Wikipedia talk:Citing sources

Lazy citations

I have been coming across many instances of citations that use the ref name format for a text, several times in an article. Editors use this as a lazy way of citing without page numbers. In this format, there seems no place to put page numbers, so these articles end up with lots of cn or page needed tags. I have been replacing these inline citations with a shortened form (eg. Smith (1999), p. 12.). However, this seemed to require a bibliography entry with the text name in full. The WP article Help:References and page numbers seems to suggest that the full text reference would be put in the references section and the shortened citation in notes. I don't see this very often, and it would surely be very confusing. A solution would be to allow the ref name format to specify a page number, but I can't find a way of doing that. Myrvin (talk) 08:57, 14 September 2014 (UTC)

I am now trying rp|12 to indicate page number, but I'm not sure readers will understand it. Myrvin (talk)

rp|page=12 seems better. Myrvin (talk) 12:21, 14 September 2014 (UTC)

@Myrvin: The {{rp}} template is disliked by a number of editors, partly because it means putting the page number some distance from the rest of the ref. One way, as you noticed, is Shortened footnotes - you can see this in use at NBR 224 and 420 Classes; another is to use a full citation for the first use of a given source, and a shorter form for subsequent uses where only the page number differs - this is what I did at Boar's Head railway station#References, refs 2 & 6. --Redrose64 (talk) 15:16, 14 September 2014 (UTC)

Myrvin: What do you mean by "use ref name format for a text"? Are you referring to the inclusion of explanatory text or comments in a note? Could you provide some examples?

Your surmise that that a shortened form, or "short cite", requires "a bibliography entry with the text name in full" is correct. That is the "full citation" Redrose just mentioned. The beauty of short cites linking to full citations is that only one full citation is needed for each source, with as many short cites as needed, each customized with specific page numbers or comments as needed. Putting these into separate sections — typically "Notes" and "References" — is quite common. ~ J. Johnson (JJ) (talk) 19:57, 14 September 2014 (UTC)

Thanks for the comments. The problem with "the first use of a given source" is that often we are changing an existing article. The new edit, or the page-less cite, may well precede the first naming of a given source. It can also be difficult to find it. I have edited many articles where the full name of the ref can be anywhere. By ref name format I mean e.g. ref name="Barber, Spanos 1974"/ for a citation. In most of the articles I have edited, there are no Notes sections; cited sources all go in References. Suddenly introducing Notes, or even Bibliography, seems a big change to a settled article. See my recent edits to Hypnosis.Myrvin (talk) 06:46, 15 September 2014 (UTC)

  What you are referring to as a "ref name format" is what we call a "named ref", referring a construction in the form of "<ref name=xxx />", which will link to another "ref" in the form "<ref name=xxx ... </ref>". (This is the "named" variant of the "<ref>" entity.) Note: "ref" tags (named or not) are not references (despite the name), nor citations. They only make a place — specifically, a note — where citations can be put. The "named ref" is just a way of linking to a "ref" elsewhere. And you are correct, it can be difficult to find the master "ref", as they can be anywhere. For that reason I find them obnoxious.

  A better solution is put all of the full citations that provide the full description of a source (typically using the {{Citation}} or {{Cite xxx}} templates) into a dedicated section as described above, and then linking to them using short cites in the notes.

  Yes, doing this can be a big change, so always ask on the article's Talk page before proceeding, in case someone else is vehemently opposed. ~ J. Johnson (JJ) (talk) 22:01, 15 September 2014 (UTC)

I'm sure you are right about the name, although I don't think my usage was confusing. I'll try to do better and "use correct term" next time. However, I note that in WP:REFNAME they seem to be called "named footnotes" - perhaps the correct term will be hard to find. I disagree with you about the use of Notes - a note is not a citation reference, it's an explanatory note. WP:Citations says

If an article contains both footnoted citations and other (explanatory) footnotes, then it is possible (but not necessary) to divide them into two separate lists, using the grouping feature described in the Grouping footnotes section of the footnotes help page. The explanatory footnotes and the citations are then placed in separate sections, called (for example) "Notes" and "References" respectively.

WP:REFNEST gives an example of where this is particularly useful. However, WP:SFN seems to agree with you. And what about Bibliography? All very confusing. See discussion below on all this. Myrvin (talk) 07:02, 16 September 2014 (UTC)

It would take a braver man than I to change, say, Intelligent design the way you suggest. Myrvin (talk) 08:58, 16 September 2014 (UTC)

I note that [[WP:SFN}} Shortened footnotes with separate explanatory notes allows Notes for actual notes, Citations for short citations, and References for the full citations. I like that, but for large, stable articles I'm not going to change them into that. I like to think I have a life. I'll stick with {{rp|page=nn}} I think. Myrvin (talk) 17:56, 16 September 2014 (UTC)

Yes, it is confusing, and the confusion of terms and concepts is perpetuated for having taken root in various places. Yet I would have you disabused of notion that "a note is not a citation reference, it's an explanatory note." Which is an ambiguous statement: do you mean "reference" in the same sense whereby it is often taken to be synonymous with "citation", meaning the bibilographic details of a source? Or are you implicitly referring to the tags used for creating notes (aka "footnotes" or "endnotes")? Either way, you are trying to make a distinction that simply is not useful. ~ J. Johnson (JJ) (talk) 22:24, 16 September 2014 (UTC)

I disagree with J. Johnson's assertion that the English Wikipedia is using the wrong terminology. But he and the regulars all know that already, because we've had this discussion many times. In my opinion, and I believe it's one that's generally shared, any words that average, non-style-expert editors can make sense of is good enough for discussions like this. Don't worry too much about getting the terms "right". If people can't figure out what you're saying, they'll ask. WhatamIdoing (talk) 22:55, 17 September 2014 (UTC)

"Wrong" is somewhat relative, but the confusion of terms is self-evident. The problem of confused terminology is that it leads to confused thinking, such as the notion that "footnote" and "citation" are synonymous, and therefore should be processed by the same code. We can't even properly talk about these things (witness the prior discussions alluded to) because we keep running into differences of meaning. ~ J. Johnson (JJ) (talk) 23:09, 18 September 2014 (UTC)

Should we improve ref tags?

I've been talking off and on to James F on the mw:Editing team about this problem. He says that (in theory, at least) it should be easy to build in page numbers to the citation system. You'd type something like <ref name=Foo page=123>(Bibliographic citation for the whole source here)</ref> – or, for the second use of the same book, just <ref name=Foo page=456 />. The cite.php system would handle the rest.

The most important question is how the page numbers should then be displayed. The simplest system is to repeat the entire bibliographic citation each time, appending the page numbers at the end. Would that be an improvement over the current system? Do people have ideas about how they'd like it work? (Realistically, we are only able to have one "built-in" system across all wikis, so people will still want to have the ability to use manual options. The goal is to find the one that will work for the most cases.)

This isn't likely to happen this year, so we've got plenty of time to think about what we actually want. If you're interested, feel free to leave a note for me or ping me. Whatamidoing (WMF) (talk) 01:54, 15 September 2014 (UTC)

I think something like what User:Whatamidoing (WMF) says would be preferable. However, if the article is to look like an academic article, the first mention of a source should be in full, and the rest, with page numbers, in short form (or even ibid). I suspect this is too much to ask of an automated system. See my comments above on where the full name is often found - anywhere. The system would have to find the full name, move that to the new first position and redo the subsequent mentions of the source. Would there need to be a Bibliography, and would the system produce that or some editor? Perhaps the best idea pro tem is to repeat the full name with added page numbers in the References. Even that might be difficult. Myrvin (talk) 08:24, 15 September 2014 (UTC)

Also see Template:Bug.Myrvin (talk) 13:03, 15 September 2014 (UTC)

A full listing only on the first appearance and short forms are ok when footnotes appear on each page of a printed book, and the short form is only used if the full listing appears on the same page, but when lists become long it becomes hard to find the full listing. That is why, if sources are used many times each, it is better to have all the citations be short forms and have an alphabetical list with the full citations. Jc3s5h (talk) 13:28, 15 September 2014 (UTC)

@Whatamidoing (WMF): In general it sounds like a good extension. As far as the concrete display is concerned I suppose tastes will differ. However may it would be possible to offer 2 versions? One for duplicating the full reference each time and one for proving a shorthand like author or title + page numer only.--Kmhkmh (talk) 13:34, 15 September 2014 (UTC)

