= can also be used, while if something more specific needed like naming a section, the CS1/CS2 |at= can be used to include a page number with an annotation, and sfnp, etc., have the equivalent |loc=.  — SMcCandlish ¢ 😼  01:07, 5 February 2024 (UTC)

Funny that you should say this. On another article I've been working on in draftspace recently one source is a 77-minute video documentary. It's obviously not enough just to cite the video; one has to cite the timestamp of the material that supports it, much as we would cite the page number or range of a 300-page book. And I found myself having to use {{rp}} ... I'd love to use {{sfn}} but there's just no way to make that work with {{cite AV media}}. Someone, in the process of creating that former template, failed to remember that people cite things other than documents, even though that should have been obvious at the time it was created. Perhaps that can be addressed soon in the Community Wishlist Survey.

Likewise, I'd love to see some equivalent version that works with {{bluebook journal}} or {{cite court}} to produce acceptable short versions of cites in Bluebook format (i.e. "AUTHOR, at PAGE" or "CASE SHORT TITLE, at PAGE" for the first time a single page is cited after the original full cite. Daniel Case (talk) 02:26, 5 February 2024 (UTC)

Cite AV media supports the |ref= field, so you can use {{sfnref}} to setup the link. Bluebook journal can be used with sfn in the normal way you would use it with {{cite journal}}, it will cause a false-positive error but that can be suppressed with {{sfn whitelist}}.
{{cite court}} doesn't support either method, but that's what {{wikicite}} is for. Wrap the 'cite court' inside the wikicite, and setup the |ref= field with sfnref. -- LCU ActivelyDisinterested «@» °∆t° 03:14, 5 February 2024 (UTC)
{{cite court}} now supports |ref=. You can use {{sfnref}} to create the anchor however you need, the same as CS1 templates. The short citation templates all have a |loc= parameter to cite things that are not pages. @Daniel Case: where were you looking for this in the documentation? It should likely be added there. And would |at= be more intuitive than |loc= for the sfn/sfnp/harv templates? Rjjiii (talk) 05:21, 5 February 2024 (UTC)
SMcCandlish recently raised the loc / at point at Module talk:Footnotes#loc, at, but that page doesn't have much traffic. It does seem more consistent, as |location= is for the general physical location of the publisher. I hate to think how many corrections would have to be made of it was changed though. -- LCU ActivelyDisinterested «@» °∆t° 15:23, 5 February 2024 (UTC)
I'd like it if loc= and at= were aliases, that would make {{sfn}} and {{cite}} more consistent to use with each other. Since loc= would remain valid, existing usages wouldn't need to be changed, as I understand it. Gawaon (talk) 16:00, 5 February 2024 (UTC)
I use loc= for all sorts of free-form usages. I've cited things I've found in old newspapers and put something like "near the bottom of column three on page 47" in the loc field, because otherwise there's no way anybody could find what I was talking about. Services like newspaper.com do provide URL-addressable clippings, but they have the same problem as PDF page numbers; they won't be of any use to somebody accessing the original source material via a different format (paper archive, microfilm, etc). But, yeah, a computer should be smart enough to see "loc=47", figure out that the digit string is a page number, and format it accordingly. RoySmith (talk) 16:25, 5 February 2024 (UTC)
Yes, I go by what I read in template documentation. It should be no surprise; we seem to be absolutely horrible at updating it. For instance, only thanks to SMcCandlish above did I learn that {{rp}} has been deprecated for over a year. There is still nothing on the documentation to advise editors of that, much less tell us what we should do instead. Daniel Case (talk) 18:49, 5 February 2024 (UTC)
It's not really deprecated though… what was decided in 2022 was to get rid of in-line MLA/Harvard style parenthetical referencing; {{rp}} barely came up in that discussion. Having multiple back to back instances like this[1]:1–4[2]:17 can get cluttered, but there isn’t a blanket proscription on this template (yet). Umimmak (talk) 18:57, 5 February 2024 (UTC)
I agree. SMcCandlish created that template and has been steadily encouraging people to develop and use better options, but it's not actually banned, and Wikipedia:Templates for discussion/Log/2021 December 17#Template:Rp (more than a year after the RFC about parenthetical citations) had a WP:SNOW-level keep response. WhatamIdoing (talk) 22:37, 7 February 2024 (UTC)
I would not use the PDF's own page numbers if there are printed page numbers visible in the doc. We have had questions about this for academic journal articles, and the advice has always been to use the printed/official ones (which could be something like pp. 124–131) rather than the PDF's (always and automatically starting a 1). WhatamIdoing (talk) 01:41, 5 February 2024 (UTC)
