Misplaced Pages

• Working on Lua to show same cite format: There have been some differences in the format of the Lua-based cite templates. Currently, {Cite_web/Lua} shows the equivalent for {Cite_web}:

• Cite: {{cite web | title=My Page | last1=Doe | first=John | publisher=Acme |location=London | url=http://www.google.com |date=5 May 2009 |volume=II | issue=3 |page=6 |accessdate=8 Febrary 2013}}
• Cite_journal: Doe, John (5 May 2009). "My Page". II (3). London: Acme: 6. Retrieved 8 Febrary 2013. {{cite journal}}: Check date values in: |accessdate= (help); Cite journal requires |journal= (help)
• Cite_web now: Doe, John (5 May 2009). "My Page". London: Acme. p. 6. Retrieved 8 Febrary 2013. {{cite web}}: Check date values in: |accessdate= (help)
• Cite_web/lua : local p = {}

local CS1 = require('Module:Citation/CS1')

p = function(frame) local newFrame = { getParent = function(self) return frame end, getTitle = function(self) return 'Template:Cite web' end, args = {CitationClass='web'} } setmetatable(newFrame, { __index = function(t, k) if type(frame) == 'function' then return function(...) return frame(frame, select(2, ...)) end else return frame end end }) return CS1.citation(newFrame) end

return p

To get parameters to match the same format, then lines of Lua script must be moved around into the same order. -Wikid77 (talk) 11:29, 19 February 2013 (UTC)

There is a similar issue with handling of the "origyear" parameter as shown by:

• Cite_web now: Cookridge, E. H. (1972) . "The Baron of Arizona". New York: Ballantine Books. OCLC 32333347. {{cite web}}: Missing or empty |url= (help)
• Cite_web/lua : local p = {}

local CS1 = require('Module:Citation/CS1')

p = function(frame) local newFrame = { getParent = function(self) return frame end, getTitle = function(self) return 'Template:Cite web' end, args = {CitationClass='web'} } setmetatable(newFrame, { __index = function(t, k) if type(frame) == 'function' then return function(...) return frame(frame, select(2, ...)) end else return frame end end }) return CS1.citation(newFrame) end

return p

--Allen3  11:46, 19 February 2013 (UTC)

• Thanks, I have switched us to Lua Module:Citation/CS1 to adjust for CS1 format, because other editors perhaps want the initial Module:Citation to show another style for origyear and such. Formerly, page numbers listed near the end, so I have shifted them back near the end, to match CS1. Values for volume=II and issue=3 are shown now, although {cite_web} had ignored those parameters. The results in Lua are very close now. -Wikid77 (talk) 14:54/17:56, 19 February 2013 (UTC)

The use of accessdate when a page is archived or dead

What is the appropriate use of accessdate when a page is archived or dead? My initial feeling was that the accessdate applies to the archival page (if it exists) and otherwise to the "main" page which has been archived or is dead. Is there guidance somewhere? --Izno (talk) 00:34, 19 February 2013 (UTC)

Access date is simply the date the link was accessed, usually the date it was added. There is no requirement to include it, and it can be hidden by registered users. It is mainly useful for a web page that changes frequently, or when there is no discernible publication date. --— Gadget850 (Ed) 02:06, 19 February 2013 (UTC)

If the page is dead, and there is an access date, the latter apparently can serve to locate an archive version that most closely matches the date the citation was placed, but such utility is IMHO debatable. If there is an online archive of the article, I think it makes little sense to leave an access date that serves a purpose no longer. -- Ohconfucius  02:22, 19 February 2013 (UTC)

I believe I'm the one who argued that, in fact, and applied just that logic today in cleaning up some citations. :) I might agree with the latter sentence, but I'd like thoughts on the below question before I do so. --Izno (talk) 04:54, 19 February 2013 (UTC)

Yes, but for which link? The archive link or the archived link? I suppose it seems a little incredulous to expect the archive link to go down... --Izno (talk) 04:53, 19 February 2013 (UTC)

however, it does happen (at least temporarily). i always archive online sources (when allowed), preferably at WebCite, and have encountered service disruptions several times. 70.19.122.39 (talk) 14:06, 19 February 2013 (UTC)

In my opinion: when the original link is dead, then there is no need for an access date, as the archive has an archive date that will normally not change. Where the original link is still live then the access date is useful if there is no publication date, regardless of whether it is preemptively archive. --— Gadget850 (Ed) 10:57, 19 February 2013 (UTC)

If you are adding an |accessdate=, it pertains to the original |url=. The |archivedate= (although being the date that the archiving service grabbed the page and not the date when you viewed it) does a similar job for |archiveurl= - we know that the archived page isn't going to change further. --Redrose64 (talk) 13:30, 19 February 2013 (UTC)

Yes, but our ability to access the archived page might . That's the only Pandora's box I'm opening here. :) If the lesser consensus here disagrees with me, that's fine, and I'll remove (or at least, not add) accessdates when the cited pages are archived. To that end, I'm not sure the below proposal captures the point I'm making. --Izno (talk) 18:09, 19 February 2013 (UTC)

Documentation

This has come up before, so we should tweak the documentation. Current:

accessdate: Full date when URL was accessed; use the same format as other access and archive dates in the citations; do not wikilink. Can be hidden by registered editors.