I'm dubious about the ability to create a good design and use more than one system. I wouldn't want to type <ref name=Foo page=456 style=short /> or <ref name=Foo page=456 style=full /> for each citation, and I expect that other people would feel the same way. Whatamidoing (WMF) (talk) 23:25, 15 September 2014 (UTC)

Thinking about it a bit more, <ref> and </ref> are used with many styles of referencing, not just citation templates. The code for these things does not examine or parse the citation; it can't, since there are no Wikipedia standards for the format of the citation. So the <ref> and </ref> thingamabobs parser tag hooks can't possibly shorten a citation, because the citation is not parsed. Jc3s5h (talk) 13:57, 15 September 2014 (UTC) modified 2:15, 16 September 2014 (UT)

I find it odd that <ref> should be used for notes and not references. I see it in many articles, but academic books will often have a Notes section and then a References section for sources that have been cited. Mixing the two seems confusing. Given that it is <ref> and not <note> and {{reflist}} and not {{notelist}}, the original idea was surely for citation references. It seems to have been extended to other things. I note that there is a {{refn...}} for notes, as well as {{notelist}} & {{efn}}. Not that I use that, but maybe I should. Myrvin (talk) 18:44, 15 September 2014 (UTC)

If Wikipedia had specified and developed a single comprehensive referencing system right from the start, it would probably have been different, and the names of the various page sections and elements would almost certainly have been chosen more carefully and enforced more rigidly. But over something like fourteen years, the several systems that we do have were built up piece by piece; and each new piece has been given a name that is not necessarily the best, but the most suitable of what's left over after everybody else has picked something.

{{efn}} and {{notelist}} are quite new; they date back to December 2011; the {{ref}}/{{note}} system is older by something like seven years, and needs more care in its use so that the right ref links to the right note. {{refn}} isn't really a system as such, just an alternate form of <ref>...</ref> without some of the latter's limitations, much like {{reflist}} is <references /> with extra features, or like {{sfn}} is really <ref>{{harvnb}}</ref> wearing a shorter coat with more pockets. --Redrose64 (talk) 20:39, 15 September 2014 (UTC)

The articles on Template:Reflist, WP:SFN and WP:Citations seem to advise different ways of doing all this. Myrvin (talk) 08:39, 16 September 2014 (UTC)

Yes; as I say, it was built up piecemeal over many years. There is a lot of flexibility, and thus a lot of variation. Where there is a contradiction between those three pages, a careful examination shows that these are merely alternative approaches, none has precedence over the others.

Without being too formal, the rules for references go something like this. 1: WP:V is policy (those other pages that you named are not). 2: if adding sections for notes, references, etc., try to abide by MOS:APPENDIX. 3: if you're editing an existing article, stick with the ref styles and conventions already established in the article (this is WP:CITEVAR), unless there are very good reasons to change it. 4: if you're writing an article from scratch, use any referencing system that you're comfortable with, but be consistent within the article. Consistency isn't required by any policy, but it looks good and does become a requirement if you intend getting an article to FA-Class. 5: don't be afraid to ask for assistance.

I primarily use Shortened footnotes, mainly because I draw information from a variety of pages on a small number of books (as with NBR 224 and 420 Classes that I mentioned earlier). --Redrose64 (talk) 14:17, 16 September 2014 (UTC)

In articles with few citations then long citations with ref tags are not a problem, but as the size of the article increases with a corresponding increase in citations then I think short citations are a help. This is for several reasons:

1. When I convert an article that has a mixture of long and short citations to a reference list with short citations it frequently happens that mixed into the ref tags are citations that are missing enough information to meet WP:V, typically missing either the page number or enough extra information to identify the edition. This gets hidden when the the ref tags create a mixed list at the bottom of the article of slightly differently formatted long citations all mixed up together, particularly if the list also includes {{cite web}}s that do not include page numbers (spotting chaff in the wheat)
2. The second one is that there can be a large saving of space on an article with around 100 to 150 citations typically about 10K in my experience, something that ought not to be ignored.
3. A "Notes" section with multiple columns of short citations I think looks visually cleaner than a mix list of long and short citations.
4. An alphabetic list of long citations is I think of more use as a bibliography than a mix list of long and short citations not in alphabetic ordering (which is what one gets with just a {{reflist}}).
5. When the first citation is a long one and short citations depend on it. A deletion of that first citation means that the later short citations are no longer supported with a long citation. This is less likely to happen if the long citation is in a separate references section. However if the References section ends up with a few long citations that no longer support short citations, then while undesirable and needs cleaning up it is a fail-safe situation.

One problem I have seen with short citations is when an editor assumes that because only one book is cited there is no need to include year in the short citation. This can cause problems down the line when another edition of the book (or another book by the same author is cited) working out to which long version a short citation refers can be very time consuming involving a trawl through the edit history trying to find it. -- PBS (talk) 18:07, 16 September 2014 (UTC)

Yes! I concur in full. ~ J. Johnson (JJ) (talk) 21:24, 16 September 2014 (UTC)

What do you feel ought to be done with explanatory notes? If they are included in Notes along with short citations, doesn't that look less clean? Myrvin (talk) 18:44, 16 September 2014 (UTC)

Once again, see here - the sole explanatory note is in a separate list with letter identifier instead of numeral. Similarly with this article - here the explanatory notes (lettered again) go first, because one of them is referenced (and the other one ought to be but isn't). --Redrose64 (talk) 20:44, 16 September 2014 (UTC)

I don't know if the various element names would have been chosen more carefully, but they should have been, because confused terminology and sloppy conceptualizing has wreaked all kinds of confusion and grief here. A prime example is this widespread notion that the "ref" in <ref>...</ref> tags is short for "reference", as in "the full bibliographical details of a work referenced". Sorry, no, "ref" is just a poorly chosen name for tags that mark a kind of bucket (otherwise called a note) that the software automagically processes in a certain way. Bibliographic details don't mean shit to <ref>...</ref> tags. Or perhaps I should say: they are not distinguished from anything else that can be put into these buckets. Nor should they. While there is good reason to keep full citations separate, there is no benefit in segregating short cites from any other comments or explanatory material.

WhatamIdoing has overlooked that facilities for handling page numbers are already built into "the citation system". E.g., her second example could be done as: <ref> {{Harv|Foo|p=456}}.</ref>. (Although it would be better to follow the standard convention of providing a year.) ~ J. Johnson (JJ) (talk) 21:36, 16 September 2014 (UTC)

@J. Johnson: writes "... facilities for handling page numbers are already built into 'the citation system'." Not exactly. The <ref> mark up is part of mw:Extension:Cite and, while optional, is widely used in articles that do not use citation templates, while {{Harv|Foo|p=456}} is only useful in articles that use citation templates. Another factor that the two facilities are, I surmise, maintained by different groups of editors or developers. Jc3s5h (talk) 13:52, 17 September 2014 (UTC)

When you say Harv templates are "only useful in articles that use citation templates", are you referring only to the specific template named {{citation}}? In that case your statement is wrong, as Harv works just fine with the {{cite xxx}} family of templates. If you mean "citation templates" generally, and are referring to articles where the editors prefer to manually format their citations without "citation" templates of any kind, then sure, Harv templates would be rather pointless.(But then the example is even easier: "<ref> Foo, page 456.</ref>") But why should we care? Where editors use the "citation" templates (citation/cite/vcite) page number handling is already available. Where they choose to not use "the citation system" they should not expect citation functionality to be added to the <ref> function. ~ J. Johnson (JJ) (talk) 20:01, 17 September 2014 (UTC)

The {{harv}} template is just one way of making a Harvard ref, but it's not obligatory. It's possible for an article to use Harvard references without using any templates. See Wikipedia:Harvard referencing, where the first half of the page (down to the Linking inline and full citations heading) uses no templates for the ref examples. --Redrose64 (talk) 20:26, 17 September 2014 (UTC)

@J. Johnson:, I was referring to any of the templates compatible with the Harv template, including {{citation}} and CS1. Citations that don't use any templates can still use mw:Extension:Cite, which is part of a citation system. Jc3s5h (talk) 21:30, 17 September 2014 (UTC)

The existence of local templates is irrelevant. My goal is a system that works as a "built-in" product on 800+ WMF wikis. Only about 1% of WMF wikis have a copy of {{harv}}. WhatamIdoing (talk) 23:06, 17 September 2014 (UTC)