I understand where you're coming from, but more often than not the PDF's numbering is more visible than that on the pages in it. Sometimes, with more recent PDFs, it seems that it is possible for the equivalent page numbers to be used by the software, but a lot of older ones don't do that, and IME too many readers take the reader's numeration as canonical. Daniel Case (talk) 02:29, 5 February 2024 (UTC)
Agreed -- where there are page numbers visible use those; where they are not, use the PDF-associate numbers. Note that some publishers have eschewed the addition of page numbers even to their printed works (looking at you Chasma Press), which presents some special issues. I believe the retirement of {{rp}} in favor of {{sfnp}} is a mistake, but I won't go into that more here as it's an aside. User:Ceyockey (talk to me) 02:57, 5 February 2024 (UTC)
On Google Books, a lot of times when it gives you Google's e-book version there are no page numbers visible. You sort of have to guess from what the URL numbered the page you landed on from "Preview". Daniel Case (talk) 18:46, 5 February 2024 (UTC)
Per WP:EBOOK If there are no page numbers, whether in ebooks or print materials, then you can use other means of identifying the relevant section of a lengthy work, such as the chapter number or the section title. -- Ebooks do not have fixed page numbers; the content per "page" dynamically adjusts based on window/screen size, font size, etc. You should not arbitrarily assign one based on how Google Books happened to format a particular title since that page number is meaningless for anyone who accesses that (e)book in any format other than the Google Books Preview. Umimmak (talk) 19:04, 5 February 2024 (UTC)
Page numbers in older books can sometimes be interesting. I guess due to the requirements of the printing technology of the day, sometimes you would find a bunch of color plates bound into a book but outside of the normal page numbers. So, page 144, 145, no-number, no-number, no-number, no-number, 146, 147. I tend to cite those as "unnumbered page after page 145". I could see somebody naively looking at such a book and saying, "I don't know why they didn't number it, but it's after page 145 so it must be 146" and cite it as page 146, which would be wrong. RoySmith (talk) 19:21, 5 February 2024 (UTC)
I would firmly believe that the page numbers as printed should be those cited. Journals are regularly consecutively paginated; a PDF of a single chapter has the same issue. Moreover, some people might try to access it through other formats (eg Westlaw or Lexis' long web pages with [*123] to mark page divisions; see also Federal Register's version of the same with side notes). These are not compatible with this very naive approach to page number assignment. Ifly6 (talk) 19:40, 5 February 2024 (UTC)
  • Why not put BOTH in the citation? We had the flexibility to do that when we formatted citations by hand (just add a parenthetical to explain). I would think it would be easy to add an extra parameter to our templates… One for printed pagination, another for PDF pagination. While we would not use both pagination fields often, having two would be very helpful on the rare occasions when there are two distinct paginations. Blueboar (talk) 20:06, 5 February 2024 (UTC)
    Why not put BOTH in the citation? — The citation originally being asked about is "Highway-Railroad Grade Crossing Collision, Commerce Street, Valhalla, New York; February 3, 2015" (PDF). National Transportation Safety Board. July 25, 2017. pp. 18–19. Retrieved March 25, 2018. Hereafter cited as NTSB Report; page numbers will be those given by the PDF software, rather than those indicated in the document's pages. — to me, this is not one of those rare occasions which would benefit from including both. This article makes use of {{rp}} so has citations like [13]:37–38; to me it seems much more intuitive to ignore the line about page numbers will be those given by the PDF software, rather than those indicated in the document's pages. and just have [13]:26–27. There's no easy way to include both sets of page numbers with {{rp}} but even if these were converted to more standard short referencing I don't think any reader would benefit more from NTSB 2015, pp. 26–27 [37–38 PDF] over just NTSB 2015, pp. 26–27 in this particular case. Umimmak (talk) 20:32, 5 February 2024 (UTC)
    Absolutely. If a doc has page numbers, there's no good reason not to use them. If not, the question of "both" won't arise. Gawaon (talk) 20:57, 5 February 2024 (UTC)
    I would only use the printed page numbers for that NTSB report. I would expect others to do the same (e.g., if a news report needs to say something like "the photo on page 12 of the report"). WhatamIdoing (talk) 22:40, 7 February 2024 (UTC)
    What would we do, however, if what we wanted to cite was on a page that is unnumbered in the document? Daniel Case (talk) 06:52, 8 February 2024 (UTC)
    Is that the case here? On the odd occasion you would have to do that, you’d treat it like you would when you cite an unnumbered page of a print report/book: |at=Front cover, |at=Title page, |at=Copyright page, etc. I guess in a pinch you could say something like |at=Frontmatter, n.p., but just because one might conceivably cite a non-numbered page from a report doesn’t mean one should completely ignore the actual page numbers printed on each page. Umimmak (talk) 07:21, 8 February 2024 (UTC)

