Wikipedia talk:Citing sources

citation generator?

What's the easy way for a new user to generate citations from archive.org or google.books for the 25 books listed on Sydney_Moseley#Works? Enri999 (talk) 05:32, 9 September 2024 (UTC)

@Enri999, there isn't an automatic way to do that correctly. You will have to look up each one of them manually. The convention is to use the first edition for such lists.

Auto-citing a source using the visual editor – look for the "Cite" button in the toolbar.

Once you have the URLs, here's an example of what you can expect as an autogenerated ref in the visual editor (though you'll have to switch to wikitext to remove the ref tags), using the first book:

• Moseley, Sydney A. (Sydney Alexander) (1916). The truth about the Dardanelles. Robarts - University of Toronto. London : Cassell.
• Moseley, Sydney Alexander (1916). The Truth about the Dardanelles. Cassell, Limited.

The first is archive.org and the second is books.google.com. (You can also generate citations from an ISBN, but these are all too old for that.) After they're generated, you can edit it to change anything that you think it got wrong. If you've not tried the visual editor, then this link will probably work for you: https://wiki.riteme.site/wiki/Sydney_Moseley?veaction=edit It works like Google Docs. WhatamIdoing (talk) 05:49, 9 September 2024 (UTC)

Thank you, I didn't know about that button, that will help. Cheers. Enri999 (talk) 05:54, 9 September 2024 (UTC)

@Enri999; WhatamIdoing: And like so, so, many automatically created citations, those are both flawed.

|first=Sydney A. (Sydney Alexander) – don't do that; write the name as it appears in the source: |first=Sydney A.

|publisher=London : Cassell – don't do that; write the name as it appears in the source: |location=London |publisher=Cassell and Company

|others=Robarts - University of Toronto – don't do that; the name of the institute that contributed the source to Internet Archive is irrelevant and may mislead our readers

|first=Sydney Alexander – don't do that; write the name as it appears in the source: |first=Sydney A.

|publisher=Cassell, Limited – don't do that; omit corporate designations unless required for disambiguation: |publisher=Cassell and Company

You cannot trust visual editor/citoid to auto-magically create correct citations; they are dependent on the quality of the metadata that can be scraped from whatever online source. Sure, use ve/citoid to fetch some of the source's metadata but you must check and correct each and every citation that the tool creates. Be responsible and don't create a mess that other editors will have to clean up.

—Trappist the monk (talk) 07:23, 9 September 2024 (UTC)

The first one also has the title in librarian-preferred Sentence case, even though most citations use Title case. No autogenerating system is going to turn bad metadata into perfect citations, and I think this is the level of imperfection that you can realistically expect. WhatamIdoing (talk) 07:37, 9 September 2024 (UTC)

That's probably true without a lot of postprocessing, and the initial algorithmic citations are still pretty bad. I'm not sure what all can be done to make it clear that any citation generation script is a first-pass tool that will almost always produce output requiring manual adjustment. I'm hoping Edit Check might help eventually, as would surfacing CS1 maint messages within the VE interface, as would additing a translation layer on top of Citoid, adding lots and lots of special cases, etc. I personally find that automatically generated citations typically require so much tweaking that it's generally not a timesave even to begin with them unless there are more than eight or nine authors. Folly Mox (talk) 22:00, 9 September 2024 (UTC)

I guess there is a reason why I just plod along building citations manually. Donald Albury 22:47, 9 September 2024 (UTC)

@Folly Mox: What is "Edit Check"? Rjjⁱⁱⁱ (talk) 03:25, 10 September 2024 (UTC)

See mw:Edit check. The first "check" is encouraging new editors to add citations, if they are adding a new paragraph. (Of course there are other times when adding a citation would be appropriate, but it's an easy-ish thing for the software to detect, and it's almost never a bad idea to add a citation if you write, e.g., a whole new paragraph.) WhatamIdoing (talk) 04:21, 10 September 2024 (UTC)

I have high hopes for future extended functionality that can provide realtime feedback about mistakes and problematic edits. Something a bit more nuanced and informative than edit filters. Rjjiii, it's not currently implemented on English Wikipedia at all, and the stretch goals are yet but dreams. Folly Mox (talk) 04:53, 10 September 2024 (UTC)