@Redrose, sorry I didn't check out your examples before. They look very neat and tidy. However, I note that you did distinguish between the explanatory note and the short citations. I too think that is a distinction worth making, although others here do not. I still would prefer that the explanatory notes went into a separate section. Perhaps in Notes while short citations go in Citations. Myrvin (talk) 12:55, 17 September 2014 (UTC)

If you work with a wide range of article types, you will see all sorts of variants of the way WP is used to cite references. There is something quite small like Hypnosis with all citations and explanatory notes put in References, and for which I have recently added a Bibliography so that I could use short citations to give page numbers. On the other hand, something larger like Intelligent design, which has a Notes section with explanatory notes often associated with citations and a References which has all other citations. Editors in ID see a reason to separate explanatory notes from general citations and have done so. It also has a Further Reading section, which is often used for the full reference for a short citation in References. There is also William Shakespeare with Footnotes (explanatory), Citations (mostly short with some explanatory notes} and References (Full references in alphabetical order}. Myrvin (talk) 16:10, 17 September 2014 (UTC)

Just because the citation tools here are flexible enough that editors can split their notes and citations into separate sections does not mean they must be so split. Reference lists can also be split into popular and scholarly works, or by language, or any arbitrary scheme. Notes added by an editor could (conceivably) be segregated from the original notes. None of these require special versions of <ref> to do what is not its job to do. ~ J. Johnson (JJ) (talk) 20:04, 17 September 2014 (UTC)

I don't think we are discussing what must be done, but what could be done to make perhaps large and complex articles, clearer and better articles. I happen to think that separating explanatory notes from citations (particularly short citations) would do that. So any change to how referencing and page numbers etc. are handled should allow that. Myrvin (talk) 20:22, 17 September 2014 (UTC)

Yes, I'd like to hear more about the ideal–and the ideal for a wide variety of needs. For example, PBS's preference for separating short citations and full citations is one that I share for articles whose sources are entirely books, but it's not one I share for articles whose sources could accurately be listed as a long string of PMID numbers. WhatamIdoing (talk) 23:06, 17 September 2014 (UTC)

If you feel that an article is improved by separating "notes" into explanatory, citation, and mixed, fine, you can do it: the "citation system" already allows that. (Just as it already handles page numbers.) What is a really bad idea is to modify a general function to do a wholly different kind of function in certain specific cases. Indeed, it seems to me that what WhatamIdoing may be proposing (the details are unclear) is just the kind of "a lazy way of citing" with which you started this discussion. ~ J. Johnson (JJ) (talk) 23:07, 17 September 2014 (UTC)

Since nothing's been decided, there are no details to share. However, since the point of the discussion is to see how we might provide built-in support for supplying (and varying) page numbers, it is very unlikely to to result in citations with no page numbers.

I think it might be possible to create a system that separates short and full citations, possibly using the ref name as the key. But it's probably not possible to create a system in which separating short and full citations is an option. <ref name="Smith 2010" page=123 /> is plausible. <ref name="Smith 2010" page=123 style=separate /> is not so plausible. Whatamidoing (WMF) (talk) 23:20, 17 September 2014 (UTC)

I think mandatory short citations created by the mw:Extension:Cite is not acceptable. I'm imagining if an existing article had a first instance of a given source such as <ref name = CMOS>''Chicago Manual of Style'', 14th ed. University of Chicago Press.</ref> and a second instance such as <ref name=CMOS/> the second instance would be rendered "CMOS". Since "CMOS" is not a universally understood abbreviation for Chicago Manual of Style, this would not be acceptable. I expect the encyclopedia is chock full of peculiar reference names that must not be used as the rendered text of short citations. Jc3s5h (talk) 00:04, 18 September 2014 (UTC) Fixed example markup 13:18, 19 September 2014 (UT)

Yes, you're right. One of the downsides to that approach is that it simplifies the wikitext, but requires anyone using it to use a refname that they want displayed to the reader. Things like Diberri's citation script produce refnames based on database numbers like "pmid12345678" or "isbn9781234567890", and short citations in the form of "isbn9781234567890 page 10" should not be inflicted on readers. The ones that default to first author's last name would require less work in renaming refs. Of course, a short citation shouldn't be displayed for any ref unless you were actually using the page-number feature, so there would be an element of choice about whether those names were exposed.

How would you provide page numbers, if not with short citations? Something like what {{rp}} does is also possible, but I'm not entirely thrilled about the way it looks (not that my personal preference actually matters, when it comes down to that). Can we come up with any other ideas for displaying page numbers without using short citations? Whatamidoing (WMF) (talk) 04:58, 19 September 2014 (UTC)

Why should we not use short citations? ~ J. Johnson (JJ) (talk) 22:06, 19 September 2014 (UTC)

Some editors don't like short citations. If the overall (multi-project) preference is to not use short citations, then we would need to find a method that (a) still supplies page numbers but (b) does not use short citations as the method for supplying page numbers. Whatamidoing (WMF) (talk) 17:53, 24 September 2014 (UTC)

The peculiar ref names result from editors thinking only of what is sufficient to distinguish the ref, and perhaps an implicit view that being internal (non-public) the actual name doesn't matter. It might be a big improvement to get editors to form a "name" as they would for a short cite. (E.g.: <ref name= "Smith 1978">.) But that reinforces the erroneous notion that the ref is the citation.

WhatamIdoing: the notion that there would not be details to share until something has been decided reeks of the attitude WMF has shown in the VisualEditor project. But in fact you have been discussing something — i.e., "handling" of page numbers — and you have some ideas of what is, isn't, or should be. It would undoubtedly aid discussion if you would provide definite and specific statements or examples, starting with the problem you that you percieve. ~ J. Johnson (JJ) (talk) 23:16, 18 September 2014 (UTC)

It's not, as your edit summary says, "decision first, then share the details". It's ideas first, find out what real people want first, and both details and decisions come much later in the process. When (if) we (=the people who are interested in this subject, both on this page and other pages and on the English Wikipedia and on other wikis) agree on some broad goals, then we can identify details. Whatamidoing (WMF) (talk) 04:58, 19 September 2014 (UTC)

At the moment some people use list within {{reflist}} where they store all the long citations using the parameter "refs=", but such lists are not directly visible. If there was a way to link named ref...tags pairs to an entry in a sorted list in the References section (as happens at the moment with the {{harvnb}} templates), then I think the proposal could well be worth investigating further. -- PBS (talk) 20:10, 19 September 2014 (UTC)

W: my edit summary was a query. My concern remains, that some of these ideas start with the pre-decided position that "the citation system" is broke, and therefore the "ref" system must be fixed. I have yet to see that there is any real problem with the tools we have, aside from editors with a deathgrip on having to do (or not do) things in a certain manner.

PBS: what do you mean by "Example text"? By "long citations" I presume you mean the full citations. By "parameter refs=" I hope you mean (yes?) the <ref>...</ref> tags by which we mark the text that {{reflist}} extracts to assemble as notes. Proceeding on that basis, note that {reflist} assembles these notes in the order they occur in the text — which is exactly as they should be. So when you refer to "a sorted list in the References section" probably you mean the typical list of full citations of the sources referred to — right? [Yes, W., I can always ask for clarification, but if we could agree on standardized terminology I wouldn't have to ask every time a term is used.] These lists are typically assembled manually by prefixing each entry with a "*".

With all that clarified (hopefully), please tell me: why should a note — that bit between the <ref>...</ref> tags — be linked to a full citation in a list of "references"? Why not have a short cite inside the note link to the full citation? ~ J. Johnson (JJ) (talk) 22:13, 19 September 2014 (UTC)

@J. Johnson: By 'list within {{reflist}} where they store all the long citations using the parameter "refs="', I understand PBS to be referring to WP:LDR. --Redrose64 (talk) 22:59, 19 September 2014 (UTC)

@J. Johnson: Yes, as Redrose64 wrote, I meant that which is described in the section WP:LDR. A full citation consists of a long and short citation. The long hold most of the information the short hold page number.One of the problems we have had in the last 8 years or so is too many ways to be able to produced citations. It is daunting for new editors and means that even experienced editors may not know all the options which makes maintenance harder than it needs to be. -- PBS (talk) 19:06, 20 September 2014 (UTC)

  I most certainly agree with your latter statement. For all that policy and the tools permit citation to done in any style imaginable, we have suffered from lack of: 1) a recommended good practice (newbies want to know a way of citing, not every possible variant), 2) clear and definite terms and concepts (that we can talk about these things without having to constantly define our terms!), and 3) an overall conceptual framework that integrates all the variant conceptions.