Proposed:

accessdate: Full date when original URL was accessed; use the same format as other access and archive dates in the citations; do not wikilink. Not required for web pages that do not change; mainly of use for web pages that change frequently or have no publication date. Can be hidden by registered editors.

--— Gadget850 (Ed) 15:00, 19 February 2013 (UTC)

the reflink doesn't link to a ref here, do you mean the note at the template page? other than that i agree. 70.19.122.39 (talk) 15:13, 20 February 2013 (UTC)

I meant to leave that out for this. It links to an explanatory note on date formats. --— Gadget850 (Ed) 15:18, 20 February 2013 (UTC)

Could we maybe add a bit stronger guidance, something along the lines of "use of accessdate is discouraged where the page has an archive and associated archive date, but is allowed where desired", per the above discussion? --Izno (talk) 22:11, 27 February 2013 (UTC)

1. ^ Cite error: The named reference date was invoked but never defined (see the help page).

 Done --— Gadget850 (Ed) 12:44, 14 March 2013 (UTC)

When a journal article has two years of publication, one electronical and one print year

It is very common these days that articles are published online before the printed version, often in the year before the printed version is published. In such cases, an article may be cited using only the year and the digital object identifier, e.g.

Doe J (2010). "Article title". Journal name. DOI: 00000000000

Then the printed version may be published in the following year (2011), and Vol., No. and pp. become available to be used in citations. Also after the printed version has been published, it's common to continue citing the article using the original (first) year of publication.

How can this problem be solved when using this template? In such cases, the citation should ideally include both original year of publication (ahead of print) and the citation of the printed article. Bjerrebæk (talk) 19:58, 19 February 2013 (UTC)

origyear: Original publication year; displays after the date or year. --— Gadget850 (Ed) 20:25, 19 February 2013 (UTC)

You should give information pertaining to the edition which you actually consulted. There may be differences between the editions. --Redrose64 (talk) 20:41, 19 February 2013 (UTC)

There are seldom any differences between a version published online before print and the printed version of the same article. It is sometimes useful to have the actual full, correct citation, e.g. in a list of publications. Bjerrebæk (talk) 15:29, 4 March 2013 (UTC)

I have always taken "original year of publication" to be that of something that was subsequently republished. The cases referred to here are actually prepublication. And it seems to me that (in the sciences at least) the "real" date of publication is that of the printed item. I think we ought to give this some consideration. ~ J. Johnson (JJ) (talk) 22:16, 4 March 2013 (UTC)

Lua

Lua versions of CS1 templates are available for {{Cite web/lua}}, {{Cite news/lua}}, {{Cite journal/lua}} and {{Cite encyclopedia/lua}}. Please test but do not use in live articles. Report issues at Module talk:Citation/CS1. --— Gadget850 (Ed) 21:31, 20 February 2013 (UTC)

• Also {cite book/lua} but timeout problems: I have also added {{Cite book/lua}}; however, there have been some serious problems with the Lua-based cites hitting the 10-second Lua timeout error (and showing cites as "Script error"), so they are not yet reliable for use in large articles. Others have reported Lua problems with many templates per page. We need to re-focus on faster markup-based templates, in case Lua cites cannot be used in large articles. -Wikid77 (talk) 22:23, 2 March 2013 (UTC)

Bug/typo in all cite templates

Hi, I noticed that seemingly all the "cite" templates have two instances of {{{seperator}}} when {{{separator}}} is clearly intended. I haven't tested what effect fixing the typo would have, so I won't implement it – I'll leave it to someone who's more familiar with the templates.

Seperately Separately, it might be nice to wrap all the templates in <includeonly>; right now it looks like the documentation on most template pages is preceded by a stray period (except template:cite web, which does have an includeonly wrapper: ).--Father Goose (talk) 21:48, 24 February 2013 (UTC)

What you are seeing is {{{separator|{{{seperator}}}, which means if there is a parameter named separator, then use it else if there is a parameter named seperator, then use it. The latter is a common mispelling; it was added before I got into these templates, but I am sure there was a reason, and it doesn't really hurt anything. These duplicated parameters are known as aliases, and are heavily used in the citation templates for a variety of reasons. Most are documented, but this one is for a spelling error and I did not see any reason to tell editors they can use it.

I hate wrapping a template in <noinclude> when there isn't a real reason. If someone makes a mistake, it is not at all obvious when the template doesn't show. {{Cite web}} has it because there is a check for title and it shows a red error message. --— Gadget850 (Ed) 22:44, 24 February 2013 (UTC)

I see. Thank you.--Father Goose (talk) 06:11, 25 February 2013 (UTC)

Slightly smaller typeface for the "retrieved" date?

