Wikipedia talk:Citing sources/Archive 37

Archive 30

←

Archive 35

Archive 36

Archive 37

Archive 38

Archive 39

Archive 40

ISBN oddity

An ISBN is meant to uniquely identify a book as a product, including edition and format (e.g. EPUB, Mobi, PDF, paperback etc.) This does not seem to be true. For example, ISBN 9781853260247 (ISBN 185326024X) is used by the publisher Wordsworth Editions for the 1850 Dicken's classic David Copperfield:

• Google Books shows the publication year as 1992, 837 pages. But the cover inside the Google Books version is different from the snapshot in the Google Books description.
• Amazon shows the same title and publisher with yet another cover, dated 5 August 1997, 768 pages.
• Goodreads shows it with the same cover as Amazon, but 1 April 1998, 750 pages
• BetterWorldBooks, an online retailer, is showing the same ISBNs, 768 pages, 1 April 1998, and yet another cover.

Looking inside the Google Books version, we find a different cover from the Amazon version but otherwise apparently the same content. They are both showing a 2000 publication year. The (partial) publication history, always using the same ISBNs and never identifying an edition number, seems to be:

• 1992 original edition 837 pages
• 1997 5 August fresh edition
• 1998 1 April fresh edition
• 1999 fresh edition with new introduction and notes added
• 2000 fresh edition with illustrations added 768 pages

Google and Amazon are both showing scans from the 2000 editions, which apparently differ only in the cover, but Google is giving 1992 metadata and Amazon is giving a mix of 1997 and 2000 metadata. The 1997 and 1998 editions are only eight months apart, so this publisher could easily put out two different editions with the same ISBN and different pagination in the same year. Is there an impact on this guideline? A warning of some sort? Some other way to identify different versions? Aymatth2 (talk) 14:38, 28 September 2013 (UTC)

What should we warn about — that the world isn't perfect? If a publisher prints another batch of books, changing only the cover, then it's not really a different edition (which implies editing), but a different printing, and trivial. Changing the pagination is certainly not trivial, and if some publisher does that without troubling to use a new ISBN — well, that is a specific problem. But the impact is limited, and does not change the usefulness of ISBNs generally. That such anomalies occur is why references should include comprehensive bibliographic detail. ~ J. Johnson (JJ) (talk) 20:37, 28 September 2013 (UTC)

I've found these oddities quite a bit, JJ, which is why I don't add ISBNs to citations. Adding name, title, publisher, year of publication should be enough to find an edition. If a hardback and paperback were published in the same year, and it's known that the page numbers changed, adding (paperback edition) after the year of publication is all that's needed. SlimVirgin ^(talk) 20:48, 28 September 2013 (UTC)

Sure, there seem to be some cases where an ISBN has been screwed up. But it is quite an overreaction to reject ISBNs totally, considering how useful they are generally. ~ J. Johnson (JJ) (talk) 20:42, 29 September 2013 (UTC)

The question is far from hypothetical, assuming Wordsworth Editions is indeed reusing ISBNs for different editions. Their titles are cited in Wikipedia hundreds of times. I think it is fair to say that once a title has been published, the ISBN will not be used by another publisher or for another title. So a minimalist who is confident the book has not really changed since first publication in 1850 could use ISBN in place of publisher, title and author, giving a cite like

• "I Am Born". ISBN 185326024X. {{cite book}}: Missing or empty |title= (help)

This avoids the pagination problem by identifying the chapter, which presumably is present in all editions, so the publication date is not needed. I am more of a maximalist, so would be inclined to give all the information I have, as

• Dickens, Charles (2000) [1850]. "I Am Born". David Copperfield. Wordsworth Editions. p. 5. ISBN 978-1-85326-024-7. Retrieved 2013-09-29.

I can't see any harm in giving more information than is needed. The sort of reader who examines the citation details at the back of an article probably wants more rather than less. But I am still stuck on how best to cite the introduction to this book written by Adrienne E. Gavin. There could be two editions of the book, both published in 1999, one with the introduction and one without. How do I tell the reader which edition I am citing? Or does it matter? Aymatth2 (talk) 11:52, 29 September 2013 (UTC)

The question of citing an itroduction has recently been raised at Help talk:Citation Style 1#Citing book intorductions (sic).

—Trappist the monk (talk) 13:17, 29 September 2013 (UTC)

The concern I have is not how to format the citation of an introduction, although I agree that is not entirely obvious and some guidance should be given somewhere. I am more puzzled about how to identify the print edition in which the introduction was found when a reader could have a copy of the book with the same year and ISBN, but with no introduction. Not a very serious issue, obviously, but maybe there is a simple answer. Aymatth2 (talk) 14:45, 29 September 2013 (UTC)

This happens occasionally with sport coaching books, or at least it does in snooker. The same coaching book gets reprinted but with a fresh introduction (read "endorsement") by the new world champion: same ISBN, same edition number etc. If you suspect this is the case, it's best to be on the safe side and clarify with something like: edition=2 (with a new introduction by Adrienne E. Gavin). Betty Logan (talk) 22:32, 1 October 2013 (UTC)

Yes. ~ J. Johnson (JJ) (talk) 21:36, 2 October 2013 (UTC)