  E.g., you said that "[a] full citation consists of a long and short citation.". This is differs from my use of "full citation", and as that causes confusion we should sort this out. In referring to the non-short element I grant that there is use of "long". But much less than "full". Using Google as a rough guide, the balance is about 16,000 instances to over 2 million. More authoritatively, the Chicago Manual of Style uses "full citation" in my sense (that is where I picked it up); I don't recall any major style guide that uses "long citation". Therefore I suggest we should use "full citation" for the record with the complete bibilographic details, but not including any short cites that point to this record.

  If all that is sorted out, perhaps we can return to my previous question: why should a note be linked to a full citation in a list of references (full citations)? Why not link a short cite inside the note to the full citation? I wonder if what you had in mind was using "list-defined references" (WP:LDR, using {{reflist ref=...}}) to display what appears to be a list of references (full citations). I say appears, because what you really have there is a list of notes. Such a list appears to be a list of references only because that is what the notes contain. This is yet another "lazy" technique which ends up being harder because the references are still scattered through out the article. It really is easier to collect all of the bibliographical information in one area and link to it with short cites, rather then trying to force the notes (<ref>s) into making those linkages. ~ J. Johnson (JJ) (talk) 22:05, 21 September 2014 (UTC)

I have not been following this (very long) discussion. Having twice spent time and effort implementing enhancements to WP:Cite.php and proposing their incorporation via WP:Bugzilla, only to have that work overtaken by the event of the incorporation without discussion of changes which had not gone through Bugzilla, I feel as if I've been through the wars on changes to the way Refs work. Generally, my opinion about that is, "That way lies madness" (hence my not having followed this discussion).

I do think, though, that tightening up this content guideline by better defining different citation styles and, possibly, by describing some particular citation styles as "preferred" would be a good thing.

Also, I think that confusion re citations, especially for newbies, grows out of WP:GTL as much as or more than it grows out of this content guideline. WP:GTL#Notes and references says: "Editors may use any section title that they choose.[10] The most frequent choice is 'References'; other articles use 'Notes', 'Footnotes', or 'Works cited' (in diminishing order of popularity) for this material. ... Several alternate titles ('Sources', 'Citations', 'Bibliography') may also be used, although each is problematic in some contexts: ...". I think that work ought to be done on that before addressing questions here such as "Why not link a short cite inside the note to the full citation?". Wtmitchell (talk) (earlier Boracay Bill) 23:39, 21 September 2014 (UTC)

While MOS:LAYOUT (aka WP:GTL) is certainly part of the problem, I think it is rather peripheral to the core issue of how editors think citation should be done. And that can't be sorted out until folks take a closer look at what they are doing. Which is the point of my question. ~ J. Johnson (JJ) (talk) 21:28, 22 September 2014 (UTC)

Your speaking of a thing called a "note" prompted me to mention GTL. Definition of terms is important, especially in explanations intended for newbies. The terms note and reference in particular have a confusing number of meanings in the world of WP -- or so it seems to me. My digression there was uncalled for (but not, methinks, irrelevant).

More on the topic of this discussion section, one lesson I learned in my past life writing software (going back to before the coining of the term "software engineering", I think) was the importance of controlling scope (of awareness and of action in this case). That lesson learned has been very useful to me outside of the world of a code geek. As I see it, the job of cite.php (Refs & friends, that is) is to put "stuff" into footnotes, to render the footnotes and their forward and back links, and to do some housekeeping in aid of footnote reuse. to make cite.php aware of such things as page numbers (or any other details about the internals of the "stuff" in the footnote) would, in my mind, violate sensible scoping restrictions. Wtmitchell (talk) (earlier Boracay Bill) 02:52, 23 September 2014 (UTC)

In support of Wtmitchell's comment, the more the capability of separate but related bits of software overlap, the more difficult it becomes to make improvements in the future, because one find that a proposed change that works with one bit of software will break a different bit that is addressing the same capability. Jc3s5h (talk) 14:56, 23 September 2014 (UTC)

Exactly. Like the Unix philosophy of "one tool, one job", meaning a tool does one job, right, rather than all sorts of things half-assed.

A slight quibble with Wtmitchell: the various cite.php templates citation templates ({{citation}}, {{cite xxx}}, {{harv}}, etc.), such as handle page numbers, do not put stuff into footnotes: they format stuff and make links, etc., but only where an editor has put them. That could be in a footnote — or in the text, or in an image caption, infobox, etc. Cite.php (<ref> tags) puts this and other kinds of stuff into endnotes, but does not process any of it. Scoping is indeed a very relevant concept here, as processing footnotes and processing citation stuff have quite different scopes. Combining those functions makes about as much sense as having my laptop brew coffee. That anyone should consider combining <ref> (cite.php) and citation functionality I attribute to very poorly understood notions of those functions. (And non-programming backgrounds. :-) ~ J. Johnson (JJ) (talk) 22:15, 23 September 2014 (UTC) Clarified. 21:35, 24 September 2014 (UTC)

@J. Johnson: I'm a bit confused here. What are "the various cite.php templates"? Who is suggesting "combining <ref> and cite.php functionality"? --Redrose64 (talk) 07:21, 24 September 2014 (UTC)

See the first comment in this subsection "I've been talking off and on to James F. ...". -- PBS (talk) 09:27, 24 September 2014 (UTC)

@PBS: Was that a reply to me, or to J. Johnson? If it was to me, I know what cite.php is: it's <ref>...</ref> and <references /> - it has no templates ({{reflist}} was created by en.wp as a wrapper for <references /> but is not itself a part of cite.php). Indeed, if you make a link like Wikipedia:Cite.php or meta:Cite.php and follow it, you are directed to mw:Extension:Cite/Cite.php, which is a redirect to mw:Extension:Cite (as is mw:Cite.php), where <ref>...</ref> and <references /> are described extensively. Reading that, I see four templates mentioned: these are {{reflist}}, {{rp}}, {{ref}}, {{note}}; of these, the last two are only mentioned in the section Comparing ref/note style and Cite.php style where it is clear that the two templates are not part of cite.php. It's also clear that <ref> is already an inseperable part of cite.php, and so the phrase "combining <ref> and cite.php functionality" has no meaning because the one is already a subset of the other, and as far as I can tell, has been right from the start. --Redrose64 (talk) 11:37, 24 September 2014 (UTC)

I assumed that J. Johnson's line about "the various cite.php templates" actually meant "the various WP:Citation templates", which are technically unrelated to cite.php. Whatamidoing (WMF) (talk) 18:02, 24 September 2014 (UTC)

Yes, I was speaking broadly, even loosely. Sorry. Yet another example of why terms should be used carefully and precisely? ~ J. Johnson (JJ) (talk) 20:20, 24 September 2014 (UTC)

 I'd rather see the page rendering system take on more of this responsibility. The renderer could easily decide, e.g., to include a full citation for the first short cite of a separate reference. Whether the editor used rp with a named ref or sfn should not affect the appearance of the result. Such presentation decisions should be made consistently across an article, or better as part of each reader's preference (with an appropriate default). I would personally be thrilled to have a preference setting that put notes and refs on a separate tab, so that when I'm just reading I don't have to see them (wait for them to be rendered) unless I want to go there.

For that reason, I just want a very shorthand way to set things up. I strongly prefer cite/sfn, as it keeps things concise and doesn't bloat an article with repetitive long cites. I just finished cleaning up a long article that cited the same few works many times. The notes section is now much shorter and the references section is only slightly longer...

Here is a proposal I recently made:

Editors could save much tricky syntax with templates like:

{{cite gb|asdf|ref=harv}}

and

{{sfn gb|asdf|p=1}}

In my dreams the former would work like {{cite doi}}, producing a global, editable reference, and the latter would link both to the cite and separately (using the page number) to the appropriate page number in Google books.— Preceding unsigned comment added by Lfstevens (talk • contribs) 08:22, 24 September 2014‎

To which I replied "please can we avoid creating yet more citation templates? See small-font comment by PBS dated 19:06, 20 September 2014 at Wikipedia talk:Citing sources#Should we improve ref tags?" The same still applies: even more strongly, because that comment by PBS is in this very thread. --Redrose64 (talk) 08:48, 24 September 2014 (UTC)

The feature request for page numbers is Template:Bug - Page number attribute for <ref> tags. The discussion there has no description of how this should look in practice. When used with Citation Style 1 templates, I don't see how the Cite page number would be injected into the COinS metadata.

