Wikipedia talk:COinS/Archive 1

Citation template
To add COinS to Citation, we need to:
 * 1) Figure out all of the formats it can handle, like books, web sites, and newspapers
 * 2) Figure out the appropriate COinS markup for each of those formats
 * 3) Add them to the individual templates, like cite web and cite news.
 * 4) Then figure out what Citation uses to switch between cases
 * 5) Copy COinS tags from each individual template into Citation, get rid of redundant content and use the switch to switch between the parts that aren't the same between formats. — Omegatron 16:30, 18 April 2007 (UTC)

Thing Citation template handles

 * Books
 * Cite book
 * Already done
 * Journals
 * Cite journal
 * Already done
 * Newspapers
 * Cite news (but also handles televised and web news?)
 * Added a tag
 * Magazines
 * Edited books or parts
 * Encyclopedias
 * Encyclopedia articles
 * Contributions, republications, or edited quotations in a periodical article?

COinS formats

 * Book
 * info:ofi/fmt:kev:mtx:book ("This Matrix represents a single document published at one time.")
 * Genres:
 * book : a publication that is complete in one part or a designated finite number of parts, often identified with an ISBN.
 * Same as cite book
 * bookitem : a defined section of a book, usually with a separate title or number.
 * conference : a publication bundling the proceedings of a conference.
 * Same as cite conference
 * proceeding : a conference paper or proceeding published in a conference publication.
 * Same as cite conference?
 * report : report or technical report is a published document that is issued by an organization, agency or government body.
 * document : general document type to be used when available data elements do not allow determination of a more specific document type, i.e. when one has only author and title but no publication information.
 * unknown: use when the genre of the document is unknown.
 * Journal
 * info:ofi/fmt:kev:mtx:journal
 * Genres:
 * journal: a serial publication issued in successive parts.
 * Same as cite journal?
 * issue: one instance of the serial publication.
 * article: a document published in a journal.
 * Same as cite journal
 * conference: a record of a conference that includes one or more conference papers and that is published as an issue of a journal or serial publication
 * Same as cite conference?
 * proceeding: a single conference presentation published in a journal or serial publication
 * Same as cite conference?
 * preprint: an individual paper or report published in paper or electronically prior to its publication in a journal or serial.
 * Same as cite paper?
 * unknown: use when the genre of the document is unknown.
 * Dissertation
 * info:ofi/fmt:kev:mtx:dissertation (" 	This Matrix represents a dissertation related to a course of study at an institution of higher education.")
 * cite paper covers this, but also other things? - 'This template is for other kinds of "papers", for example a thesis, or an essay or paper that has been separately published (including papers on arXiv).'
 * Patent
 * info:ofi/fmt:kev:mtx:patent
 * Already in US patent reference, Ref patent

Webpage references
'' Moved from my talk. — Omegatron 06:17, 21 April 2007 (UTC)''

So as far as I can tell, the whole COinS/openurl thing doesn't seem to support webpage references at all yet, and not expected to for a really long time. Would it be madness to suggest that a modified form of the COinS metadata for cite journal be used for cite web? (that is, see User:Gwern/test and User:Gwern/test2). Seems to work, and it's the closest in concept anyway. --Gwern (contribs) 04:48 21 April 2007 (GMT)


 * That's what I thought at first, and why I used the journal format for the "Cite this article" page, but then I was looking at Zotero's website with OpenURL Referrer turned on, and they have COinS tags with things like "blog post" that they are apparently doing with the Dublin Core format. We should look into that.  There will also be the hCite microformat in the future, which i think does basically the same thing but in a different way. — Omegatron 05:56, 21 April 2007 (UTC)


 * And now that I look again, they're talking about us! And claiming that all of our citation templates have COinS!  Uh oh.  ;-)  Time to get to work... — Omegatron 05:59, 21 April 2007 (UTC)


 * Actually, it's even worse than that: the main COinS page lists Wikipedia as one of the sites implementing COinS! (With it implied that all our stuff does it, because how could a big professional site like Wikipedia have implemented something like COinS on only some of its articles? :) --Gwern (contribs) 17:45 21 April 2007 (GMT)


 * Hehe. We need to get to work!  Amusingly, I broke all of them by URLencoding the wrong part the other day, and then was too busy to work on it.  We need more than just me working on these... — Omegatron 00:35, 24 April 2007 (UTC)