Citations using cite web may have a date or an accessdate, or both, specified within each entry. The date may appear near the beginning of the citation if author details are present, or near the end of the citation if no author details are shown. In many cases there may be a mix of date styles such as February 27, 2013 and 27 February 2013 within a list of citations.

When citation lists are presented as two or more columns, and the text of each citation therefore wraps to multiple lines, I don't find it easy to look through the list. I am sometimes looking for an entry with a particular date, or looking for entries before, or after, a particular date. I often find I am looking at the "retrieved" date rather than the article date, or vice versa. The accessdate is dependant on the editor adding the entry, and is not a part of the data created by the author or publisher of the original article.

I notice that several non-English Wikis have already overcome this difficulty in a very simple manner. They present the "retrieved" date (and the preceding word) in a slightly smaller typeface than the rest of the citation text (but still quite a bit larger than the "subscription required" or "registration required" text that appears on some entries).

Could this simple change be considered here in the English Wiki?-- 86.136.109.181 (talk) 20:55, 27 February 2013 (UTC)

We do wrap the accessdate text in a span with a CSS class attached, which you could use your browser (or a login plus your personal CSS) to even hide if you wanted. I don't think for IPs that this will change otherwise, however. (No comment on whether it should or should not.) --Izno (talk) 22:07, 27 February 2013 (UTC)

I updated Help:Citation Style 1/accessdate to show how to make it show smaller if you have an account. --— Gadget850 (Ed) 03:04, 28 February 2013 (UTC)

Cite web - full parameter

Hi! At Template:Cite web I noticed that the "Full parameter set" (either horizontal or vertical format) does not include parameters regarding the editor(s), but it does discuss the issue in detail (Template:Cite_web#Editors). I tried to use editor-first and last - and they actually did work. Shouldn't they be included in the full set? Sorry, if I asked something stupid here! Zoli79 (talk) 16:03, 2 March 2013 (UTC)

I have updated a lot of the CS1 documentation, but I have done only minor work on the copy/paste parameters sets. Please update as desired. --— Gadget850 (Ed) 16:45, 2 March 2013 (UTC)

OK, done. Zoli79 (talk) 19:10, 2 March 2013 (UTC)

Deprecated parameters

In the {{cite episode}} template, there are several deprecated parameters, such as "serieslink" and "episodelink" which are listed under "Parameters" as deprecated; however, they are still used in the copy & paste example at the top of the /doc page, under "Usage: Full parameter set in horizontal format". They really shouldn't be listed there, if they are deprecated. This needs to be updated. --Funandtrvl (talk) 17:18, 2 March 2013 (UTC)

See previous discussion. --— Gadget850 (Ed) 18:40, 2 March 2013 (UTC)

Comparison of markup and Lua-based cites

The plan is to switch {cite_encyclopedia}, soon, to use the Lua-based template, as a first phase, to ensure proper handling of parameters. The output has been tested for the Lua version to match the basic functionality:

• Parameters: {{...| |last=Doe |first=J.B. |authorlink=John Doe|editor-first=Edie |editor-last=Tor |editor-link=Editor |encyclopedia=My Encyclopaedia |title=Mein Artikel |trans_title=My Article |url=http://www.aeiou.com |accessdate=5 March 2013 |language=German |edition=2012 |date=1 December 2012 |year= |month= |publisher=AEIOU Staff |volume=IV |location=Vienna |id= |isbn=1234567890 |oclc=Oclc 45 |doi=10.DOI_number |pages=77 |quote=This is a direct quote |ref=harv}}
• Cite encyclopedia/old: Doe, J.B. (1 December 2012). "Mein Artikel". In Tor, Edie (in German). My Encyclopaedia . IV (2012 ed.). Vienna: AEIOU Staff. pp. 77. doi:10.DOI_number. ISBN 1234567890 Parameter error in {{ISBN}}: checksum. OCLC 45 Oclc 45. http://www.aeiou.com. Retrieved 5 March 2013. "This is a direct quote" 
• Cite encyclopedia/lua: local p = {}

local CS1 = require('Module:Citation/CS1')

p = function(frame) local newFrame = { getParent = function(self) return frame end, getTitle = function(self) return 'Template:Cite encyclopedia' end, args = {CitationClass='encyclopedia'} } setmetatable(newFrame, { __index = function(t, k) if type(frame) == 'function' then return function(...) return frame(frame, select(2, ...)) end else return frame end end }) return CS1.citation(newFrame) end

return p Among the slight differences, the option "trans_title" seems more applicable to follow the article title in an encyclopedia, but either way, a user can enter the title of the encyclopedia with an appended translation in square brackets "". The parentheses around the publisher name should probably be removed, as too many items in the curved brackets. -Wikid77 (talk) 18:06, 5 March 2013 (UTC)

format PDF in the Template:Cite journal/doc examples

I am deleting format=PDF from the examples (four instances) in Template:Cite journal/doc. This specification is rarely used and should not be encouraged, as it clutters up the endnote with a "(PDF)" that is wholly unnecessary; the  symbol provides the relevant information. Besides being a distraction for the reader, this usage often leads to inconsistent citation style, since other .pdf references in the same article seldom use this parameter. Peter Brown (talk) 17:35, 7 March 2013 (UTC)

The pdf icon isn't displayed when the url does not include the .pdf extension and not all urls do even though they address a pdf document.