I have my doubts about whether Clippy would be welcome, but there are a few things that could be handy (e.g., pre-warning about URLs that are on the Wikipedia:Spam blacklist). WhatamIdoing (talk) 05:15, 10 September 2024 (UTC)

Right, nobody wants Wikipe-tan intruding into their editing interface all Comic Sans "Looks like you're populating an infobox! Do you want to navigate away to a tangentially related Help: page instead?" I'm conceptualising the feature I've done exactly zero work on as more like a fully automated Twinkle, dropping boilerplate modals at rookie mistakes that established contributors tend over time towards becoming increasingly frustrated and bitey about.

Our documentation is... not really presented in a way that minimises common errors for newer editors. Presenting applicable guidance on an as-needed basis feels like it should be mostly positive. Folly Mox (talk) 06:31, 10 September 2024 (UTC)

One of the lessons from the original Clippy was that newbies appreciated its assistance only for a very brief period of time. When you had just bought your first-ever computer, at a cost of almost a month's income, any friendly-looking help was appreciated, especially if you'd never used a computer before, and might not even know how to type.

But after the first jitters wore off, most people learned how to use their new computers quickly, and they equally quickly wanted to get rid of anything that treated them like a newbie.

With that in mind, it's possible that we should design for universal use (e.g., autogenerated refs, because even though they're imperfect, they are very popular with editors of all experience levels), or for bots that autofix the rookie mistakes (e.g., we don't have to revert newbies dropping Facebook links into articles, because XLinkBot does it faster than humans can). WhatamIdoing (talk) 07:49, 10 September 2024 (UTC)

I agree (I think you're advocating this; please correct me if not) that improving Citoid's output (in addition to that of scripts like reFill and Citation bot) is probably a more fruitful avenue for raising the quality of citations added across experience levels. I think this starts with some sort of community configured functionality that hooks into Citoid or VE somewhere, which there was a subthread about at the recent VPR thread on Edit Check.

As stated above, No autogenerating system is going to turn bad metadata into perfect citations, so for a "full solution", one necessity seems to be some sort of acculturation into the practice of double checking code output rather than blindly trusting it.

As to the other issue we appear to have differing perspectives on – the hypothetical future usefulness of / annoyance with potential Edit Check cases – I'm not sure if we actually disagree or if we're not understanding each other's prior assumptions.

I feel like I would derive value as an editor from a script that warned me if I e.g. left a date or copypasted superscript numeral in an author name field, similar to the thing that prompts me if my edit summary is blank (which doesn't function in Minerva). And I would derive value as a patroller from a process that prompted people e.g. not to change a shortdesc manually set to none into a vacuous / pleonastic synonym of the article title because they think having no short description is erroneous and are unaware of WP:SDNONE, or altering variant English spellings to their preferred lect's because they're unaware of valid alternatives and MOS:ENGVAR, or altering era styles from BCE/CE to BC/AD or vice versa because they have feelings about it and are unaware of MOS:ERA, or altering ordinals from English words to numeric representations because they're unaware of MOS:ORDINAL, etc.

All of these are pretty common, typically reverted, and not bot-addressible because they'd all fall under WP:CONTEXTBOT, and could be the valid result of a talkpage discussion, conformance with a "Use Regional English" template, or something similar, although the likelihood is low.

Having thought on it a bit instead of getting ready for work, I suppose the initial Edit Check message about adding a reference to new uncited paragraphs might be encountered frequently enough to generate annoyance in users (and maybe the idea of warning for impending 3rr violation, although a warning for 1rr on affected articles would probably be more valuable), but most of the things I'm envisioning should probably display once, create a moment of education, and then not be triggered again for the same user unless they are disruptively editing against consensus practice, in which case we should want to annoy them, which might help them stop without technical restriction. Folly Mox (talk) 11:25, 10 September 2024 (UTC)

I don't remember what the config plan was for Reference Check, beyond it only being shown to people with less than 100 edits. So that puts an absolute maximum of 100 encounters, though there was some talk about a maximum number of times it could be displayed, which could be set as low as once.