 * Strangely, will import, but  does not for me.  Their COinS are identical in format as far as I can tell .  I can copy and paste the following, though, which is totally excellent:




 * — Omegatron 06:16, 21 April 2007 (UTC)


 * Speaking of which, are there any plans to change the export format? I was a little surprised to see that it didn't produce a bulleted & sorted list but rather line-separated entries. --Gwern (contribs) 17:45 21 April 2007 (GMT)


 * I think it's fine. Sometimes you want them inside ref tags, sometimes you want them in an ordered list.  Sometimes an unordered list.  It still does a lot of the work for you. — Omegatron 20:29, 21 April 2007 (UTC)


 * But if I include the same COinS over here, neither are recognized by Zotero: O .  So is there something else on the page that it is reading?  I can't find anything. — Omegatron 15:57, 21 April 2007 (UTC)


 * Are you sure Wikipedia isn't in someway mangling or eating the stuff inside the  tag? Also, I notice that each one includes a relevant URL inside the domain. --Gwern  (contribs) 17:45 21 April 2007 (GMT)

<span class="Z3988" title="ctx_ver=Z39.88-2004&amp;rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&amp;rfr_id=info%3Asid%2Focoins.info%3Agenerator&amp;rft.title=About&\ amp;rft.aulast=&amp;rft.aufirst=&amp;rft.subject=News&amp;rft.source=Zotero                                                                      - The Next-Generation Research                                                                                                                    Tool&amp;rft.date=2006-10-02&amp;rft.type=blogPost&amp;rft.format=text&amp;rft.identifier=http://www.en.wikipedia.org/about/&amp;rft.language=Eng\ lish">You have to put something inside the span or Mediawiki strips it

Aha! After intensive trial and error over at User:Gwern/test2, I have discovered the difference, apparently. If you copy and paste the "WP/ZOT" blog COinS to an empty page, it will not work. But if you edit the title of the blog from "Wikipedia & Zotero:..." to "Wikipedia and Zotero:...", omitting the ampersand, it seems to work fine. As far as I can tell, the start of one field is delimited from the end of the previous by... an ampersand. So I guess the problem was that the ampersand was not being escaped. And why neither was working here I don't really know, but it's probably just how you embedded them into your post. --Gwern (contribs) 18:41 21 April 2007 (GMT)


 * Ampersand makes perfect sense. Thanks. — Omegatron 20:27, 21 April 2007 (UTC)

OK, so the breaking-in-Wikipedia issue is resolved. Shall we figure out how to get into a form we can stick into the template? --Gwern (contribs) 22:25 21 April 2007 (GMT)


 * They were using an older software version.  So theirs is fixed now ([1] [2] ) and the mistake I made the other day is fixed now.  On to more templates... — Omegatron 00:35, 24 April 2007 (UTC)

About COinS on Wikipedia
Could someone add some notes about the use of COinS in Wikipedia (what they are, why we're using them, progress so far, examples, and future plans, say), at this page's parent, please? Andy Mabbett 09:25, 25 April 2007 (UTC)


 * Want to move User:Omegatron there and reword it? — Omegatron 13:42, 25 April 2007 (UTC)


 * Done, thank you - please check. I also left a pointer on your user page. Andy Mabbett 14:18, 25 April 2007 (UTC)

How to mark up article about a book
How would you add COinS/ citation templates to an article about the various editions of a book, such as Handbook of British Birds? Please feel free to do so! Andy Mabbett 15:27, 25 April 2007 (UTC)


 * Hmmm... Well, you could copy and paste the tag from cite book and replace all the fields.  But if this is something we were going to do on all articles, it would be better to have a template for it.  I'm not sure if this is the way COinS is meant to be used, but it's probably a legit thing to do. — Omegatron 17:23, 25 April 2007 (UTC)


 * Have you had a look at the article? It refers to several impressions, a concise volume, and several editions (one of which has an SBN) and a supplementary volume of the latter. Think of it as a test case ;-) Andy Mabbett 12:11, 26 April 2007 (UTC)