If there is a real push to improve cite.php, then page numbers would not be my first priority. We have a long list of bugs and feature requests. We have older bugs that keep getting rediscovered by new editors (and sometimes I forget and have to relearn why something is broken). There are several issues with error checking, the new Template:Agrl is messy, but the most egregious would be Template:Bug - Nested refs fail inside references block.

If we really want to improve cite.php, then my recommendation would be to get together a task force of editors who technically understand how it works. Then clean up the bug list (I've been working on this for a while), start fixing the broken issues, then work on feature requests. --  Gadget850 ^talk 09:29, 24 September 2014 (UTC)

the reflist at the bottom of the page could be integrated into the current layout quite easily by changing the layout of WP:APPENDIX to recommend that the notes/footnote section is placed as the last section on the page. I used to place a Footnotes section at the bottom of the page, but a discussion (if memory serves) among less than a handful of editors decided to place it above the References section simply because the majority of those editors liked the notes adjacent to references but also wanted further reading adjacent too. A big advantage of placing footnotes [lets call them endnotes] last is that all those footer templates can then also contain reference tags, which at the moment tend to be treated as an exception to WP:V for the convenience of the WP:FNNR layout. -- PBS (talk) 12:06, 24 September 2014 (UTC)

If you browse through Template:Agrl you will find that there are five related bug reports. And I was just using it as an example of broken tings we should fix before moving to improvements. --  Gadget850 ^talk 13:06, 24 September 2014 (UTC)

The fundamental problem with doing citation at WP is not that cite.php needs fixes, improvements, or (yikes) new features, but that we don't have a clear concept of what we are trying to do. ~ J. Johnson (JJ) (talk) 21:50, 24 September 2014 (UTC)

My later comments were in the spirit of thinking about the bigger picture. I understand that the real picture is bigger still (how indeed should we link content to sources...endnotes seem a bit dead tree- suitable for printing only, don't they? E.g., why not have the source appear when you hover over/touch the text to which it applies? That way it wouldn't matter where in the text a cite appeared. The hover/touch would produce the cite material, not the sfn smidgeon.) Thinking out loud. Lfstevens (talk) 08:35, 25 September 2014 (UTC)

@Lfstevens: They can already do exactly that. At Preferences → Gadgets, enable "^(D) Reference Tooltips: hover over inline citations to see reference information without moving away from the article text (does not work if "Navigation popups" is enabled above)". --Redrose64 (talk) 09:31, 25 September 2014 (UTC)

There is Template:Bug - Footnotes content should appear in a tooltip (add ReferenceTooltips to Cite as default/option). Code was added to cite.php for this but never enabled and recently removed. We now have the ReferenceTooltips gadget or the newer Preferences → Beta features → Hovercards. --  Gadget850 ^talk 11:25, 25 September 2014 (UTC)

The tooltip shows the sfn content, not the related cite content. Wrong! While we're at it, even though the cite has no page and the sfn does, if I click on the cite it should take me to the correct page. And if you have the option, there is no reason to add the cites to the bottom of the page. Lfstevens (talk) 12:09, 25 September 2014 (UTC)

I had thought that Lfstevens was talking about a way to find out which part of the Wikipedia article was supported by the citation. I believe that @Anthonyhcole: had arranged for such a system, but I don't remember whether he is testing it in an article at the moment. Whatamidoing (WMF) (talk) 19:21, 9 October 2014 (UTC)

Template:Bug --  Gadget850 ^talk 20:48, 9 October 2014 (UTC)

 Your <ref> block}} is the note (or footnote) containing comments, citations, etc. This is understood to support the "text/statement" it is placed next to (usually but not always following).

 The idea of the specificiation (such as a page number) of a short cite connecting to the specific passage in the source is interesting. And has been done in some cases. E.g., some of the IPCC cites in Global Warming#Citations link chapter, section, and even page numbers to the specific online text. But this is entirely dependent on such text being available, and can get sticky real fast if (say) you're trying to go to specific page in a pdf which first has to be downloaded. Note that this can be done readily enough using a standard external links following the short cite. There is absolutely no need to "improve" ref (or anything else) to do this. ~ J. Johnson (JJ) (talk) 23:20, 9 October 2014 (UTC)

Indicating exactly which bits of the text are supported by a citation

(Thanks for pinging me, above, WhatamIdoing.) User:Makyen has made this template ("Ref supports2") for article segments supported by more than one source, where each source supports different parts of the segment.

Example (hover your mouse pointer under the footnote marker):

At any given time, about half of all patients with malignant cancer are experiencing pain, and more than a third of those (and two thirds of all patients with advanced cancer) experience pain of such intensity that it adversely affects their sleep, mood, social relations and activities of daily living.   

References

Graham87 has pointed out that users of screen readers like JAWS will not benefit from this feature because it relies on tool tips, and JAWS doesn't read tool tips. --Anthonyhcole (talk · contribs · email) 22:45, 23 October 2014 (UTC)

Separating reference format from content

 I advocate moving the job of presenting references/notes from the editor to the page renderer. The editor should provide the source, and let the renderer decide how to present it. Logic:

• The idea of tasking the editor with formatting makes little sense in this era of multiple display platforms. E.g., what sense does
• Articles would look/behave more consistent/ly.
• Editor work would be reduced. E.g., the renderer could automatically create an alphabetized list of sources, with endnotes displayed separately.
• Endnotes could appear instead as page footnotes when an article is printed.
• Refs/notes could be placed on a separate tab instead of appearing with the article text (as could other "standard" sections, such as See also and External links.
• At some point, the display style could be made subject to a reader preference. E.g., a reader could decide to have notes adopt the ibid/opcit convention instead of name, year, page.
• Performance could also possibly be improved if "special" sections were fetched only when the appropriate tab was accessed.

Note, the details of these display choices are not the point, which is that separating the content from the format makes such details much easier to address. Feedback welcomed! Lfstevens (talk) 01:24, 11 October 2014 (UTC)

You are proposing a whole program of stuff, much of which is controversial. And rather beyond the current topic. I suggest that you propose these individually, in new sections. ~ J. Johnson (JJ) (talk) 21:59, 13 October 2014 (UTC)

Separated into a new section. The basic concept is what I'd love to see discussed. Once that is sorted, more detailed discussions can follow. Lfstevens (talk) 05:31, 15 October 2014 (UTC)

Lfstevens, I believe that the English Wiktionary is placing citations (quotations showing the word being used) with some kinds of references (such as sources to support etymology) on a separate tab. It doesn't seem to be a very popular method, though. Whatamidoing (WMF) (talk) 18:00, 14 October 2014 (UTC)

Thanks for the comments. Didn't quite understand "quotations showing the word being used". Lfstevens (talk) 05:31, 15 October 2014 (UTC)

Look at wikt:Citations:abiding-place to see an example. Whatamidoing (WMF) (talk) 17:19, 15 October 2014 (UTC)

• Template:Bug --  Gadget850 ^talk 17:32, 15 October 2014 (UTC)

Does references accept parameters?