Why is all this not in the documentation?

It has been such a pleasure to get dragged into this discussion and learn about all these things, rather than in, oh, the appropriate template documentation, you know, that big wall o' text on green background that most users consult when they want to know how to cite things? Template:Cite web/doc#In-source locations, for instance, says nothing about using a PDF's internal page; nor does its book counterpart explain anything about what WP:EBOOKS tells us (In fact, I can't find where EBOOKS takes me other to the main WP:CITE page ... there is no boxed shortcut I can end up at. I realize that template documentation is mainly technical, but there is absolutely nowhere on them that I can remember ever being directed to one of these links to know what I've been doing wrong all these years of not participating in discussions here because I was too busy creating content and blocking vandals.

There's WP:PAGENUM, but it says nothing that would lead an editor to conclude that the printed page number is preferred over the one given by the reader.

Also, there is WP:PAGELINKS ... what number do we put in the URL? The printed one or the one the reading software uses (which may not always be the same)? Daniel Case (talk) 07:14, 8 February 2024 (UTC)

I suppose no one just ever expected an editor to not just use the page numbers printed on the page for the page numbers. This is the appropriate venue because one can cite pages with the CS1 family of citations, {{citation}}, or without any template at all — this is a question about citing sources in general, not about a specific template, although the documentation in various templates could be changed if this is deemed necessary. WP:PAGENUM reads If there are no page numbers, whether in ebooks or print materials, then you can use other means of identifying the relevant section of a lengthy work, such as the chapter number or the section title. — to me that suggests if there are page numbers then no additional instructions are needed because one would just use those. Umimmak (talk) 07:32, 8 February 2024 (UTC)
I decided to stop being snotty about this and ponder why we started doing it that way. I think it goes back to the late 2000s, when we only had the now-long-deprecated {{PDFlink}} (which had no page number support IIRC) for PDFs and we didn't have all the {{cite}} ones yet. At the time, also, I remember, a lot of PDFs were often created specifically for the web, usually from Word. They were rarely more than one-page documents, and those that were often weren't numbered within the documents because no one expected them to be used anywhere but online. And further, clicking on one usually triggered Adobe Acrobat Reader or whatever other software you had installed to read them, externally entirely from your browser.

I remember over at WP:NRHP we often had to deal with scanned copies of old nominations to the National Register which had never been numbered because no one expected them to be so easily readable when they were created and reviewed. That may have been where we used software-assigned pagination.

Nowadays, with a lot of online PDFs being versions of hard-copy documents even though they may be read much more online than off, that's less of an issue than it was, but I must admit that even in 2017 I was still holding to the old way.

For things that clearly have a real-world existence, like journal articles or court decisions (anything using consecutive pagination, really), I have always used that pagination.

Of course there's another issue ... PDFs comprising a collection of documents from disparate sources whose on-page pagination isn't sequential, so you might have several possible page 7s. Yes, you could give the title of the collection of documents within the PDF, but a reader might not know that and wonder where your source is and, when they don't find it, plant a {{failed verification}} tag that really wasn't justified inline and give you a small headache. In that case it would make more sense to use the software's pagination; the issue that prompted Roy to start this thread would not be a problem because the document represented by the PDF wouldn't/doesn't exist in real life.