It is because no means of identification is perfect that we provide as much bibliographic detail as possible. Most of the time there is no problem. Then there are a few (very few, even rare) extreme "oddities". E.g., I have heard of publishers correcting embarassing errors in already printed and bound books by cutting out and replacing individual pages. So it is conceivable that two editors might have books that are exactly identical in ALL respects except for one word, and the only way of distinguishing them is (say) a close physical examination for a glued in ("tipped") page. Or perhaps an introduction was left out of the first hundred copies. Well, if one editor finds an introduction, and another doesn't, and careful (and collegial!) investigation shows the items to be otherwise entirely and undistinguishably identical, then you have an extreme case which cannot be idenftiied except by the difference or omission itself. In such a case I would simply add a note warning of the difference. E.g.: "Some copies of this book lack the Introduction." (And perhaps check out rare book prices?) ~ J. Johnson (JJ) (talk) 21:31, 29 September 2013 (UTC)

The editor making the cite is almost certainly unaware that the edition they are citing is not the only one for that year and ISBN. The incorrect Google Books metadata in this example does not help at all. I have no idea how common it is for publishers to re-use ISBNs for different editions. The example given is of a static work, but if the publisher of a dynamic work like a travel guide did this, with repagination between editions, it could cause serious confusion. I would not assume this is rare. The trend is towards electronic-only publications, often with no ISBN, so the problem is likely to get much worse.

I have no solution, beyond perhaps adding something to the guideline like "provide as much information as possible to help others who want to find the exact version of the work you are citing." This is not specific to books, but could go up front somewhere. And maybe something like "when viewing a book online using Google Books, if possible take the publication date, edition and other information from a preview of the relevant front-matter page. Use the bibliographic information provided by Google only when no preview is available." Aymatth2 (talk) 01:26, 30 September 2013 (UTC)

It's generally a good idea to include a small relevant quote in book citations (usually less than a full sentence), especially in-line citations. This not only provides verifiability that the source backs up what it claims to back up (or that the editor is lying through his teeth, sigh) but it also gives those with access to electronic copies of the book something quick to search on, even if their copy has different page numbers. Such searches may also trigger hits in "Google Books" even if only a snippet preview is available or even if no preview is available at all. In short: "|quote=" is your friend. davidwr/(talk)/(contribs) 02:44, 30 September 2013 (UTC)

It's a good idea to do this. As David says, it gives the reader some words to search for and that will pin down the edition if there's doubt about it. SlimVirgin ^(talk) 03:06, 30 September 2013 (UTC)

It seems that there is always doubt about the edition, including with books that have ISBNs. Using "|quote=" would not work for me. I often start articles that are almost entirely based on books, using information from several pages in each book. See Lake Uniamési for an example. I can't see sticking a quote or quotes into each source definition. I think quotes from copyrighted works should only be used for legitimate purposes, like critical commentary. Anyway, suppose I add a quote to a citation, as in

• Kenya eBook edition. Lonely Planet. 2013. p. 102. ISBN 978-1-74220-355-3. Retrieved 2013-07-30. The downtown area [of Nairobi] is unthreatening and comparatively trouble-free...

A reader searches their copy of the book, or searches the Google Books copy to which I have provided a convenience link, and does not find that quote. Am I lying about what I found in the book? Or did Google replace an older version with a newer one, same year and ISBN, that omitted the quote? Aymatth2 (talk) 13:39, 30 September 2013 (UTC)

In the rare cases where an ISBN covers more than one edition it is likely, as you said earlier, that the first editor is unaware of this (hypothetical) other edition. That comes out when a second editor finds a discrepancy with his edition. And (hopefully) does NOT immediately go into "liar, liar, pants on fire!" mode, but assumes good faith and collegially works with the first editor to resolve the discrepancy. Time enough to sort out such problems when they actually appear. Hopefully the first editor also specified where in the source the material is located. ~ J. Johnson (JJ) (talk) 21:26, 30 September 2013 (UTC)

I am not convinced it is all that rare. Agencies that issue ISBNs would like publishers to buy a new one for every little change, but publishers and booksellers probably only want a new ISBN when it really is a new product. That is something the publisher will decide. I don't think we should count on ISBN too much as a unique identifier. It would be nice if it was, but people break rules and the world is changing. See "Book-keeping". The Economist. 2 March 2013. If citations give as much information as possible it is more likely that the exact source will be found, or that someone checking will realize they may not be looking at the same version as the editor, but sometimes there will just be no way to distinguish versions. We will indeed have to assume good faith. Aymatth2 (talk) 14:15, 1 October 2013 (UTC)

Not rare? You only suspect so; you have no evidence that ISBN "oddities" are more than vanishingly rare. Yes, discrepancies happen, which is why we should never be absolutely certain of uniqueness, but that is the way to bet. (In the event of a discrepancy: AGF! and investigate.) The rare exceptions do not warrant any warning that ISBNs are not absolutely perfect. ~ J. Johnson (JJ) (talk) 20:10, 1 October 2013 (UTC)

JJ, I don't have evidence either, because I've never thought to gather any, but ISBN mistakes are certainly not vanishingly rare. I used to encounter them regularly (when I paid attention to ISBNs), although not frequently. Apologies that I can't be any more precise.