I wonder if <references /> accepts any parameters like for example the font size. I don't know a better place to ask this question, so I asked it here. —  Ark25  (talk) 20:51, 15 October 2014 (UTC)

Only the group parameter. It is a parser tag not a HTML element. --  Gadget850 ^talk 20:58, 15 October 2014 (UTC)

Use {{reflist}}, it has far more flexibility than <references /> which I regard as obsolete. BTW I can't think of any legitimate reason to change the default font size of a reflist. Roger (Dodger67) (talk) 06:21, 26 October 2014 (UTC)

If you're not using any of the template's features, then all {{reflist}} does is slightly slow down the page (and make sure that you get an error should you copy it to another wiki, but few people do that). In fact, it would probably be better to replace all of the plain and/or empty {{reflist}} templates with the faster, native references tag. WhatamIdoing (talk) 18:04, 26 October 2014 (UTC)

{{reflist}}, with no parameters, is exactly the same as

<div class="reflist " style=" list-style-type: decimal;">
<references group=""></references></div>

The empty element <references group=""></references> is exactly equivalent to <references group="" /> and since that empty group"" attribute does nothing, it further reduces to <references />. The enclosing <div class="reflist " style=" list-style-type: decimal;">...</div> is served up virtually unchanged (one or two superfluous spaces are removed, but that's it); it is valid HTML 4.0, and browsers post IE3 just breeze through it. Any difference in speed between {{reflist}} and <references /> is minimal, because almost all of the processing is the construction of the ordered list that is served to your browser instead of <references />. --Redrose64 (talk) 19:12, 26 October 2014 (UTC)

Yes, it is indeed very minimal, after you've processed the transcluded template to determine that all you needed is the plain tags. For cached pages, it's probably so minimal as to be non-existent. For uncached pages (including every page sent to a logged-in user), and for all pages if anyone changes anything at all in the template since the last time the page was cached, it's probably measurable, but still small. That's why I said it's effect was only to "slightly" slow down the page. On a project with millions of pages, it is noticeable in aggregate, but I wouldn't expect editors with high-speed internet connections to notice any significant difference. WhatamIdoing (talk) 19:41, 26 October 2014 (UTC)

I can't prove it, but I expect that if people were to "replace all of the plain and/or empty {{reflist}} templates with the faster, native references tag" that itself would be counterproductive because of the way MediaWiki works: the whole page needs to be reparsed just for that small edit. --Redrose64 (talk) 20:13, 26 October 2014 (UTC)

And the change would cause an increase in the database size (I have no idea how much) due to the extra revision being stored. Comparatively, {{reflist}} has extremly low overhead. --  Gadget850 ^talk 22:19, 26 October 2014 (UTC)

As you both know, if the change is made when editing anyway, then it has zero such effects. WhatamIdoing (talk) 03:03, 27 October 2014 (UTC)

I don't think that advocating the replacement of {{reflist}} with <references /> is at all constructive or even useful. Any performance gains are negligible, and <references /> does nothing that {{reflist}} cannot also do. At one time, they had different font sizes ({{reflist}} was 90%, <references /> was 100%) but for a couple of years now, both have yielded 90%. --Redrose64 (talk) 10:57, 27 October 2014 (UTC)

Until the OP replies, I'm not sure we answered the question. --  Gadget850 ^talk 12:23, 27 October 2014 (UTC)

Legitimate use of fake sources

There has been a discussion in WikiProject Video games regarding the use of fake sources. It looks like this practice is acceptable in some situations and, according to some members, it has happened many times during nine years.[1][2][3]

I would like to know if this is common practice in other WikiProjects. Indeed, this is a powerful tool that could improve the whole Wikipedia. This methodology should be explained in this content guideline for the benefit of future editors. 84.127.82.127 (talk) 09:36, 25 October 2014 (UTC)

No way. If a "source" is fake, it isn't a source, and WP:V is violated, probably WP:NOR also. --Redrose64 (talk) 10:01, 25 October 2014 (UTC)

The procedure seems to be to figure what probably appeared in a magazine article by reading quotes from talk groups, and creating a citation that makes it appear the editor who adds the "source" actually read the source. This is dishonest and I would have no qualms about supporting a community ban for the editor who does this. Jc3s5h (talk) 14:42, 25 October 2014 (UTC)

I would appreciate anyone pointing out concrete evidence of any user (including JimmyBlackwing) actually doing this. The subsequent blocks will be immediate and indefinite.—Kww(talk) 15:06, 25 October 2014 (UTC)

OK this is silly. I'm mystified by 84.127.82.127's claim subsequent to the above-linked discussion that "It looks like this practice is acceptable in some situations. He's been directed multiple times to the guidelines which sit at WP:VG/RS. If there is anything in the guidelines that even hints that this practice is acceptable please point it out and it will be changed immediately. That goes for any of you. Likewise it has been explained to 84.127.82.127 that WP:VG/RS is controlled by WP:RS and WP:V. 84.127.82.127 is complaining about a user essay. And if we look at the words of the user essay we see this: "This underhanded method thoroughly violates some Wikipedia policy or other" (emphasis added for the easily confused). Why anybody would imagine that this practice is acceptable especially after he's been told that it isn't goes beyond my understanding. JimmyBlackwing might want to shift the "Fake It" section below the "Google it" section and into the "terrible, horrible" section, but that's entirely up to him. -Thibbs (talk) 16:43, 25 October 2014 (UTC)

Seconded. The IP seems to be having a hard time distinguishing a user's personal essay subpage (which doesn't even advocate it, but rather says he suspects people do it) from the Wikiprojects actual stance. None of the WikiProjects literature, or editors, are actually advocating it. Judging by the IPs edit history, I believe it just enjoys stirring up trouble. Sergecross73 msg me 18:13, 25 October 2014 (UTC)

Okay, just to clarify: This is not about using actually fake sources. That's a misleading descriptor I wish hadn't been affixed to this issue. This is about using sources when you don't actually have access to them, but you can figure out what's said based on other sources - for example, if there's a video game review you don't have access to, but Metacritic summarizes it, you can extract the appropriate details from Metacritic and cite it as though you have the actual source. This may sound underhanded, and it's definitely not ideal, but consider the case of citing one scientific paper referenced by another scientific paper when all you've got is the second one - that's gotta be done constantly and without issue. The sources are perfectly legitimate; it's more that the access to them isn't. That's all. Tezero (talk) 18:11, 25 October 2014 (UTC)

No, Tezero, when you cite something, it means that you, the author of the text, have personally examined the material and can assure the reader that it supports the material in the article. Anything short of that is fraud.—Kww(talk) 18:29, 25 October 2014 (UTC)

That's dumb. There's nothing preventing people from seeing the real McCoy, deciding they don't like it, and lying about it (or just embellishing the facts) because no one'll ever know, but we're focusing on those who just don't want to shell out a hundred dollars for an old French magazine through a sketchy Russian website, so they place their trust in a review aggregator or a helpful forum community? Calling that a respectable sourcing policy is the real fraud. Tezero (talk) 18:37, 25 October 2014 (UTC)

Occasionally the author of an article in a respectable journal will not have access to an obscure reliable source, but will have access to a summary of it in some more accessible reliable source. In this case, the author would cite the accessible reliable source, and explain that the accessible reliable source summarized the obscure reliable source. Your user profile states you are a college student. I suggest you look up in the style manual required by your professors how to do this, otherwise you are in danger of getting low grades on your papers, or perhaps more serious penalties. (If you college lets you get away with this, perhaps you should find a better college.) Jc3s5h (talk) 18:52, 25 October 2014 (UTC)

Your user profile states that you are under no obligation to obey Wikipedia's verifiability policies that include a distinction between first- and third-party information. Are you really in a position to be didactic about Wikipedia policies? Tezero (talk) 19:29, 25 October 2014 (UTC)

This IP user has a history of troublemaking, and he appears to be continuing it here. The essay section discusses an undesirable sourcing practice that some editors, in WP:IAR fashion, have used in an attempt to improve articles. There is no malicious motive behind it, and it is used, from what I've seen, only for information that would otherwise be inaccessible. Should they be blocked for thinking they'd made a necessary, justifiable exception to the WP:SAYWHEREYOUGOTIT guideline? Depends on whether the administrators here want most of Wikipedia to be blocked. JimmyBlackwing (talk) 19:40, 25 October 2014 (UTC)

No, they would be blocked for falsifying sources. That they would consider falsifying sources justifiable is simply a sign that they are incapable of making honest and competent contributions to the project.—Kww(talk) 23:35, 25 October 2014 (UTC)

Consider the alternatives, though: either you don't include the information at all, you leave it unsourced, or you attribute it to an unreliable source. Not only could it be of benefit to include the information by falsifying your access to a source (as opposed to one of these options); it could be the honest thing to do, if leaving the information out would contribute to WP:UNDUE or otherwise be misleading. Tezero (talk) 23:39, 25 October 2014 (UTC)

No, Tezero, it is simply fraud. "Don't include the information at all" is a quite viable option.—Kww(talk) 23:52, 25 October 2014 (UTC)

No sources are "falsified", whatever that means. The contents of an unavailable source are reconstructed indirectly, and that reconstructed document is then itself treated as a source—a standard technique in off-Wikipedia research. Hardly "fraud". The only guideline violated is WP:SAYWHEREYOUGOTIT, because the many steps involved in reconstructing the source are not included in the citation template. JimmyBlackwing (talk) 00:20, 26 October 2014 (UTC)