Anyone? Andy Mabbett 09:19, 8 June 2007 (UTC)
 * I'm not sure what you are requesting. COinS is not as rich as FRBR, but you can have a separate entity for most of the manifestations.  COinS doesn't use SBN, but "affixing a zero (0) as prefix to a 9-digit SBN creates a valid 10-digit ISBN."  COinS can identify a specific edition & the supplemental volume would, of course, be a different book.
 * COinS cannot be used to simultaneously and richly describe all manifestations of a work. You'd have to use only those fields in common & omit others.  However, there are OpenURL resolvers that will do their best to find other manifestations when given a single one. --Karnesky (talk) 17:42, 8 April 2008 (UTC)

Category
I created Category:Templates generating COinS, to match the existing Category:Templates generating hCards, etc., but only one of the templates I added it to is showing up. Any ideas why? Andy Mabbett 09:32, 8 June 2007 (UTC)

rfr_id
Is there a reason we don't have rfr_ids in COinS-generating templates? --Karnesky (talk) 17:32, 8 April 2008 (UTC)


 * Not that I know of. What should it be?  rfr_id=info:sid/wikipedia.org:books for cite book?


 * http://ocoins.info/cobg.html


 * http://info-uri.info/registry/OAIHandler?verb=GetRecord&metadataPrefix=reg&identifier=info:sid/ — Omegatron (talk) 17:56, 7 June 2008 (UTC)


 * The most important thing is to have "wikipedia.org" in there, but I do like your idea of having some sort of id related to the specific citation template that generates the COinS. --Karnesky (talk) 18:18, 7 June 2008 (UTC)

COinS bloat
Editors here may be interested in Template_talk:Cite_journal.LeadSongDog come howl  18:18, 8 October 2009 (UTC)

Discussions in various places (e.g., Talk:Pain and WT:Citing sources/example style) indicate that a major reason editors do not use citation templates is the performance hit (page generation and page bloat), which has been shown to be due largely to COinS (see Cite journal/Archive 2009 October). The supposed benefit of COinS (facilitating interlibrary loan?) seems small, even trivial, compared to detriment of discouraging use of citation templates. So I ask: is this really a good feature? If there is a benefit to having it in some special cases, would it be better that this be done where an editor specifies it, instead of being done all the time in all cases of these templates? - J. Johnson (JJ) (talk) 22:04, 20 February 2011 (UTC)
 * References are about making it easy for readers to verify information and to find out more about a topic. COinS allows not only easier access to ILL, but easy addition of the references to reference management software such as Zotero and faster access to articles their library subscribes to.  If the bottleneck is page generation, then it should be addressed architecturally (better hardware, better caching, whatever).  If the marginal extra bandwidth needed for COinS is an issue, I would argue that it might be acceptable to strip them from low bandwidth (e.g. mobile) versions of the page, but that removing tangentially related photos or splitting up large articles is likely to bring much larger savings.  --Karnesky (talk) 23:02, 20 February 2011 (UTC)
 * Whether COinS adds some utility (or not) is moot if editors do not use the citation templates that implement COinS. And while the suggestions you make might be worthwhile in any case, all of them together hardly offset the effects of COinS.  Now if it were entirely user optionable (with the default "off"), or implemented independently of the citation templates, or not applied unless the editor so specified, then I think there would be no issue.  But as it is, the very little benefit COinS provides appears to be offset by the loss of benefit where editors are deterred from using citation templates. Wherefore I ask:  is this feature really a net benefit? - J. Johnson (JJ) (talk) 23:49, 21 February 2011 (UTC)
 * People choose amongst the variety of citation mechanisms for any number of reasons, including things as trivial as the punctuation used by compared to the various cite templates.  I choose to use the cite templates partly (though not exclusively) because of the COinS support, but I am not deluded enough to think that a ton of people share my opinion.  However, I'd also point out that, while there are a few who don't use these templates only and specifically due to the COinS, I doubt that description applies to a ton of editors. --Karnesky (talk) 02:35, 22 February 2011 (UTC)
 * I am not aware of any editors who expressly name COinS for shunning citation templates. But performance issues (speed and bloat) are one of the main reasons cited for not using citation templates.  And that does come back to COinS being the biggest factor, and therefore (albeit indirectly) a principal factor in discouraging use of templates.
 * Note that I am not asking that COinS be eliminated. I am asking if it could be made an option (i.e., not shackled to the templates), just as we have accommodations for different styles of punctuation. - J. Johnson (JJ) (talk) 21:41, 22 February 2011 (UTC)