Can you say when it is helpful to add them to citations? I often have to borrow and re-borrow books as sources, and the same edition doesn't always arrive – with inter-library loans I take whatever is offered for the sake of speed – so I sometimes end up using different editions. That's especially likely on an article I've been writing over a long period. Usually the page numbers haven't changed, so it doesn't matter. If they have, I try to guess which edition most readers will have access to, and I adjust the page numbers accordingly. But then, months or years later, another editor will add ISBNs to the citations without making sure they match the edition that was used. That isn't helpful. The only reason it isn't harmful is that readers almost certainly never pay attention to ISBNs. So I'd be interested to see examples of when ISBNs do add information that would help a reader or editor. SlimVirgin ^(talk) 20:31, 1 October 2013 (UTC)

• I see no evidence that ISBN oddities are unusual. There is clearly a possibility that a high-volume publisher like Wordsworth Editions may not always use new ISBNs for minor changes like adding an introduction or repaginating. When I searched for ISBNs that did not match editions the majority of results said firmly that this cannot be. But a scattering of results, like distant laughter, gave examples:

• With Steve Biddulph's ISBN 9780007153695 Raising Boys: Why Boys are Different - and How to Help Them Become Happy and Well-Balanced Men the cover shown by Google is interesting but not relevant, so I will not mention it. Amazon gives it 224 pages, Google gives it 199 pages, both 2003.
• Goodreads has The Jane Austen Book Club as ISBN 0670915580, 288 pages, published 2005 by Penguin, first published January 1st 2004. Amazon has the same book as 304 pages, Viking, October 7, 2004.
• Google gives ISBN 000647988X A Game of Thrones 835 pages, 1996, while the Amazon entry gives it 864 pages, January 6, 2003.
• Anyone interested in citation styles should be careful with the Publication Manual of the American Psychological Association, Sixth Edition. There were numerous minor errors in the first printing, corrected in the second printing. But the ISBN is the same.

I do not take the extreme position that ISBN has no value because it is not always precise. The ISBN is one bit of information about the source. Anything else that can be supplied is more information. I am in favor of encouraging editors to supply all the information they have. The combination may be enough to uniquely identify the source. It may not, but there is no harm trying. Aymatth2 (talk) 01:00, 2 October 2013 (UTC)

  SV: I certainly agree that ISBNs are not always useful, for the reasons you cite. But they do provide linkage to the most detailed and authoritative publication information about a book (including where it is available). Why should that be disdained? Sure, some publishers are lax. But I doubt if (e.g.) pagination would be changed without other changes to the content. Such changes or revisions might be reflected in the ISBN. If not, well, they are even less likely to be noted in any other identifier. And providing an identifier that does not perfectly resolve all variations does not do any harm, unless there is some expectation that they really are perfect.

  The problem of subsequent editors adding the ISBN of an edition different from what the original editor consulted is vexing. But I would blame the original editor for not fully specifying his or her source. Where multiple versions of a source are consulted I would cite each version, noting different pagination or such.

  BTW, in these example of books with different page counts — do these really have different pagination? Or do Amazon, etc., perhaps just differ on how many pages they count? ~ J. Johnson (JJ) (talk) 21:41, 2 October 2013 (UTC)

Pagination is easy to change when the book is mostly pure text, as with the Wordsworth Editions oddity that started this thread. Wordsworth seems to have changed pagination from the first edition. They could argue that since it is still a paperback edition of the same classic book, same text, sold at the same low price, it is not really a different product. The consumer is unlikely to complain that they did not get what they were led to believe they were getting. With the same example, Goodreads seems to have taken page count as the last page number in the book (750), while Amazon included front matter and introduction (768). I think these are the same edition. The 1992, 837 page edition before the introduction was added seems to have really been different.

The people updating the catalogs may make mistakes. Google obviously does. I smell a rat with the Penguin/Viking Jane Austen Book Club example. Librarians make mistakes. Publishers make mistakes. Editors make mistakes. I confess to having happily used the Google Books metadata from the Wikipedia citation tool for Google Books for years, only checking the metadata inside the book when something seemed obviously wrong. I must have put in a lot of incorrect metadata, but if someone checks a citation I added, following the convenience link I supplied, they should find the information I cited, unless my eyes got crossed for a minute. Editors citing books almost always try to report what they found, and to give an accurate citation. But it might be sensible to say somewhere "give as much information as you can in a citation". That should be the subject of a separate debate. Aymatth2 (talk) 01:23, 3 October 2013 (UTC)

I am begining to think these "oddities" are (for the most part?) little more than confusion of page counts and editions. E.g., the Google Books link at the top cites the Wordsworth 1992 edition, but the preview clearly indicates it is of the "1992 and 2000" edition, including an introductory essay copyright 1999. I suspect the pagination (past the prefatory material) is identical, and the different page counts due entirely to the introductory material. Similarly, The Goodreads link says published "April 1st 1998 by Wordsworth Classics", and even provides both ISBNs. But these apply only to the source of the text (Wordsworth), not to the Goodreads version. I suspect the ePub/Kindle/PDF downloadables differ in format and pagination, but this is quite beside the point: it appears they don't have ISBNs. Which takes us back to where I think we all agree (or should!): "give as much information as you can". ~ J. Johnson (JJ) (talk) 22:01, 4 October 2013 (UTC)