A warning for WP:3RR could be valuable to people of any experience level. It would be easy to figure out cases where it obviously shouldn't trigger (e.g., 0–4 edits have been made to this page during the last 24 hours; your first edit to this page during the last 24 hours; the most recent edit was made by you; nobody except you has edited this page during the last 24 hours) but impossible to detect all the cases where it should trigger. WhatamIdoing (talk) 19:39, 10 September 2024 (UTC)

Yeah, I've seen new editors describe the three-revert rule as a kind of a trap. Nearly all experienced editors are aware of it, and new editors who break it sometimes claim they've been tricked. Rjjⁱⁱⁱ (talk) 12:27, 21 September 2024 (UTC)

Especially after we changed the rule (many years ago now) from primarily covering straight-up reverts to include trying again in a different/better/collaborative way, I can easily imagine people feeling like they've been tricked. WhatamIdoing (talk) 17:15, 21 September 2024 (UTC)

@Enri999: I'm late to the party, but have you tried Zotero? I know that you can easily capture citations into your Zotero library and then export them from the program as Wikipedia citations.--Gen. Quon_[Talk] 13:15, 1 October 2024 (UTC)

Hey, I was wondering if anyone can help me with citing sources!

Hey, I'm new, I recently joined today and I've been working on a particular article, I finished recently Bianca Giovanna Sforza, (Early Life). But I don't want to look like I'm spreading misinformation so, can anyone show me how I can improve and how I can cite my sources. Hectorvector7 (talk) 08:42, 21 September 2024 (UTC)

@Hectorvector7: Yeah, thanks for expanding that article. I see citations added to search results.[1] Do you already have sources that you are working from? Or is step one finding sources about Sforza? Rjjⁱⁱⁱ (talk) 12:29, 21 September 2024 (UTC)

Bundles

Hey @ActivelyDisinterested, I saw that you reverted me. However:

1. I just wanted to demonstrate one way to achieve the appearance of what the section meant. I didn't mean to advocate for a particular template, and I didn't use naked <ref></ref> because that wouldn't work if the reftags were real.
2. The paragraph is about the case in which each source applies to the entire sentence. The case for using {{multiref}} is for if the sources each support a different portion of the preceding text, which is in the next paragraph.

Aaron Liu (talk) 21:25, 27 September 2024 (UTC)

There are many times that multiple references are for the same sentence and multiref could be used. You'll see it quite commonly used for contentious statements. The use of {{refn}} is really only needed if refnames are being used. -- LCU ActivelyDisinterested «@» °∆t° 21:54, 27 September 2024 (UTC)

To be clear you could use any of the templates from Template:Multiple references to achieve the same end, not just multiref. All the options are valid regardless of paragraph. -- LCU ActivelyDisinterested «@» °∆t° 22:00, 27 September 2024 (UTC)

Some examples of refn and multiref usage from my sandbox[2] For some reason the multiref formatting is tiny (at least for me), this isn't the usual formatting. Whether the final reference looks like ref #5 or ref #6, they are both valid ways of bundling the references. -- LCU ActivelyDisinterested «@» °∆t° 22:18, 27 September 2024 (UTC)

@ActivelyDisinterested: on Chrome and Firefox I see the same font size. There is less spacing between the lines in an {{unbulleted list}}:

{{Multiref}}	{{Unbulleted list}}	{{Bulleted list}}	{{Multiref2}}
One Two Three	One Two Three	One Two Three	One Two Three

If that helps to see it side by side, Rjjⁱⁱⁱ (talk) 03:53, 28 September 2024 (UTC)

The font size could be an issue with the page, my sandbox contains some odd things. -- LCU ActivelyDisinterested «@» °∆t° 10:12, 28 September 2024 (UTC)

Maybe, but I mean on your sandbox page I see exactly the same font size. Rjjⁱⁱⁱ (talk) 13:36, 28 September 2024 (UTC)

Archive date AND access date

Random question that I can't seem to find a good answer to: If I'm citing a source and I include an archive date, including the original access date seems unnecessary, no? The citation should at that point indicate a) when it was originally published and b) when a snapshot was 'captured'. This leads me to think that the addition of an access date is just extra information that doesn't really tell a reader anything. Any thoughts?--Gen. Quon_[Talk] 13:17, 1 October 2024 (UTC)

We had a big discussion that touched on this topic last year at Help talk:Citation Style 1/Archive 91 § Do we need |access-date ? (disclosure: I participated).