Discussion about removal of COinS from Citation/Core
There is a discussion about removal of COinS data at Template talk:Citation/core, this will impact on most citation templates.--Salix (talk): 22:44, 13 November 2012 (UTC)

Viewing COinS
Is there a way to view the COinS data emitted by a template, particularly a cite template? I'm thinking of a method that can be used directly in a browser without having to add obscure plug-ins, writing a program that invokes an API, etc. This would just be to see how what is emitted relates to what what the template parameters are. Jc3s5h (talk) 19:46, 9 August 2013 (UTC)
 * "view source", or whatever your browser uses as an equivalent. Andy Mabbett ( Pigsonthewing ); Talk to Andy; Andy's edits 21:06, 9 August 2013 (UTC)

Editor's comment moved from project page
I have moved this following editor's comment from the project page to here because such comments rightly belong on talk pages


 * Template:Cite book/doc explicitly allows the use of the authors parameter. The error is not in using the parameter, the error is in attempting to emit COinS metadata. Any attempt to deprecate the authors parameter would require changes to a large number of existing citations, and would require a well-advertised RFC. Jc3s5h (talk) 21:33, 13 May 2014 (UTC)

—Trappist the monk (talk) 16:32, 12 June 2014 (UTC)

Possible accessibility issue
There may be an accessibility issue with the COinS metadata that is appended to citations emitted by the template as well as all the  templates. See Village pump (technical)/Archive 130. -- Red rose64 (talk) 22:57, 24 September 2014 (UTC)

Zotero
Is anyone here involved in Zotero? I'd like to find someone who knows how to fix the Google Books and BBC News descriptions. Google Books isn't returning all of the information that it used to. Whatamidoing (WMF) (talk) 19:45, 26 March 2015 (UTC)

Language metadata pollutes COinS
Now I've actually bothered reading about COinS and looking at the output (which isn't all that easy, given nothing appears to be around to provide a user-friendly UI for COinS metadata), I finally understand why adding lang to arguments within H:CS1 templates is evil-and-wrong.

But that leads me to wonder how we might get the benefit of both language-tagging and of COinS. For ease of editing, I guess embedding lang within the CS1 templates is probably most convenient (?) and having Module:Citation/CS1 remove the language metadata before adding the parameters to COinS.

Does that make sense? Thoughts anyone? — OwenBlacker (Talk) 20:18, 2 January 2016 (UTC)


 * This is a longstanding feature request for which a solution is still at large.


 * Because this talk page is the right place to talk about its associated project page, and because your base question appears to be about implementation of COinS in cs1|2 templates, perhaps this topic is best closed here and taken up at Help talk:Citation Style 1 which is the place where most of the discussions regarding cs1|2 implementation take place.


 * —Trappist the monk (talk) 16:18, 3 January 2016 (UTC)


 * Agreed. I've reopened a dormant discussion at Module talk:Citation/CS1/Feature requests, if anyone has strong feelings on the matter. — OwenBlacker (Talk) 15:20, 7 February 2016 (UTC)

Nowrap in references
What happens when nowrap is used in a CS1 template? Sometimes, I use it to nowrap name initials separated by hyphen (don't know if this is already being done by modules) or to nowrap publisher that contains hyphen because hyphenated words — after they are broken where hyphen is — should have another hyphen below in order to eliminate ambiguity.

all- star

is allstar, for example.

all- -star

is all-star.

However, All-star (without nowrap) is rendered as first example if it is near the right margin. Is there some dictionary or something that proposes hyphen should or must be written again if hyphenated word is broken at the hyphen location and the rest written in new line?

I use in the articles generally nowrap for all hyphenated words, too.

Even if there is a guide, sometimes it is necessary to use nowrap template...--Obsuser (talk) 03:40, 6 June 2016 (UTC)
 * produces this:
 * all-star
 * from all-star, Module:Citation/CS1 makes this metadata:
 * when it should be:
 * In English, I'm pretty sure that I have never seen published writings that use repeated hyphens as you describe (except to emulate an emdash). I have seen oblique double hyphens replace a single hyphen when the line break occurs at the hyphen.
 * —Trappist the monk (talk) 11:41, 6 June 2016 (UTC)
 * In English, I'm pretty sure that I have never seen published writings that use repeated hyphens as you describe (except to emulate an emdash). I have seen oblique double hyphens replace a single hyphen when the line break occurs at the hyphen.
 * —Trappist the monk (talk) 11:41, 6 June 2016 (UTC)


 * What's  used exactly for i.e. why is   bad? I did not understand the main point explained in the COinS or COinS articles.