Convenience break

For several years i worked extensively editing and moderating the ISFDB, where noting minor differences in editions and printings is a core concern. Based on this, I can assert that at least in fiction publishing, reuse of ISBNs for different printings is more or less the rule, and reuse for truly different editions is not all that rare. Each ISBN costs the publisher a fee, and unless both versions are to be on sale at the same time, the publisher has no incentive to change as often as the published standards suggest should be done. (this does mean that different bindings of simialr date generally ahve different ISBNs, since they have different prices.) Many publishers do make such changes, but many, particularly smaller publishers (but including some larger publishes intent on cost-cutting) routinely reuse ISBNs. There is, I fear, no single utterly reliable identifier for a given edition, and much less for what collectors call a "state" (two books in the same state should be essentially identical, same cover, same binding, same pagination, same contents -- a misprint of a single word may define a variant state). That said, the ISBN is probably the single most useful identifier for Wikipedia's purposes, although citing a LCCN for US-published books, or a similar national number for thsoe countries that have them is often helpful also. DES ^(talk) 22:41, 4 October 2013 (UTC)

What DES says is consistent with the evidence above. Some discrepancies may be due to difference in the way of counting pages, but often there really is a different edition. In the example that started this, there is no way that the 1992, 837 page edition was the same as the 2000 768 page edition with new introduction and notes added. On a different tack, it occurs to me that my misguided faith in Google metadata, which seems quite error-prone, and the Wikipedia citation tool for Google Books has led me to insert thousands of dubious citations. I hate to say this, but in these cases the url, or perhaps url+accessdate, may be the best clue to what I was actually citing. That should not be - I should have looked at the preview of the front matter, assuming it was available, but I did not. Many editors are likely to make the same mistake. Aymatth2 (talk) 23:10, 4 October 2013 (UTC)

Thank, David, that's very interesting. When you talk about ISBNs as useful identifiers in citations, what information (or other benefit) would an ISBN add to John Rawls, A Theory of Justice, Harvard University Press, 1971? SlimVirgin ^(talk) 22:59, 4 October 2013 (UTC)

Well if you want to buy a used copy online, the ISBN is still often the key datum. Particularly for books a bit newer than 1971, it can also be a key to finding a library copy via OCLC. (An OCLC accession number can also be helpful if that is your metadata source.)

And in general, an ISBN will normally (not absolutely always) uniquely identify a particular work, if not a particular edition, and I have seen two different works with the same author and title. It can be a useful clue to which edition is being used, particularly if someone cites a copyright date instead of the edition publication date, as very often happens. Not definitive, but a clue, and it may often give at least an earliest possible date for the edition involved -- provided it does NOT come from amazon.

Amazon metadata, I am sorry to have to say, is often little more than a joke. Amazon routinely shows a current ISBN with an older edition, shows as the publisher of an older edition the newer publisher which the older publisher has merged into, shows a current cover for an older edition, and in general commits anachronisms freely and often. It also routinely invents page counts and publication dated for newly published or pre-announced books if the data has not yet been provided by the publisher, and sometimes these do not get over-written by valid data. If the page count is an exact power of 2 at Amazon, it is quite probably bogus. Amazon's "Look inside" views are usually accurate, but are often for editions other than the one from which they were reached. The ISFDB, after long and painful experience, uses Amazon data extensively, but never trusts it until it is confirmed by a more reliable source, such as an OCLC record, a trusted reference, or better yet, a known user who has the actual book in hand. DES ^(talk) 02:43, 5 October 2013 (UTC)

ISBN provides a convenience link to the "Book Sources" page, which then provides links to vendor indexes like Amazon or Google, Worldcat for library data, Goodreads for reader reports and so on. Maybe as DES says these indexes refer to different editions, but they are usually close enough in practice. The citation can also be enhanced with links to the author and publisher, and a link to an online copy, as: Rawls, John (30 June 2009) [1971]. A Theory of Justice. Harvard University Press. p. iv. ISBN 978-0-674-04260-5. None of the links are required, but they all may be useful. If Amazon metadata is poor, Google also has problems, see doi:10.1080/19386389.2012.652566. But editors will use the Google metadata anyway, and must use it when Google fails to provide a preview of the publication data. We may as well give as much information as possible in the hope that someone who cares can figure out what we were actually referring to. Like keeping possible evidence from the crime scene, even if we are not sure it is relevant. Aymatth2 (talk) 13:26, 5 October 2013 (UTC)

  Re your earlier comment: Especially look at the front matter!!! Particularly the title page and copyright page, as that's where the key publication data is. Access date doesn't seem useful here (Google's image of the book is unlikely to change, though that is a possibility), unless you are referring (as we are in this discussion) to Google's metadata itself. (E.g., what Google says the page count is, versus the actual page count.)

  But where you say an editor "must use" the Google metadata: no. The original editor — who presumably can view the material, or he wouldn't (shouldn't!) be using it — must provide a certain minimum of information. Short of some kind of documentation (or psychic abilities) no one else can provide that, and it is wrong to do so. ~ J. Johnson (JJ) (talk) 19:58, 5 October 2013 (UTC)

An editor will often decide to research and write up an article using only sources available online. It could be a straightforward subject like a lake where no particular background knowledge is needed. They find sources that seem reasonably authoritative, extract the relevant information and provide citations to the sources they have used. One source may be a book written by a reputable author and published by a reputable publisher. The visible page on Google Books says the lake is at an elevation of 325 meters above sea level, and that roughly agrees with what they see on a topographic map. But Google Books does not show them the front matter in preview mode. It is not wrong for them to record what they found, and to record the Google metadata in the citation, even if the metadata may not be exactly accurate. It is certainly not realistic to expect that editors will not use Google books preview for research.