Daniel Case (talk) 07:01, 13 February 2024 (UTC)
Just to state the obvious, usage of {{rp}} (and {{r}} as well?) and whether or not it is desirable/deprecated or whatever should be in documentation. Generally, I think a lot of the grief over referencing is due to a lack of clear, up-to-date and well thought out documentation that is accessible to the ordinary editor. ThoughtIdRetired (talk) 08:39, 13 February 2024 (UTC)
I agree that much of the documentation around citations is just plain terrible. I suspect much of it makes sense to the insiders, but to people coming at this new, it's impenetrable. I also think (and I know I'm not going to convince anybody of this) we need to standardize on one common reference style that's used everywhere. Even if we went with a style that I don't like, at least it would be a common target that everybody could aim for, ranging from end users, to the people who document templates, to the people who write tools like Visual Editor or Citoid. RoySmith (talk) 15:44, 13 February 2024 (UTC)
You will get no argument from me when you say that documentation around citations is just plain terrible. If you think it needs improving, improve it. None of the cs1|2 documentation is protected; anyone who has the skill to improve the documentation may do so. If you know how to improve the citation documentation, don't complain about it, improve it.
If the past is any predictor of the future, what happens now will be absolutely nothing ... until the next time it is convenient to complain about citation documentation. Whereupon, I shall reissue this challenge, get the same result; wash rinse repeat.
Trappist the monk (talk) 16:02, 13 February 2024 (UTC)
With all due respect, that's blaming the victim. How can somebody who doesn't understand how something works improve the documentation which explains how it works? RoySmith (talk) 16:27, 13 February 2024 (UTC)
I don't think so. My experience suggests that newbies who don't understand ask for help rather than complain about documentation quality. It's the experienced editors who do understand that complain about documentation quality. I suspect that you are a member of that latter group. Because I am so close to the cs1|2 templates from a technical point of view, I tend to write from that point of view which is not really accessible to non-technical editors. In the real world, there are people (there were when I was living in the real world) who translate writing written by technical people into something that a user can understand. That is what we need here. Wikipedia has lots of people who are skilled at writing for those who don't 'know'. They are the people who can make citation documentation accessible. Will they? Will you?
Trappist the monk (talk) 16:59, 13 February 2024 (UTC)
It keeps going back to the fact that we are all volunteers here. I have experience at creating and presenting training materials for bureaucratic procedures and for computer use, but I edit WP because I enjoy finding good sources and adding content. That, and trying to follow the selection of project pages that are on my watchlist, already eats up more time than I am comfortable with spending on WP. I've also seen editors put in a lot of time revamping policy or guidance pages, only to see the community flatly reject the changes (although, what I'm thinking of happened many years ago). So, while I might consider helping on a rewrite of the citation procedures, I would want to be sure first that the community wants it to happen, and there was some clear consensus on what would constitute an improvement. That means there would have to be a lot more discussion than we have seen here. Donald Albury 17:57, 13 February 2024 (UTC)
If there were to be a revamp of referencing advice/methods/etc. it would make sense to start with analysing whether what we have actually meets the hopes and aspirations on the subject. A particular problem is the way the long form reference displays the page numbers in the text of an article. Fine if it is just one page number, but if you have two page ranges that support the article text, or a non-numerical location identifier, it takes a lot of article space. I am no fan of {{sfn}} (pain in the neck to edit) but it gives a much better result for the reader. Surely something like that could be replicated with long form references? ThoughtIdRetired (talk) 08:38, 14 February 2024 (UTC)
Should our aspirations on referencing include a target of reducing the number of ways of doing this? We already have parenthetical referencing deprecated. Is this a clue that we should perhaps target just two main referencing styles (long and short form) and achieve those with a limited number of templates that would give style consistency? The advantage would be for the reader, as they would consistently see better technical displays of referencing info (getting mouse-over displays, etc.). I suspect that it would also weed out some "problem" references  for example ambiguous identification of the work cited in a short referencing arrangement. I am not suggesting a massive conversion of all of Wikipedia to a new system, just preferring a more limited range of methods for new content. Clearly, a slimmed down methodology would be much easier to document for new and existing users. I appreciate this might upset the "Betamax man" editors who are wedded to an older system, but I think that is just a feature of a changing world. ThoughtIdRetired (talk) 08:53, 14 February 2024 (UTC)

Established citation style

What to do if original website was replaced by predatory dangerous website

Separate section for explanatory notes

Editor changing citation style from short to long formatting.

Splitting notes is an arbitrary Wikipedia-made standard with unknown consequences for readers

Discussion in Wikipedia:Village pump (policy)

Year of publication differs from that generated in citebook

Source citing question

Is this a record for excess cites?

Related Articles

Wikiwand AI