Yes, JimmyBlackwing, editors that do that are falsifying sources. No doubt about it: the editor is using a forum as a source, but pretending that the information has been taken from a reliable source. That the forum may have correctly reported on the contents of the source does not make the forum reliable, and representing the original as the source when the editor is actually relying on the forum is lying.—Kww(talk) 05:07, 26 October 2014 (UTC)

The debate isn't over whether this practice is technically untruthful, but over whether that's a good enough reason to forbid it given the advantages. And for what it's worth, the intermediary source is often more than just a forum, but an actual reliable source - e.g. Metacritic or Rotten Tomatoes - that just doesn't happen to link to the full version of the desired source. In this case, trust is being placed in the reliable source - which we're already doing by definition when we cite a source classed as reliable for anything. Tezero (talk) 05:17, 26 October 2014 (UTC)

Kww, I'm not defending their rationales on this matter, but I think that they mean something similar to a WP:SOURCEACCESS case; in a WP:SOURCEACCESS case, a Wikipedia editor might not have the source and might get assistance from another Wikipedia editor who relays to them the contents of the source. That has happened to me more than once on Wikipedia, where I trusted a fellow Wikipedia editor's word on the matter. And it happens in WP:Good article and WP:Featured article cases all over Wikipedia, where, for example, a reviewer might not have access to the source and is trusting the nominator to report on the sources accurately if asked about them. I know that the forum instance is not exactly the same as either of those WP:SOURCEACCESS examples, but it's similar. Flyer22 (talk) 05:20, 26 October 2014 (UTC)

I'm interested to know Kww's view on the magazine scans obtained from non-editors, such as those found on enthusiast sites like SonicRetro and Amiga Magazine Rack. Such scans are used regularly in VG FAs. It seems that his interpretation of SAYWHEREYOUGOTIT invalidates them as sources. In any case, to suggest that editors are lying when they reconstruct sources is ridiculous. I suppose scholars who cite Socrates or the Q source are lying as well. Indeed, all historians work inductively to rebuild events and artifacts—it's absurd to say that the conclusions of their induction are lies just because they're uncertain. An editor who reconstructs a source may potentially be accused of original research, of using less-than-reliable informants and of violating SAYWHEREYOUGOTIT. However, accusing them of fraud is laughable. They believe that they've rebuilt an unavailable document such that it has become usable and citable again, like Q. The question is whether the practice fits through Wikipedia's narrow verifiability and reliability guidelines, not whether the editors are being honest or engaging in valid research. JimmyBlackwing (talk) 06:11, 26 October 2014 (UTC)

Indeed, the basis for the series of complaints 84.127.80.114/84.127.82.127 has made against JimmyBlackwing's essay appears to be related to the counterproductive efforts of 84.127.80.114 to remove all offline sources the editor has been unable to locate at Ghost in the Shell (video game) (where he has taken over the GA/FA-directed work of topic banned editor Lucia Black) as well as this editor's misguided efforts to link to copyright violating digital scans of the few hardcopy magazines he has located online. -Thibbs (talk) 12:26, 26 October 2014 (UTC)

Yes, Jimmy, you need to examine actual sources, and accusing people that participate in such behaviour as participating in fraud is not laughable, simply accurate. Your analogies are strained, as no legitimate that "reconstructed" something would pretend that his conclusions came from anything but the reconstruction. It's not that our verifiability and reliability guidelines are narrow: they track reputable academia. As for an FA that uses a copyright violation as a source? It should be immediately delisted.—Kww(talk) 14:37, 26 October 2014 (UTC)

You're not being rational. If a source violates copyright, that source can be removed, just like if there's any other problem with it. FAs aren't immediately delisted even for extensive sourcing problems; one source containing a copyright violation - even several - wouldn't be cause for immediate delisting. But with the Ghost in the Shell article, the scans are the only copyvio element; removing the scans allows the sources to conform to Wikipedia policy for both verifiability and respect of copyright instead of only verifiability. This is all rather off-topic, though; using sources you don't have access to - regardless of how ethical it is - isn't a copyright issue, but a verifiability one. Tezero (talk) 20:12, 26 October 2014 (UTC)

@Kww: Feel free to delist Flight Unlimited III, Secret of Mana, Thief II: The Metal Age, Lost Luggage, Flight Unlimited II, Mischief Makers, Drakengard, Anachronox, Sega CD, Sega Genesis and most of WPVG's other FAs. All of them use "copyright violations" as sources. A good chunk of my essay directs editors to "copyright violations": AMR, SMS Power, Digital Press Library, CGW Museum, DLH's Commodore Archive, Retromags and Replacementdocs. It explains how to obtain "copyright violation" scans even beyond those "copyright violation" websites. Go ahead, Kww—let the purge begin. JimmyBlackwing (talk) 21:07, 26 October 2014 (UTC)

Yeah, Sonic Retro, which was mentioned earlier, I've used in probably half of my GAs and FAs for the scans it provides. In fact, earlier today I used it to look up a scan of the manual of Sonic Riders: Zero Gravity to source a fact in the Knuckles the Echidna article, because I don't own that game, only the original Riders. Should I have ponied up $20 or whatever on Amazon and waited a few days just to confirm with my own eyes that the scan was legitimate? (Of course, I'd have the game, too, but that's beside the point.) Of course not; it's ridiculously unlikely that the scan was fabricated. Not only would creating a convincing fake be difficult; there'd be no reason to. Really, this is no different from going to the library and looking at their microfilm, or watching a YouTube upload of a TV documentary. Tezero (talk) 23:53, 26 October 2014 (UTC)

We're not talking about a concrete case here and this isn't something worth getting riled up about in the abstract. Yes sometimes there are serious abuses such as when users fabricate sources out of thin air to bolster a personal opinion. There are also times when it is clear that an unavailable source is actually saying something despite one's lack of access to it:

• e.g. when a Google Books preview only shows the table of contents and you see chapter 4 is entitled "Criticism from Japan" and chapter 5 is entitled "Criticism from Europe" then it's no huge crime to claim that "the topic saw criticism in both Japan and Europe" and then to cite Chapters 4 and 5 of the book instead of the table of contents despite not having read the chapters.
• e.g. when a now-inactive wikipedia editor added a set of claims cited with a malformed ref to an offline source a while ago then you should be banned when in trying to repair it today you carefully determine the date and issue number yourself from workable non-RS sources despite not having access to the original.
• e.g. when a fellow editor sends you a transcript of an article you need then it's not worth anyone's time dragging you to ANI if you take the transcript as accurate and cite the article despite not holding the magazine in your own fingers.

There are varying degrees of "faking it" and most reasonable editors know where the limits are. As much as I like the reliable if dogmatic "The rules are the rules, end of story - you bend them and you're banned" approach, I don't think this is a case where immediate indef blocking would usually be a more appropriate way to start than say a warning. Again the most we should be saying in the abstract is that this is an underhanded method that violates Wikipedia's policies. We don't need to add such a warning to the guidelines as suggested by 84.127.80.114 at WT:VG, we don't need to suggest that this is a powerful tool to improve Wikipedia or explain the methodology in the guidelines as suggested by 84.127.82.127 just above, and we don't need to change any userspace essays unless there is evidence that their currently included warnings are insufficient to prevent actual problems. -Thibbs (talk) 15:15, 26 October 2014 (UTC)

Yes, this is something that requires using judgement after considering all the facts and circumstances. We've had discussions about this issue before, most memorably with silly claims like, "If you read it at Google Books, then you have to say that, because Google Books might fake the contents!" The general rule is that accurate copies—defined as any copy that a typical editor would believe is unadulterated—are acceptable, regardless of provenance.

Also, the goal is to write the encyclopedia, not to punish people for their sins. If someone is making a mistake, then we need to educate them and clean it up, not punish them. WhatamIdoing (talk) 18:37, 26 October 2014 (UTC)

WhatamIdoing, surely you don't endorse forging citations based on input from forums. There's an enormous difference between Google Books and a fan-forum for video games.—Kww(talk) 00:29, 27 October 2014 (UTC)

I see zero (non-hysterical) claims that anyone is forging anything. I see a claim that some people have looked at apparently accurate and authentic copies of articles—copies whose only fault is that the poster had no right to post a copy of the article—and cited the reliable source that they actually read, which is the newspaper article. We should not accept a link to the copyvio, but the editor can (honestly, ethically, legitimately) claim to have read the source—because he did read the source!