—Trappist the monk (talk) 17:44, 7 March 2013 (UTC)

I did not know that. Nevertheless, the use of format=PDF should be discouraged when the .pdf extension is present, as in the Template:Cite journal/doc examples:

• Aries, Myriam B. C.; and Newsham, Guy R. (2008). "Effect of daylight saving time on lighting energy use: a literature review" (PDF). Energy Policy. 36 (6): 1858–1866. doi:10.1016/j.enpol.2007.05.021. Retrieved September 1, 2009.{{cite journal}}: CS1 maint: multiple names: authors list (link)

• Woodin, W. Hugh (2001). "The Continuum Hypothesis, Part I" (PDF). Notices of the AMS. 48 (6). Providence, RI, USA: American Mathematical Society: 567–576. ISSN 1088-9477. OCLC 34550461.

Perhaps Trappist the monk or someone else can update the documentation with an example without the .pdf, making clear that format=PDF is only appropriate in such cases. Peter Brown (talk) 18:16, 7 March 2013 (UTC)

Format=PDF is also useful for installations that don't load images and for WP:ACCESSIBILITY reasons. I think Format=PDF should be compulsory for all pdf files. -- Magioladitis (talk) 18:31, 7 March 2013 (UTC)

See Help:External link icons for details. The link icons have no alt text, thus do not meet accessibility standards. --— Gadget850 (Ed) 18:41, 7 March 2013 (UTC)

External link icons are surely not what WP:ACCESSIBILITY#Images and other policies and guidelines have in mind. Misplaced Pages:Image use policy, for example, requires that images have an image description; this is not applicable to these icons. Anyhow, if  and  were images in Misplaced Pages's sense, then, even if format= were used, their use would violate WP:ACCESSIBILITY since they don't have alt text. Peter Brown (talk) 21:34, 7 March 2013 (UTC)

I have for a long time been routinely deleting format=PDF from references when the URL ends in .pdf, but not when it doesn't. Although, I must say I don't really see why people need to be warned anyway that a document is in that format. Surely all browsers nowadays have a PDF reader built in? -- Alarics (talk) 21:44, 7 March 2013 (UTC)

Only Firefox and Chrome have PDF readers built in. Not every computer or tablet can read PDFs out of the box. Not everybody needs to have a PDF reader installed. Also, not all PDFs can be "read" via a screen reader. While many PDFs will work fine, only PDFs that are PDF/UA complaint will be guaranteed to work with screen readers. Only a few non-free products can create PDF/UA and not by default. Anything scanned and turned into a PDF will not work with screen readers. Bgwhite (talk) 23:42, 7 March 2013 (UTC)

Whether or not a tablet can read PDFs, (PDF) and  are redundant. If the user doesn't have the PDF software, either indicates that the source is not directly available. Deleting format=PDF when the URL ends in .pdf still makes sense.

Screen readers are another matter, and many of us are not familiar with the technology. Bgwhite, would you explain further? Can screen-reader users follow external links at all? If not,  and (PDF) are equally useless, so including format=PDF is of no help. Suppose, though, that they can follow links. Do the icons register the presence of an external link that might be followed? If they don't, some text indication should be provided in connection with  as well as with . Peter Brown (talk) 00:31, 8 March 2013 (UTC)

I'm certain that screen readers can follow external links; but I don't know what happens when the target is a PDF file. I shall invite Graham87 (talk · contribs) here, who probably knows more than most people about the capabilities and limitations of screen reader software. --Redrose64 (talk) 09:08, 8 March 2013 (UTC)

Thanks for letting me know about this discussion. Yes, screen readers work with many, but not all PDF documents, as Bgwhite said above. I've always been a stickler for adding format=pdf or similar to links to any PDF files, because I do appreciate a warning that a link is in PDF format (the Adobe PDF plugins are slower than browsers with my screen reader JAWS, and the layout is obviously different). Until I read this conversation, I had no idea that a PDF icon came up automatically when a PDF file is directly linked. Alt text should be added to these icons, which would benefit not just screen reader users but also people who don't use images. Once the alt text is added, the format parameter can be depricated for all direct links to file formats with automatically generated icons, and a bot should go through and clean up all the redundant format parameters. Graham87 09:50, 8 March 2013 (UTC)

Template:Bug - External link icons should have alt text. --— Gadget850 (Ed) 10:33, 8 March 2013 (UTC)

Even if the link icons gained alt text and that rendered the parameter largely redundant, we'd still have the visual inconsistency that some links have an icon and other links will have the text. I'm sorry, but I'll stick to consistently adding |format=PDF to all citations to PDFs (or any other non-HTML-based source). Imzadi 1979 → 10:38, 8 March 2013 (UTC)