In this case the url seems important to me. Whether or not the metadata exactly matches the scanned page image the editor has used, at least the url lets someone check the citation - if they have access to the Google page image, which may not be true. Not perfect, but better than nothing. If we said every article has to be based on viewing printed books or journals, most contributions would grind to a halt. Aymatth2 (talk) 00:54, 6 October 2013 (UTC)

  I think you need clarify what you are arguing about. E.g., is your hypothetical editor researching the elevation of a lake? Or the page count of some book? The former is content in the book, the latter is the metadata about the book. Also: your last paragraph is entirely immaterial. Of course the url is important; I have said nothing to the contrary. Nor has anyone said anything like "every article has to be based on viewing printed books or journals".

  However, the basis of all those scanned images at Google Books is in fact printed books. Google provides a convenient on-line view of the book, but what you cite is the book itself, not Google. If the Google preview does not show the title page and other front matter it is exactly like a printed book with those pages torn out. So how do you know who wrote it and published it? From the Google metadata? That is someone else telling you what it is, not you reading it yourself. And in regard of ISBNs (the subject here, remember?) I say that if you don't see it in the book itself, or in the image of the book, then you should not use it. A prime example is where you referenced (above) the Goodreads "edition" of David Copperfield: an ISBN was mentioned in the metadata as the source of the text, but the Goodreads text has no ISBN, and therefore it is an error to assign it an ISBN. ~ J. Johnson (JJ) (talk) 22:32, 7 October 2013 (UTC)

If Google says the book is Lacs du Sahara (1967) by Professor J. Dollaeans, the visible page heading/footing says "Lacs du Sahara - Mali page 45", and the text is about a lake in Mali, it is an image of page 45 from Lacs du Sahara. A typical editor would use the rest of the metadata supplied by Google, e.g. publication date, ISBN and author, even if they cannot see an image of the page holding this information. There could be an error in the ISBN, publication date or even the name. Maybe Dollaeans should be Dolléans. There could be a typo: maybe the lake elevation is 328 meters, not 325 meters. But I am willing to say "here is what I found, where I found it, and what Google said it was. I think it is credible, but you make your own judgement." I would include the ISBN that Google provided for those who want to follow the link to Book Sources. I should probably take the trouble to check that some other book index agrees with the title and author for the ISBN. We could recommend that. Aymatth2 (talk) 00:04, 8 October 2013 (UTC)

Again, you need to clarify what you are arguing about. You started by asserting that ISBNs seem to not uniquely identify different books ("including edition and format"), based on "oddities" of different page counts reported in the metadata. Now you assert that it is "not wrong" for editors to rely on metadata "even if the metadata may not be exactly accurate." That some metadata is incorrect, or an editor unwisely relies on it anyway, or simply misunderstands it, is not a fault with the ISBNs generally. It seems to me your concern is really about the metadata. The safest course in that regard is, as I said before: don't rely on what you don't find in the source itself. ~ J. Johnson (JJ) (talk) 19:19, 9 October 2013 (UTC)

I disagree with J. Johnson's comments above. Frequently when I research a subject online I will use google books and find a relevant citation in a google books preview. I will cite the book based on this. Insofar as possible, I will derive bibliographic info from the displayed book images itself. But where these are incomplete (particularly if the front matter is not included in the preview) I will routinely use the google metadata, or even the Google links to vendors od the book (esp if two or more vendors agree) to supply such meta data as the publisher, year of publication, and ISBN. TO simply leave these out would seem to me highly remiss. I would say it is not merely desirable but morally obligatory to supply this information from the best sources available. There is never 100% surety, even the printed book can have a misprint. By including the preview URL I make it clear exactly what I viewed to form the cite. DES ^(talk) 22:06, 9 October 2013 (UTC)

  We are in agreement on several aspects: that nothing is perfect (even printed books), that including the preview URL is strongly recommended, and even that as much information as possible should be included. But note carefully: it seems to me that the crux of the issue initially raised here — the "oddities" — are not so much regarding ISBNs themselves as the use of possibly unreliable metadata (which is where ISBN issue came in).

  There are several problems with doing research by Google snippets (such as not seeing enough of the source to assess its reliability or even applicability), but let's focus on the narrower domain of possibly unreliable publication data (e.g., author, date, edition, publisher, isbn, etc.). And we have a prime example at hand: Aymatth2's attribution, apparently based on the metadata, of the Goodreads edition of David Copperfield to ISBN 9781853260247. This is incorrect, as that ISBN does not apply to the Goodreads edition itself, but (apparently) only as the source of the text (formatting and pagination differing). The problem is not with the ISBN, but its relationship (which Goodreads is not very clear about). To cite this edition with this ISBN is simply incorrect.

  Now consider the Google Books image. The page with the ISBN, which is authoritative, is not shown, barring direct verification. We might presume the metadata to be correct, but that is relying on some one else. We are essentially quoting Google Books, which strictly speaking requires its own citation. There's the bind: does including as much information as possible include unverified and possibly misinformation?