Perhaps an example would be clearer:

• On Monday, you buy a copy of The Times, and you read an article in it. You cite the article that you read. Everything's good, right?
• On Tuesday, you hear that The Times had another related story. You are unable to buy a copy of Tuesday's paper. However, you look online, and their website gives some information and the start of the story. What you're interested in is not shown. You go to your favorite web search engine and look around. You find what purports to be and appears to be multiple exact copy-and-paste copies of that story. You read one of them. You find the information you're looking for in it. A few minutes later, you look in your e-mail inbox, and you find that a friend who has a subscription has just e-mailed yet another copy of this article to you.
  You want to update the article. Do you:
  1. cite the newspaper article in The Times that you actually did read?
  2. cite "email from my friend", because even though what you read was read the newspaper article from The Times, you didn't read a legal copy of the newspaper article?

I'm thinking that option #2 is silly. If you read the newspaper article, even if the copy you read was not an original, licensed copy, then you should cite the newspaper article. WhatamIdoing (talk) 03:15, 27 October 2014 (UTC)

I'm sorry that you think being honest about the source of your material is silly, WhatamIdoing. I will also point out that the material in the essay under discussion includes "basically, you cite a source that you don't have by cobbling together a citation from quasi-reliable snippets of information on the Internet" as a citation technique.—Kww(talk) 04:48, 27 October 2014 (UTC)

Kww, if you have actually read the reliable source yourself, then I see nothing "dishonest" about citing the source that you actually read yourself. I do not think that it is "dishonest" to fail to say "Oh, and by the way, the accurate and true copy of this source that I read was sent to me in e-mail" (or borrowed from a library, or read in a bookstore, or read on a web forum, or however else you might have received a copy). I'm not responding to some essay; I'm responding to what I wrote: accurate copies are acceptable. If you have read an accurate copy, then you have read the source and you should cite the source. Nothing that I wrote about "accurate copies" says or even implies that the editor has cobbled together bits and pieces or guesses from hearsay. WhatamIdoing (talk) 19:49, 27 October 2014 (UTC)

Seconded. So long as an editor reads the source and cites to it appropriately when authoring material, verifiability has been satisfied; how the editor obtains the source should have no bearing on whether the source can be cited. Obviously, editors who violate copyright by illegally duplicating sources can be prosecuted for doing so, but that's a separate issue from reading such a copy and then citing to the source. In such a scenario, the citation to the source is honest, even if the means of obtaining the source aren't. Of course, Wikipedia shouldn't provide a link to any illegal copy, but that's another separate issue. –Prototime (talk · contribs) 20:33, 27 October 2014 (UTC)

Again there is a spectrum of examples of this technique. Some uses really shouldn't be a problem. Others are probably good places for a warning. And yes some examples would be block/ban-worthy offenses. I appreciate your call for honesty in sourcing, Kww, but this is a tempest in a teapot. Most established editors know how to stay on the right side of the line and - let's be completely frank - those who do step over the line are often quite capable of disguising their transgressions. That's not to say that they are excused, or that such transgressions should be promoted, but that is the reality. Ultimately the only things we can control are the official guidelines (none of which currently support or link to this technique), and enforcement of specific incidents falling within the spectrum of "faking it" that cross the line. Should the essay be reworded? Honestly I would have worded it differently myself, but I doubt there will be any problem because it seems to give clear disclaimers regarding the utility of "faking it" (i.e. "This underhanded method thoroughly violates some Wikipedia policy").

The concerns shared by 84.127.82.127 appear to be based on serious confusion over the relations between policy, guideline, and userspace essay, and not over any specific violations he is seeing. They also coincide immediately with this editor's misguided efforts to strip out all sources he cannot personally verify from an article he is working on. From some of his statements there (i.e. "I have something to say and I will in due time"), it appears he is waiting for the outcome of this discussion to provide justification for his removal of a source that he can personally only locate in the form of a scan hosted at a situationally reliable website. -Thibbs (talk) 12:05, 27 October 2014 (UTC)

This seems to be partly about what is sometimes called "citation plagiarism", which was an issue in the Dershowitz–Finkelstein affair. --Boson (talk) 23:34, 26 October 2014 (UTC)

Correct me if I'm wrong, but citation plagiarism occurs when an original source cites a second source, and then a third person puts that citation into their own work without checking the second source. That's altogether different from citing to a source after reading an illegally-duplicated copy. –Prototime (talk · contribs) 20:33, 27 October 2014 (UTC)

Agreed, but some of the discussion above seemed (to me) to be suggesting that only directly reading the original source justifies citing (only) the original source (which I tend to agree with, at least in principle). An even more venial "sin" would be failing to cite the intermediate source when you have read the original source (but are possibly using it in a way that fails to credit the intermediate source with some insight that caused you to cite the original source in the first place). --Boson (talk) 00:15, 28 October 2014 (UTC)

Kww asked for concrete evidence. I have located sources that are damaging the article. Regarding the first one, Next Generation review, I have found what may be a fake page without a page number. I cannot ask Lucia Black, the editor who added the source, because she is topic banned. The editor that allegedly owns the issue 35, JimmyBlackwing, has decided to take a long wikibreak. Thibbs insists that page numbers are not required and that I should be the one providing the page number. Thibbs has collapsed this part of the discussion.

I would appreciate Kww's guidance; for instance, I would like to talk with Lucia Black about these sources. 84.127.82.127 (talk) 08:30, 29 October 2014 (UTC)

@84.127.82.127: The concrete evidence Kww spoke of was in regard to editors who had faked sources. Are you accusing Lucia Black of faking this source? What leads you to the idea that the source has been faked? The scan of the article that you located seems to corroborate the claims in the Wikipedia article. Why do you believe it "may be a fake page"? Is the missing page number the only reason you wish to exclude the reference? Perhaps you could make a request for the missing page number at WT:VG or WP:REREQ. -Thibbs (talk) 13:00, 29 October 2014 (UTC)

I see. This is not an article for me, but for everyone. If the community wants articles with fake sources, I will respect that, but I need to make sure this is what the community wants. 84.127.82.127 (talk) 02:39, 4 November 2014 (UTC)

Guidelines for citing foreign language sources

When translating a foreign-language Wikipedia page to create or expand a corresponding page in the English-language Wikipedia, how to style the citations? List the original (romanized if from a non-latin alphabet) and provide a parenthetical translation of the title only, or what? -- Deborahjay (talk) 23:32, 31 October 2014 (UTC)

Reference 24 is an example. 84.127.82.127 (talk) 20:33, 1 November 2014 (UTC)

Deborahjay, your plan sounds good. The rule is to do something that you think is sensible for that article, and then not to worry too much about it. If you want, you may use the optional templates in Category:Language icon templates like {{ru icon}} that let you tag the language of the original source, but it's neither encouraged nor discouraged. Do what you think is best. WhatamIdoing (talk) 15:48, 2 November 2014 (UTC)

In Citation Style 1, a foreign language citation title can have up to three components.

For languages that use Latin-based writing systems, |title= is the title of the resource as it appears in the original resource. |trans-title= holds an English language translation of the value in |title=.

|title=Jacques Lapin |trans-title=Jack Rabbit

For languages with writing systems that are not based on the Latin alphabet, Arabic, Chinese, Greek, Hebrew, Russian, Thai, etc), CS1 has an additional parameter. |script-title= holds the resource's title in its native writing system. Unique to this parameter is the language prefix, an ISO639-1 two-character language identifier followed by a colon. The identifier is used to help browsers correctly display the non-Latin text. When |script-title= is used, |title= holds a transliteration of the title in |script-title=. There are often strict rules that govern transliteration of a non-Latin script. As with the previous example, |trans-title= holds an English-language translation of the title in |script-title=.

|title=Tōkyō tawā |script-title=ja:東京タワー |trans-title=Tokyo Tower

It is helpful to readers if you include the language of the resource in the CS1 template. For that, use |language= where the value assigned to the parameter should be either the language name or the two-character code from the List of ISO 639-1 codes.

|language=cy or |language=Welsh

Do not use icon templates within CS1 templates.

|language={{cy icon}}

—Trappist the monk (talk) 16:31, 2 November 2014 (UTC)