That suggests a larger problem with pdfs. We'd serve users better by distinguishing the purely scanned-image pdfs from those that have accompanying text. This applies not just to users of screen readers, but also to bandwidth-limited and mobile users. A scanned-image pdf of a book can be tens of megabytes.LeadSongDog come howl! 14:25, 8 March 2013 (UTC)

Imzadi1979, your response is puzzling. There is currently an inconsistency in that format= is sometimes used for  but never for . Graham87's proposal, providing alt text for both and using a bot to eliminate format=PDF where it is redundant, would get rid of the inconsistency. Are you opposed? Peter Brown (talk) 15:54, 8 March 2013 (UTC)

No, it shouldn't be. |format=PDF is telling the reader that the link is to a PDF, just as |format=MrSid tells a reader that a map is in the MrSid format. Both are potential warnings that additional software may be required to access the link. The fact that the MediaWiki software includes an icon for PDFs is irrelevant to that notice. My point is that if half of the links in an article's references have the PDF icon without "(PDF)" appearing, and the other half lack the icon but contain "(PDF)", there is still an inconsistency in visual display, and I'll err on the side of always including the format notation regardless if the link generates a graphic or not. Imzadi 1979 → 00:45, 9 March 2013 (UTC)

Well, if that view is accepted we should surely abolish the PDF icon altogether since there is no need for both, and it just becomes more irrelevant clutter. -- Alarics (talk) 07:46, 9 March 2013 (UTC)

abolish ALL such icons, not just the PDF. the |format= and its output provide more exacting and understandable information. in the doc, suggest to editors, that the "format" field's data should be linked, at least in first use. i do not understand the argument about "citation style inconsistencies". surely this is an article-style problem that should be fixed at the article, not at the citation system level. 70.19.122.39 (talk) 15:22, 9 March 2013 (UTC)

ORCID

ORCID identifiers disambiguate journal contributors with similar names; and unite records for authors who write under more than one name. Think of them as an ISSN for people. For example. my ORCID is 0000-0001-5882-6823 and the corresponding URI is http://orcid.org/0000-0001-5882-6823

I believe that we should add an option for an ORCID parameter for each author of a cited work, to CS1 templates. Whether or not to make that a clickable link can be debated. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:33, 8 March 2013 (UTC)

I believe ORCIDs are a good idea within the {{Authority control}} template on Misplaced Pages articles about individual researchers. I am less convinced that we should clutter our bibliographies with them, especially in cases where we have a wikilink to the author's article. —David Eppstein (talk) 22:17, 8 March 2013 (UTC)

I suspect (and would welcome evidence) that in the vast majority of cases, we have no article about the author, and thus no such link. ORCIDs would not be clutter, but valuable information. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:53, 9 March 2013 (UTC)

I support this. It should be a clickable link and the template should check the format of the ORCID for mangled ones. Stuartyeates (talk) 06:12, 9 March 2013 (UTC)

I agree with David, they should be in the authority control system and not in the citation templates themselves. I don't see how they help a reader locate a copy of the source to verify the information being cited. I do see how the ISSN would help a reader locate a library that maintains a copy of the journal/magazine, but I'm not sure the same can be said for ORCIDs. Imzadi 1979 → 06:37, 9 March 2013 (UTC)

You're quite right. But then no such claim has been made. An ORCID doesn't help find the publication; it helps find the - and unambiguously to identify - the author; and thus to find that author's other works, affiliation, reputation, etc. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:51, 9 March 2013 (UTC)

The purpose of a citation is to allow a reader to find the publication being cited for verification purposes. The other works of the author, etc would best be spelled out in an article on that author, which conveniently enough, |authorlink= would provide. From that article then, the ORCID and other authority control information would be linked. Anything else in the citation is clutter. That isn't to say that the ORCID isn't valuable, just that it isn't needed, in this context. Imzadi 1979 → 12:31, 9 March 2013 (UTC)

Please refer to my above comment, where I have already addressed the |authorlink= red herring. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:35, 9 March 2013 (UTC)

one solution would be to have "authorlink" output either a wikilink or an ORCID? add logic in the template so that if the field's data start with a number link to ORCID. i realize there could be situations where people (legally) use numbers as names. in which case my idea would not be optimal. 70.19.122.39 (talk) 15:37, 9 March 2013 (UTC)

If we introduce data like this (sounds useful for cross referencing, metadata), we have to be able to strike a balance between the additional data and a default display of the reference that is not too detailed for an average reader. Not sure how we would do this/how it would be configured though. Rjwilmsi 10:17, 12 March 2013 (UTC)

If we must, we could hide ORCID numbers behind an icon, or the text "ORCID"; but displaying them would be better. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:31, 12 March 2013 (UTC)

.39's solution would be technically easy, just "if authrolink exists, do not display ORCID". Numeric names would not be a problem. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:31, 12 March 2013 (UTC)

your amendment is more elegant and takes care of names that are numbers. the only potential problem i can see is that #ifexist is expensive, or more so than plain #if statements that don't check article namespace. conceivably an article could bump into limits when there are many citation templates? 70.19.122.39 (talk) 14:00, 12 March 2013 (UTC)

And it would mean that when printed or ported, the ORCID would be hidden. --— Gadget850 (Ed) 14:15, 12 March 2013 (UTC)

That's easily dealt with by CSS, as we do with external links in citations, for instance. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:22, 12 March 2013 (UTC)

Doesn't Lua resolve this? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:22, 12 March 2013 (UTC)

Would someone like to knock up a test implementation in a sandbox, please? If you need a sample ORCID, mine is on my user page. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:26, 12 March 2013 (UTC)

|publication-date= does not pass data to {{harv}} when verbose

pertinent examples are emphasized. i suspect whitespace is the prob.