There are differing opinions here. My opinion aligns somewhat with yours, Gen. Quon, in that I find |access-date= to be entirely superfluous when an archive snapshot is provided and the original URL is dead.

Other editors find |access-date= in these cases still to have value for other purposes, like prioritising verification of text–source integrity, on the basis that citations added a long time ago have had more opportunity to have the claims citing them altered by editing over time.

Other editors find |access-date= to be unnecessary clutter in most cases.

Probably the reason you aren't able to find a good answer is this disagreement. If you don't feel like adding |access-date= when adding an archive snapshot, you don't need to, but don't go around removing |access-date= from other citations with archive snapshots, and don't be surprised if bots later modify your citations to include an |access-date= (this happened to me; best to let bots do what they want, since they'll just do it again if you revert them). Folly Mox (talk) 17:58, 1 October 2024 (UTC)

Folly Mox Thank you for the comprehensive answer! Those were exactly the sort of thing I was curious about, especially with regard to dead URLs that are archived.--Gen. Quon_[Talk] 21:52, 1 October 2024 (UTC)

I have long regarded access-date as useless clutter in most cases. What I find particularly irritating is news citations that give an access date but not the publication date, which is much more important. -- Alarics (talk) 07:56, 2 October 2024 (UTC)

a) I find |access-date= helpful when trying to find an archived snapshot if the original link has died. b) Not every web page provides a date of its publication; |access-date= gives at least some indication. -- Michael Bednarek (talk) 09:25, 2 October 2024 (UTC)

Linking google drive

Here's a fun one for the crew: I'm working on rescuing someone else's draft, and I'm citing some publications in a theological periodical where the official archives (as accessed through the link just a few words back) are hosted on google drive.

For these sources, do I:

• ⓐ direct link the google drive pdf for the cited article like it's no big deal
• ⓑ link the issue containing each article, so it's possible to navigate back to the archive root and more clear I'm not just tossing a random Chinese pdf onto google drive myself (adding an |article-number= or other navigation instructions since obviously the pdfs are not titled the same as the articles)
• ⓒ no link
• ⓓ some better idea I haven't thought of

I'm leaning ⓑ, but it's a weird case I haven't encountered before. Folly Mox (talk) 22:43, 1 October 2024 (UTC)

ⓑ sounds plausible, as it may well be more robust than ⓐ, plus the link will look more "official". Gawaon (talk) 06:58, 2 October 2024 (UTC)

It wouldn't hurt to leave a hidden note so that anyone coming across it later knows why a google drive link is being used. If I came across this I would be concerned that someone had posted it to google drive without copyright permissions. -- LCU ActivelyDisinterested «@» °∆t° 14:20, 2 October 2024 (UTC)

I've got an html comment in one of the four articles cited, but I might add additional clarity with something like |via=Google drive (official archives). Folly Mox (talk) 16:15, 2 October 2024 (UTC)

Citations for individually authored chapters in edited books

Are citing individual chapter authors in edited books a requirement? Or is it optional? I have thought citing the entire edited book is acceptable [3], but that might not be the case [4]2.4.2? Bogazicili (talk) 17:16, 6 October 2024 (UTC)

Edited books tend to have chapters/sections/contributions authored by someone other than the editor. If you are citing an author's chapter/section/contribution in an edited book, state that in your citation; don't make our readers hunt for whatever it is that you claim supports our article. Be specific down to page number: Title (name the editor(s)) → section (name the author(s)) → page(s) that support our article.