 * Can you tell for sure if here it is about selfemployment or it is about self-employment [this is just one example but words can be with different meaning if they are hyphenated or non-hyphenated]?
 * There's shy that can be added to any article in order to designate where can each word get broken, and after that — it is not possible to determine if it is a half-compound word or a single compound one-piece word.--Obsuser (talk) 19:15, 6 June 2016 (UTC)
 * There are software tools that can harvest citation data from Wikipedia pages. We support that with COinS.  If I used  to cite a book called All-star, Module:Citation/CS1/COinS will package that title with author, publisher, date, etc data into  tags that are appended to the rendered citation.  A list of the currently supported COinS keywords is at Module talk:Citation/CS1/COinS (also links to the COInS documentation such as it is).  As an example:
 * All of that stuff that isn't a title ( and  ) or isn't part of the necessary COinS formatting does not belong in the   keyword's payload.
 * Because selfemployment is not a legitimate English word, I know that if the right margin were moved sufficiently, the meaning would remain the same as should the hyphen. This place really isn't the proper place to discuss hyphenation.  You might want to read Manual of Style and Manual of Style.
 * —Trappist the monk (talk) 19:58, 6 June 2016 (UTC)
 * I know it isn't; that's why I noted it is just an example. Better example would be re-dress/redress mentioned right in the Manual of Style but that's—as you said—not meant to be discussed on this talk page.--Obsuser (talk) 20:11, 6 June 2016 (UTC)
 * Because selfemployment is not a legitimate English word, I know that if the right margin were moved sufficiently, the meaning would remain the same as should the hyphen. This place really isn't the proper place to discuss hyphenation.  You might want to read Manual of Style and Manual of Style.
 * —Trappist the monk (talk) 19:58, 6 June 2016 (UTC)
 * I know it isn't; that's why I noted it is just an example. Better example would be re-dress/redress mentioned right in the Manual of Style but that's—as you said—not meant to be discussed on this talk page.--Obsuser (talk) 20:11, 6 June 2016 (UTC)
 * I know it isn't; that's why I noted it is just an example. Better example would be re-dress/redress mentioned right in the Manual of Style but that's—as you said—not meant to be discussed on this talk page.--Obsuser (talk) 20:11, 6 June 2016 (UTC)

What to use instead of chem or ?
I'm working through the references on Hydronium and finding several that contain things like which produces ""…which is not easily replicable using just   and   ! However, the rubric on chem says that it should not be used in CS1/2 templates because it pollutes COinS. What is the recommended alternative? TIA HAND —Phil | Talk 18:08, 16 March 2017 (UTC)
 * Alas, not much can be done unless there is a way of writing chemicals as plain text. It used to be that MediaWiki allowed Module:Citation/CS1/COinS to analyze the content of  tags so that it could write a text version into the metadata.  A change to MediaWiki broke that functionality so the content of  tags is not available and renders in the metadata as something like   which is meaningless to those who consume cs1|2 citations via the metadata.  More on this in the cs1 archives – a discussion about Math ML, but quite related.


 * cs1|2 has never been, and probably will never be, able to render clean metadata for simply because that template produces so much html and css.  For your example, the  template output is:
 * Consumers of cs1|2 metadata are then left to decode that on their own.
 * —Trappist the monk (talk) 10:53, 17 March 2017 (UTC)
 * Is it impossible to simply strip out the HTML tags from that text (which I have taken the liberty of prettifying to emphasise the tags)? That would at least produce some kind of usable result. I would have thought that would be a fairly basic operation… —Phil | Talk 16:19, 17 March 2017 (UTC)
 * Is it impossible to simply strip out the HTML tags from that text (which I have taken the liberty of prettifying to emphasise the tags)? That would at least produce some kind of usable result. I would have thought that would be a fairly basic operation… —Phil | Talk 16:19, 17 March 2017 (UTC)