70.19.122.39 (talk) 14:24, 13 March 2013 (UTC)

The five {{harvnb}} are generating links to, in order, href="#CITEREFAuthor2000" href="#CITEREFAuthor2001" href="#CITEREFAuthor2002" href="#CITEREFAuthor2003" href="#CITEREFAuthor2004" whereas the generated anchors are id="CITEREFAuthor2000" id="CITEREFAuthor2001" id="CITEREFAuthorJanuary_2002" id="CITEREFAuthor1_January_2003" id="CITEREFAuthor2004" But the |harv=ref code only uses |publication-date= as a fallback, when both |year= and |date= are absent:

|publication-date= is intended for use in addition to, not instead of, |date=, so when you only have one, it's best to use |date=. --Redrose64 (talk) 15:30, 13 March 2013 (UTC)

Correct: It is the date of publication when different from the date the work was written. A real example:

White, T.H. (1941). The Book of Merlyn. University of Texas Press (published 1977). {{cite book}}: Invalid |ref=harv (help)

--— Gadget850 (Ed) 16:00, 13 March 2013 (UTC)

i think you guys misunderstood.

|publication-date= passes data to hrefs when (a) only the year is given (b) the href is input with date formated exactly as the data in the "publication-date" field, or (c) {{harvid}} is used

when a citation is formatted with a verbose date (eg Jan 1, 2012) template code tries to extract year from |date=, and this is passed to |ref=harv.

but this does not happen with |publication-date=.

this is a bug. pls fix.

the suggestion to use |date= when only the publication date is known presents several problems.

what happens in online sources (eg {{cite web}}) when neither work date nor pub. date are known? do we input the access date in |date=?

if not, there's an inconsistency. .

on the other hand, if the access date should become the variable data for |date= then the dependency between |accessdate= and |url= should be undone to avoid errors.

and what about archival dates when |deadurl=yes? another clarification should be provided.

secondly, |publication-date= has obvious semantic significance for editors/citation providers. substitution with |date= should not be encouraged. instead |publication-date= should be promoted to the same status as |date= within the template when |date= is absent. similarly for online sources |accessdate= should be promoted when |date= and |publication-date= is absent. and so on for |archivedate=.

or, just do the lazy thing and remove |publication-date= altogether from CS1.

70.19.122.39 (talk) 00:19, 14 March 2013 (UTC)

Yes, it does not happen with |publication-date= but it is not a bug; it is a side-effect of the non-use of other parameters. Quite frankly I don't know why we attempt to extract a year from |publication-date= in the first place, since |publication-date= (see documentation) is intended to be supplementary to either |date= or |year= (see documentation).

Access dates must only be placed in the |accessdate= parameter (see documentation), no other parameter is intended for this.

The |ref= parameter (see documentation) is pretty much free-form. It is provided so that a manually-constructed anchor may be attached to the {{cite book}} when the special value |ref=harv does not give a suitable anchor. Use of the {{harvid}} template is not obligatory: it is a tool for constructing an anchor which is consistent with the links generated by templates like {{harv}}, {{harvnb}} and {{sfn}}. --Redrose64 (talk) 10:58, 14 March 2013 (UTC)

I need to fix the doc for this parameter. Somehow I thought the field did not display if year or date were not defined. --— Gadget850 (Ed) 12:46, 14 March 2013 (UTC)

No, that is how it works. If publication-date but not date, then publication-date is used as date and the anchor is formed:

This should work. Let me look at the original examples again. --— Gadget850 (Ed) 13:23, 14 March 2013 (UTC)

Here is what is happening: {{harvnb}} is including the year only in the link and {{cite book}} is including the full date in the anchor, causing a mismatch:

If you include the full date in {{harvnb}}, then it works:

At this point, we can debate whether to include the full date in {{harvnb}} or to change the cite templates to include only the year in the anchor. --— Gadget850 (Ed) 13:43, 14 March 2013 (UTC)

It would be a massive change to put a full date into {{harvnb}} and related templates; people are very much used to using the year alone. As things stand, the cite templates already include only the year in the anchor if you use |date=. It is only |publication-date= where there is inconsistency. See my post of 15:30, 13 March 2013