—Trappist the monk (talk) 17:35, 6 October 2024 (UTC)

I always give page numbers in references for books. When not available, for example in dictionary-like sources with entries, I use the |loc parameter. Are individual authors still required for chapters? Bogazicili (talk) 17:38, 6 October 2024 (UTC)

Provide the chapter contributor if you have that information. Folly Mox (talk) 17:41, 6 October 2024 (UTC)

Of course I have that information. I'm mainly asking this for Turkey. In the Turkey#Sources section, I have a lot of edited books with individually authored chapters, like [5].

I was thinking if I should switch to a style such as Climate_change#Sources, where individual chapters are cited and are grouped by report/book. Bogazicili (talk) 17:49, 6 October 2024 (UTC)

If these are for short form references you can use {{harvc}}. It creates a link between the short form and the full cite in which you can put the chapter details. -- LCU ActivelyDisinterested «@» °∆t° 17:55, 6 October 2024 (UTC)

For example see the second cite in Bible#Works cited. The cite is for The Literary Guide to the Bible, while the two entries below it are chapters in that work. -- LCU ActivelyDisinterested «@» °∆t° 17:59, 6 October 2024 (UTC)

{{harvc}} was exactly what I was thinking. But I was thinking about grouping everything in Turkey#Sources by Books, Collections, Reports and Others. Collections would be edited works with individually authored chapters. But your example actually looks much cleaner.

Are all of these acceptable, or is any one of them preferred:

Grouped by various criteria: Climate_change#Sources

Chapters under source with {{harvc}}: Bible#Works_cited

Chapters separate, might be multiple citations to same book. For example: "Smith, Kirk R.; Pillarisetti, Ajay (2017). "Chapter 7 Household Air Pollution from Solid Cookfuels and Its Effects on Health". In Kobusingye, O.; et al. (eds.) ..." Sustainable_energy#Sources? Bogazicili (talk) 18:11, 6 October 2024 (UTC)

I don't believe there's a dictated way it has to be done, just options you can choose from. -- LCU ActivelyDisinterested «@» °∆t° 18:48, 6 October 2024 (UTC)

Saying that the situation at Climate change, and other articles using the IPCC reports, are like that because of how the IPCC reports get published. They are forever having to be fixed, as it's such an unusual way of linking cites. So personally I wouldn't use that method, if I was doing it I would use {{harvc}}. -- LCU ActivelyDisinterested «@» °∆t° 18:57, 6 October 2024 (UTC)

I tried replicating it with harvc but I don't think you can get the desired output given all the authors are listed as IPCC. There doesn't seem to be a way to identify a custom chapter reference ID, similar to book reference ID with |ref={{harvid... Bogazicili (talk) 20:47, 6 October 2024 (UTC)

Yes harvc is limited to the default author/year. -- LCU ActivelyDisinterested «@» °∆t° 21:07, 6 October 2024 (UTC)

There's no "official" preference for one of those three. If you're instead asking for personal preferences, I prefer that like under Bible (source with sublist chapters) (eg in Julius Caesar). Ifly6 (talk) 18:51, 6 October 2024 (UTC)

An important point for the reader is if one of those individual chapter authors is also cited in another work. If you want to track the opinions of various authors and how they contribute to the article, then it is better to see the bibliography ordered by those authors. The grouping of authors in one edited volume destroys the ability to quickly assess where the article content comes from and may conceal a reliance on one author. (That reliance may be appropriate in some subjects, but that in itself may be useful to the reader.) ThoughtIdRetired _TIR 20:00, 6 October 2024 (UTC)

Thank you for the answers everyone!

Now, if multiple chapters within a single edited book need to be cited, what's the preference? Citing the entire book with a single reference, or using multiple separate references to each chapter? Bogazicili (talk) 20:50, 6 October 2024 (UTC)

There was a rather rambling discussion (I did not reread it) at Help talk:Citation Style 1/Archive 59 § Proposal for an "in-title" ("In" + title) parameter that I think was what caused me to create this sandbox page which may be of interest.