  I point out a possible resolution. In bibliographic practice it is conventional that when information not found on a title page is included in a citation (such as expansion of initials, or the original date of publication) it is placed in square brackets. Perhaps this would be useful to indicate where publication data has been taken from a secondary source such as metadata. ~ J. Johnson (JJ) (talk) 20:36, 10 October 2013 (UTC)

From the above discussion, it is indisputable that some publishers repeatedly use the same ISBN for distinctly different editions. That should not come as a big surprise. Big Sur Biorganic Books & T-Shirts may be casual about the rules, as may larger cut-price publishers. It is also indisputable that the metadata provided by Amazon.com or Google Books is often wrong. When citing a book viewed online we should give the url, ISBN and the metadata provided by the online source if we cannot see the book page that holds the metadata. Sometimes this will be incorrect, but it is the best we can do. By giving the url we are saying that this is the information provided by the website. Sticking stuff in square brackets is not going to help. Aymatth2 (talk) 00:20, 11 October 2013 (UTC)

  Your repeated statement "we should give the url" is a strawman argument, as no one has suggested otherwise.

  Your demonstration of supposed ISBN "oddities" is weak, especially where (as I have shown) you misattribute the ISBN of one publisher to the edition of a different publisher. The problem shown is not in the use of ISBNs, but in misuse of metadata.

  What you have apparently failed to grasp is that relying on someone else's metadata increases the chances of error (either in the data itself, or understanding its proper application). The use of square brackets would be a caveat that the datum came from a non-authoritative source, and could use verification. ~ J. Johnson (JJ) (talk) 22:13, 11 October 2013 (UTC)

Order of Information in Citing Newspaper Articles

I changed the listed order of information suggested for citing newspaper articles because the page previously seemed to suggest that the name of the author and title of the article not be given first. I changed this without discussion as it should not be controversial; having the author and article name first is consistent with the all the other examples given on this page for Wikipedia citations in which there is an author, and also is consistent with the Chicago Manual of Style citation guide. - Embram (talk) 15:53, 29 September 2013 (UTC)

• That seems consistent and uncontroversial, and is what template:cite news does:

Jardine, Cassandra (5 August 2004). "The return of the secondary modern". The Daily Telegraph. London.

Aymatth2 (talk) 16:03, 29 September 2013 (UTC)

Except that is not the order of the new revision. It says: 1) author, 2) title, 3) name of newspaper, 4) date, 5) city of publication. Which has several defects. E.g., why should the date come between the newspaper and place of publication? And: where references are ordered by author and date, it is common to have the date immediately follow the author. It also suggests that where an author is not given the reference should be ordered by title, whereas typical practice is to use the newspaper name. I don't say that the original formulation was correct, only that the current formulation is inconsistent with {{cite news}}, and not fully satisfactory in itself. ~ J. Johnson (JJ) (talk) 21:59, 29 September 2013 (UTC)

• I didn't want to change it beyond what Wikipedia already says, especially since I'm not a member of the Style Manual "club" here. Personally, I like the Chicago Manual of Style method (but with Wikipedia-style embedded links), in which the date generally comes at the end, except for page number(s), if any. For example:

Daniel Mendelsohn, “But Enough about Me,” New Yorker, 25 January 2010, p. 68.

or

Daniel Mendelsohn, “But Enough about Me,” New Yorker (25 January 2010), p. 68.

Sheryl Gay Stolberg and Robert Pear, "Wary Centrists Posing Challenge in Health Care Vote,” The New York Times, 27 February 2010, accessed 28 March 2013.

There was a recent RFC about where to put the date in a citation – see Help talk:Citation Style 1/Archive 3#RFC: Consistent date location. I don't think anything much came of it, but it looks like there's broad consensus for having the date as the second element (immediately after the author). DoctorKubla (talk) 07:19, 30 September 2013 (UTC)

There are two obvious problems with this "broad consensus"¹ that the date should be the second element. First, it's inconsistent with the citation methods accepted by the rest of the literary world (including The Bluebook and The Chicago Manual of Style.) Second, it's inconsistent with Wikipedia's own citation templates (like {{cite news}}, for instance), which generally put the date at or near the end of the citation, consistent with the rest of the literary world. - Embram (talk) 20:16, 30 September 2013 (UTC)

Unless the template has been changed recently, I was wrong about the date in {{cite news}}. It's unfortunate the date is put after the byline because it makes it seem like the date refers to the author, when actually it refers to the date the newspaper edition containing the article was published. - Embram (talk) 00:44, 3 October 2013 (UTC)

¹ By a vote of 8 to 3 out of 11 amateur editors, with discussion closed when the discussion leader (who agreed with that vote) decided to close it. That's a "broad consensus"? Really?

• Also, as an aside, I notice that the current {{cite news}} seems to use the European punctuation style of putting the period outside the quotation marks surrounding the title of the article. Is this done on purpose?

- Embram (talk) 02:18, 30 September 2013 (UTC)

Embram, your comments broadly incorrect.

• There are indeed two general styles of where to place the date: after the author, and at the end of the reference (not counting the specific page numbers part of the reference). If you will dig into CMOS a little more you will see where it acknowledges both. Bluebook is specifically for legal use, and not appropriate here.
• "[T]he rest of the literary world" (however you define "literary") is not the arbiter of citation style and practice on Wikipedia, nor even scholarly/scientific writing at large.
• Your statement that {{cite news}} puts the date at the end of the citation is contradicted by example at the top of this section.
• The change you made does not put the date at the end of the citation.