If we should build the harv anchor from |publication-date=, we should do so in the same manner as |date=. That is, instead of:

|Year={{{year|{{    <!-- attempt to derive year from date, if possible -->
           #if: {{{date|}}}
           |{{
              #iferror:{{#time:Y|{{{date|}}} }}
              |{{#iferror:{{#time:Y|{{{publication-date|einval}}} }}||{{#time:Y|{{{publication-date|}}} }}}}
              |{{#time:Y|{{{date|}}} }}
            }}
           |{{{publication-date|}}} <!-- last resort -->
         }}
      }}}

we should put

|Year={{{year|{{    <!-- attempt to derive year from date, if possible -->
           #if: {{{date|}}}
           |{{
              #iferror:{{#time:Y|{{{date|}}} }}
              |{{#iferror:{{#time:Y|{{{publication-date|einval}}} }}||{{#time:Y|{{{publication-date|}}} }}}}
              |{{#time:Y|{{{date|}}} }}
            }}
           |{{#iferror:{{#time:Y|{{{publication-date|}}} }}||{{#time:Y|{{{publication-date|}}} }}}} <!-- last resort -->
         }}
      }}}

--Redrose64 (talk) 14:05, 14 March 2013 (UTC)

I agree. I just got publication-date added to the Lua version, and it needs some more fixes including the anchor, so this is a good time for that. --— Gadget850 (Ed) 14:27, 14 March 2013 (UTC)

yes, this is a good solution imo, and goes some way into normalizing date output across CS1. i still have problems with the rationale imposing limits on the use of access date when sources are online (eventually ALL sources may be online). the doc reflects these limitations. imo both should be changed, but that's another issue. Redrose, i don't mean to nitpick, but when software does not perform as expected, eg forming anchors in unpredictable fashion, then it is buggy software. but that's also another matter. 70.19.122.39 (talk) 14:45, 14 March 2013 (UTC)

Now in {{cite book/sandbox}}:

This would break any instances where someone used a full date in a harv template, but I would expect this to be nonexistent or rare. Do need to update the anchor documentation to include publication-date. --— Gadget850 (Ed) 15:03, 14 March 2013 (UTC)

It would only break the harv linking if:

• the harv template had used a full date (the docs explicitly state that the year of publication should be given)
• and in {{cite book}} they had specified
  • |publication-date= with a full date
  • and |ref=harv
  • and had omitted |date= or left it blank
  • and had entirely omitted |year= (not just left it blank)

Pretty big set of ifs. --Redrose64 (talk) 15:50, 14 March 2013 (UTC)

I searched the archives and don't see this was every reported before, so I think it is pretty safe to go forward. --— Gadget850 (Ed) 16:21, 14 March 2013 (UTC)

Multiple ISBNs

Is there some way in cite book etc. to indicate multiple or alternative ISBNs e.g. paperback vs hardcover? Thanks. EEng (talk) 21:05, 16 March 2013 (UTC)

This is asked quite often; full answers are in the archives, but in brief: give the ISBN for the single edition/version which you actually consulted. --Redrose64 (talk) 21:37, 16 March 2013 (UTC)

That seems like a cop-out. Part of the purpose of ISBN is to allow the reader to find a copy he can consult for himself, so assuming that e.g. paperback and hardcover have identical content, it's useful to include both. Maybe there's a "misc info" parameter I can just stick the other ISBN into as free text? EEng (talk) 22:59, 16 March 2013 (UTC)

As a source we only need the one version - That said let me help - I do it like the following for bibliographies - not sure its ok but been doing it for 7 years - you dont need ISBN's in the {{cite book}} template itself to work - <ref>ISBN 978-0-8020-5016-8</ref> = ISBN 978-0-8020-5016-8

So this coding that has the ISBN's outside the {{cite book}} but with in the <ref> </ref> parameter will work as a reference.

• <ref>{{cite book |last = Taylor |first = Martin Brook|coauthor= Owram, Doug|year =1994|title =Canadian History|volume=|publisher= University of Toronto Press}} Hard cover pp 12, ISBN 978-0-8020-5016-8 and Paperback pp 24, ISBN 978-0-8020-2801-3</ref>
  

Will render this...

• Taylor, Martin Brook (1994). Canadian History. University of Toronto Press. {{cite book}}: Unknown parameter |coauthor= ignored (|author= suggested) (help) Hard cover pp 12, ISBN 978-0-8020-5016-8 and Paperback pp 24, ISBN 978-0-8020-2801-3

--Moxy (talk) 23:22, 16 March 2013 (UTC)

That's what I was doing -- was hoping for something more elegant. But thanks. EEng (talk) 23:30, 16 March 2013 (UTC)

Except that "pp" is the abbreviation for "pages", and since only one page number is given, it should be Hard cover p. 12, ISBN 978-0-8020-5016-8 and Paperback p. 24, ISBN 978-0-8020-2801-3 --Redrose64 (talk) 23:50, 16 March 2013 (UTC)

And that stuff outside the template will not render metadata in the new Lua templates as we migrate. --— Gadget850 (Ed) 05:06, 17 March 2013 (UTC)

Millions and millions of entries not in templates - any plans to address this?Moxy (talk) 05:30, 17 March 2013 (UTC)

No. ISBN is not required by any major citation style, and is not completely reliable as a unique identifier. We could provide the capability to add 16 different ISBNs, but it would not directly identify the particular edition used in creating the citation. --— Gadget850 (Ed) 13:43, 17 March 2013 (UTC)

Pardon my ignorance, but:

• In what way is ISBN not completely reliable as a unique identifier? If you're going to say that there are occasional glitches such as the same number issued twice, or multivolume works assigned a single number, then by that token we might as well not bother with any kind of identifier of anything.
• What does it matter that no major citation style requires ISBN? Here at WP it's recognized as a highly useful thing to include, so it's worth giving attention to making its inclusion as easy as possible.