—Trappist the monk (talk) 21:55, 6 October 2024 (UTC)

In climate change, we follow the reference suggestion in reports (see part about Referencing this report: [6]) So they all have IPCC as the author. I think the way for harvc to work in this case would be a |ref-chapter, a unique custom ID for the chapter. But I doubt it's worth the trouble if no other page needs it. Bogazicili (talk) 22:06, 6 October 2024 (UTC)

I've had some edges cases before like citing works preserved fragmentarily in citations within a different work that has survived transmission to the present day. For these, {{harvc}} doesn't work because the publication dates for the fragmentary works strictly precede that of the work they're now found within. I'll just create regular citations with custom citeref values and format them similar to chapters, indented one level and underneath the full work they appear in. A similar approach might work for your IPCC chapters. The metadata probably ends up kinda garbagey, but it makes sense to someone actually reading the article. Folly Mox (talk) 22:14, 6 October 2024 (UTC)

Thanks everyone for the answers. By the way, one suggestion I would make after the above discussion is adding something here: Wikipedia:Citing_sources#Books

Currently: Citations for individually authored chapters in books typically include:
Suggestion, something like: Citations for individually authored chapters in books are recommended. These typically include:

As I said, I previously thought this was optional, but there seems to be a strong preference for it Bogazicili (talk) 22:17, 6 October 2024 (UTC)

Remove citation tools from this page

I propose that most of the § Citation templates and tools be removed from this page. First, I don't think a content guideline page is a good place for this content. Second, it duplicates an existing page, Help:Citation tools. To be precise, I propose that the following sections be removed from this page, and merged into Help:Citation tools insofar as they are useful there:

• § Citation generation tools
• § Duplicate reference finders
• § Programming tools
• § Reference management software

The text directly under the section § Citation templates and tools and the § Metadata section should stay. Daask (talk) 18:56, 9 October 2024 (UTC)

That sounds like a good idea to me. Duplicated pages mean we have to update everything twice. WhatamIdoing (talk) 01:13, 11 October 2024 (UTC)

Agree Moxy🍁 01:14, 11 October 2024 (UTC)

How to cite something in newspapers.com?

What's the right way to generate a URL for a publicly-viewable clipping in newspapers.com? Cannonball (Milwaukee Road train) had a reference that linked to https://www.newspapers.com/image/1066814482 but that gets you to "You need a subscription to view this page" if you're not logged in. So I logged into my account and generated a clipping, which has a URL of https://www.newspapers.com/article/the-waukesha-county-freeman-cannonball-c/159032901/ which is only marginally better; if you're not logged in, it gets you to an image of the page that's too small to read the type, and if you click on it, you're back to "Create a free account, or sign in". I thought the idea of a clipping was that it was publicly viewable. Am I just doing it wrong? RoySmith (talk) 22:00, 14 November 2024 (UTC)

@RoySmith, so far as I can tell, a clipping image is always the same width for logged-out viewers. So, if you're clipping one column, even if it's a long one, then the legibility is good. Clipping a whole page across will come out fuzzy. Wikipedia:Newspapers.com says that we're meant to use clippings rather than "/image/" links, so I've been doing it that way. Rjjⁱⁱⁱ (talk) 00:57, 15 November 2024 (UTC)

Per Rjjiii, clipping image can be seen by non-logged or logged-out viewers and you should take a news block for clipping instead of the whole page and use the "/article/" link. Here is an example (taken from a citation in WXYZ-TV)

<ref>{{cite news |last1=Johnson |first1=L.A. |date=February 3, 1995 |title=Channel 4 newscasts take the ratings lead in Detroit |url=https://www.newspapers.com/article/detroit-free-press-channel-4-newscasts-t/120083876/ |url-status=live |archive-url=https://archive.today/20241012091020/https://www.newspapers.com/article/detroit-free-press-channel-4-newscasts-t/120083876/ |archive-date=October 12, 2024 |access-date=March 3, 2023 |work=[[Detroit Free Press]] |pages=3F, [https://www.newspapers.com/article/detroit-free-press-channel-4-news-wins-r/156927271/ 6F] |via=[[Newspapers.com]]}}</ref> Vulcan❯❯❯Sphere! 09:08, 19 November 2024 (UTC)

In my case, the original article was laid out so as to span the full width of the page. RoySmith (talk) 15:54, 19 November 2024 (UTC)