In otherwords, what you wanted to do is dubious, and what you actually did is incorrect. You should consider reverting your edit. ~ J. Johnson (JJ) (talk) 22:16, 30 September 2013 (UTC)

J. Johnson, you should read more carefully before criticizing, so you don't make strawman arguments. For example, I never said The Bluebook method is appropriate here, and "at or near" is not equivalent to "at," especially when the issue is whether the date should be way up at front, the second element, when citing a newspaper article. - Embram (talk) 12:06, 2 October 2013 (UTC)

  Your own statement explicitly invokes Bluebook in support: "First, it's inconsistent with the citation methods accepted by the rest of the literary world (including The Bluebook and The Chicago Manual of Style.)" If you agree that Bluebook is not appropriate here perhaps you so state. Or at least not suggest otherwise.

  The second part of your run-on sentence reflects some basic confusion. You don't want the date as the second element, and as an alternative cite a general style of putting the date at the end of the bibliographic details that describe the source, including place of publication. (Or near the end, if you include page numbers, which really pertain only to the note form. Read your CMoS.) Well, your change puts the date close to the end, but it misses the style you ascribe to; it is incorrect.

  As your change does not reflect the actual ordering of elements in the citation templates and is contrary to established practice here, and having a contrary example only confuses the readers of this page, it ought to be reverted. Which I will do. ~ J. Johnson (JJ) (talk) 21:33, 2 October 2013 (UTC)

You're just playing rhetorical games. I'm not invoking Bluebook in support of anything. I just mentioned it as yet another real-world citation system that is inconsistent with what you're proposing. For you to seize on that and try to turn it around as though I'm proposing we use the Bluebook system is disingenuous – either that, or you don't understand what a strawman argument is and why you shouldn't use it.

You also seem to be a lot more interested in slapping down what I said than you are in considering appropriate solutions. Your decision to revert the smaller change I made even though all it did was to make the description more (if not perfectly) consistent with what the {{cite news}} template already does seems rather petty: are you changing it back to what is even more inconsistent with a Wikipedia template just to indulge in a power display against me? You could have at least changed it to make it more consistent with the cite news template, which is what I've just done. (Now I see someone else has put the date back near the end - you can argue with them for a while.) - Embram (talk) 00:26, 3 October 2013 (UTC)

If you want sympathy you should avoid making contra-factual statements. And cut-out the name calling. ~ J. Johnson (JJ) (talk) 22:08, 4 October 2013 (UTC)

The text in that section does not determine the order for any particular citation, it only suggests information that might be included. We do not have any house citation style, and in particular each individual article can establish its own order for information in a reference. — Carl (CBM · talk) 00:30, 3 October 2013 (UTC)

That's interesting, and not bad if it's true, but could you cite where that's said? - Embram (talk) 00:50, 3 October 2013 (UTC)

Um, the section "Citation style" of this guideline? — Carl (CBM · talk) 00:52, 3 October 2013 (UTC)

Thanks. In that case, it's perfectly fine in Wikipedia for the original contributor to the article (and those who follow) to put the date of the article publication after the name of the newspaper in which it was published, as is more logical and is in accordance with the Chicago Manual of Style and the rest of the literary world, instead of putting it immediately after the name of the author as though it were some date related to his life, or the pre-publication date the article was actually written (which it certainly is not since we don't know when he wrote it). - Embram (talk) 14:28, 3 October 2013 (UTC)

Yes, the original author can choose any order she likes. But it is worth noting that in CMOS the year can go just after the author name, with only the month and day portion of the date goes later; this facilitates author-date references. Click the "author-date" tab at [1]. Example: Stolberg, Sheryl Gay, and Robert Pear. 2010. “Wary Centrists Posing Challenge in Health Care Vote.” New York Times, February 27. Many of the Wikipedia style templates were designed to be compatible with author-date referencing. — Carl (CBM · talk) 15:06, 3 October 2013 (UTC)

That is indeed a relevant factor. Thank you for making that clear. I like that: year after the author, month and day of publication after the newspaper. - Embram (talk) 15:16, 3 October 2013 (UTC)

  Embram: what you have failed to notice is that in the general style you espouse the date goes after not only the name of the newspaper, but also the city (place) of publication. That is, putting the date between those two elements is not a "style" recognized by anyone.

  I point out to everyone that the original ordering of elements here, with the name of the paper first, and date second, is also very common, as newspaper news articles are often unsigned and collectively attributed to the paper. (See CMOS.) It also results in less confusion where some articles are signed, and some are not. ~ J. Johnson (JJ) (talk) 22:13, 4 October 2013 (UTC)

Question on quotation citation location

I thought at one point our polices required that if you included a quote (with or without in-prose attribution) that the citation immediately had to follow the either the quote or the sentence that used it. I'm running in a situation that an editor claims our policies aren't set up for that, so that as long as the citation for the quote is "near" (in this specific case, after the sentence that includes the quote) that's sufficient. I recognize that this is completely appropriate for two more consecutive sentences that do not include any quotes or contentious material, but I could have remembered the higher standard needed for quoting. --MASEM (t) 22:16, 9 October 2013 (UTC)