EEng (talk) 15:07, 17 March 2013 (UTC)

Well, an editor should be specifying the exact edition of a work being referenced. The pagination and typesetting could differ between editions printed in the UK and and the US, for instance. The pagination normally differs between hardcover and paperback editions of the same edition of a book. An editor could be consulting a revised or updated edition as well. Each of these editions will have differing ISBNs, and the principle is to cite the actual source consulted. Imzadi 1979 → 15:58, 17 March 2013 (UTC)

I understand all that, and certainly any provision for multiple ISBNs should include some way to single out the particular ISBN that was consulted, because the core imperative is that the reader be certain he's looking at precisely the work to which the citation refers. It's still useful to include additional ISBNs, with the caveat that different ISBNs may be subject to the sorts of variation you mention. On the other hand, it's often the case the a publisher will explicitly state that two ISBNs are the same page images in different bindings, and it would be useful to be also able to designate those when that's known. EEng (talk) 17:19, 17 March 2013 (UTC)

Good— you have identified the issues with ISBN. JSTOR, doi and others use well verified systems with a central point of authority.

The purpose of a citation is to help the reader identify the source used to make a point in the content. If you have a pressing need to identify versions other than the one used, then put it in a further reading section. Alternate sources that were not consulted do not belong in references. --— Gadget850 (Ed) 22:38, 17 March 2013 (UTC)

Oh, honestly! Pressing need? It's amazing the amount of verbiage being wasted to characterize as illconceived something that's obviously useful. We've even had one person interject a correction about pp vs p, as if somehow anyone here needed that pointed out.

Let me state the situation again: Frequently a work is issued in paperback and hardcover, utterly identical except for the binding. Sometimes I'll have both on my bookshelf, so it doesn't make sense to talk about "the one actually consulted." It would be stupid to have two separate entries, one for the paperback, one for the hardcover. I could arbitrarily pick one or the other to use in the cite, but that unnecessarily reduces the ease with which the reader can find a copy for his own consultation. So if you can find it in your heart, please make some way accommodate the entirely sensible desire to have two or more ISBNs in the same entry. EEng (talk) 05:10, 18 March 2013 (UTC)

Error conditions

As far as I know, {{cite web}} is the only citation template that requires title= not be blank, e.g.

{{cite web | title= | url=http://www.foo.com/ | first=James | Last = Ford }} = Ford, James. http://www.foo.com/. {{cite web}}: Missing or empty |title= (help)

{{cite news | title= | url=http://www.foo.com/ | first=James | Last = Ford }} = Ford, James. http://www.foo.com/. {{cite news}}: Missing or empty |title= (help)

while at the same time cite web, for some reason, allows the URL to be omitted:

{{cite web | title=Bob | url= | first=James | Last = Ford }} = Ford, James. "Bob". {{cite web}}: Missing or empty |url= (help)

{{cite news | title=Bob | url= | first=James | Last = Ford }} = Ford, James. "Bob".

This situation makes no particular sense to me. Shouldn't something called "cite web" require a URL? Also, isn't the lack of a title generally a problematic omission in all cases? For example, using {{cite book}} without naming the book seems silly.

In working on the Lua migration of the citation templates, I'm wondering what to do about the error conditions. We can, of course, enforce a rule that {{cite web}} must have a title but not the other templates, if that is what's needed. But that seems rather arbitrary. Do people here have suggestions for what error checks you would like to see? Dragons flight (talk) 04:57, 17 March 2013 (UTC)

I find it sensible that when a URL is produced, there should be a forced title for all of the templates. cite web might be the special case where the title is expected to be an online one (not sure that's relevant).

Forcing a URL though? Not sure about that. Even if it is cite web. --Izno (talk) 13:39, 17 March 2013 (UTC)

Cite encylopedia - Lua

{{Cite encyclopedia}} now uses the Lua version. Please report any issues. --— Gadget850 (Ed) 14:48, 17 March 2013 (UTC)

{{Cite Nuttall}}, which relies on this template, is broken (editorlink param). Bob Burkhardt (talk) 22:08, 17 March 2013 (UTC)

I fixed it, but it was never a Lua issue. That template was broken due to an erroneous edit last June. Dragons flight (talk) 22:26, 17 March 2013 (UTC)

Thank you. Bob Burkhardt (talk) 01:49, 18 March 2013 (UTC)