You're right – a previous version of this guideline required quotes to be immediately followed (or preceded) by an inline citation. This rule was added by SlimVirgin in May 2006, and removed by SlimVirgin in September 2010, during an extensive rewrite. I can't find any discussion about this in the talk page archives; perhaps SlimVirgin can tell us more. DoctorKubla (talk) 07:59, 10 October 2013 (UTC)

And just reading to be clear: I have sentence A that includes the quoted language from source X, and sentence B that also relies on source X but is quote-free. I had it (based on what I thought was in the language) as "Sentence A.(ref to X) Sentence B.(ref to X)" while someone else, because we no longer have the "immediate" requirement, believes this can be simplified to "Sentence A. Sentence B.(ref to X)" , which if there was no quoted material floating around, I'd completely agree. --MASEM (t) 14:32, 10 October 2013 (UTC)

Normally I would use "Sentence A (ref). Sentence B." Whether a reference is needed on the exact same sentence as a direct quote depends, though, on exactly how the paragraph is worded. If the article says something like Smith (1992, p. 5) claims the opposite. He writes that alien abductions are not hallucinations, but instead are "misperceived stimuli"., then the reference on the thesis sentence of the paragraph is sufficient for verification purposes. It is usually the case that things can be worded in a way to avoid repetitive citations, perhaps with some revision. — Carl (CBM · talk) 23:45, 10 October 2013 (UTC)

I agree that there may be rearrangement, but I do want to get to the core issue is if we have to have "Sentence A (ref)." because of the quote in sentence A. As DoctorKuble points out, the language that required that was removed in a rewrite and if that was an intentional removal or not. I certainly can work the sentences around to try to avoid dup refs, but I need to make sure that, at least as best I remember, we need the ref right after the quoted material. --MASEM (t) 02:13, 11 October 2013 (UTC)

DoctorKubla pinged me about this. It used to be the case that refs were expected directly after a quotation, but that hasn't been the case for some time. It can cause problems if the previous text was from source A, then a quote from source B, then more from A, so you're jumping back and forth adding unnecessary footnotes to the text. If the material's contentious, it might be a good idea to do this, and some editors do it anyway, but others prefer to place footnotes only after sentences or paragraphs. If there are multiple citations at the end of a sentence or paragraph unit, and you need to make clear which source the quote is from, that can be handled with citation bundling. SlimVirgin ^(talk) 03:16, 11 October 2013 (UTC)

Okay, that's fine, and I won't worry about it (though I personally prefer it). I just wanted to make sure if I completely misread something or if there was an actual change. --MASEM (t) 22:16, 11 October 2013 (UTC)

Guideline for how to cite "borrowed" articles?

Perhaps I have missed it, but is there a guideline for how to handle situations in which a newspaper or other news medium publishes an article that was originally developed by a different news agency or a wire service? This often happens with articles from the Associated Press, AFP, and Reuters, for example. Which of the article's identifying items should be included in the citations? A few examples:

• an article published by The Telegraph in which the author is given as "Associated Press"[2]

• an article published by the Los Angeles Times, which includes a named author from the Washington Post[3]

• an artice published by the Houston Chronicle from Reuters with a named author[4]

• an article published by the Houston Chronicle from Reuters with no named author[5]

If you are using the cite news template, which parameters should be included in each of these situations? Dezastru (talk) 01:07, 14 October 2013 (UTC)

Use the "agency" parameter for Reuters and the Associated Press. In the case of newspapers selling on their stories I would try to track down the original source; if you can't, in the case of the LA Times publishing The Washington Post story I would just treat the WP as an agency since it is acting in that capacity. Alternatively, if you just cite it as an LA Times publication it is still a valid cite in terms of providing enough bibliographic details, although you lose an aspect of the original publication. Betty Logan (talk) 01:22, 14 October 2013 (UTC)

Thanks for the speedy response, Betty. Is there a guideline somewhere that states this? I ask not in disagreement but to note that these types of references are very common for certain types of articles and the recommended citation style is not obvious. If there isn't a guideline, or a suggested example, shouldn't there be one? Dezastru (talk) 09:22, 14 October 2013 (UTC)

Also, I would like to propose a couple of tweaks for the cite-news template. Where is the most appropriate place to make the proposal? Dezastru (talk) 09:22, 14 October 2013 (UTC)

I think Help talk:Citation Style 1 (which is where Template talk:Cite news redirects) is the place to go. Thincat (talk) 11:05, 14 October 2013 (UTC)

VisualEditor's reference dialog

VisualEditor's reference dialog, especially the way it handles citation templates, is being redesigned. Suggestions and opinions are wanted at mw:VisualEditor/Design/Reference Dialog. (It's over at the sister project, Mediawiki.org, but you should be able to use your Wikipedia username/password to login there.) This is a practical, focused workshop to improve the design. Views from people who regularly edit at other Wikipedias are also needed, since some Wikipedias do not use the same citation templates (or don't use citation templates at all).

If you aren't using VisualEditor regularly, it may be helpful to turn on VisualEditor (at the bottom of Special:Preferences#mw-prefsection-editing) and try to add or edit a couple of refs in your sandbox, or to read Wikipedia:VisualEditor/User guide first. Whatamidoing (WMF) (talk) 19:01, 16 October 2013 (UTC)