Help talk:CS1 errors/Archive 1

This is an archive of past discussions about Help:CS1 errors. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 1

Archive 2

Archive 3

Archive 4

Unknown parameter

The "Unknown parameter |????= ignored" error would be useful in other templates. Is the code fairly short? It would also be quite useful if we could flag duplicated parameters. — kwami (talk) 01:56, 17 April 2013 (UTC)

It relies on the template being Lua-based. This is because with normal Wikicode-based templates, there is no means for testing parameters in a general manner. The presence or absence of any parameter has to be tested explicitly and separately. For example, if we know that |accessdate= is often mis-spelled as |acessdate= we can add a test like this:

{{#if: {{{acessdate|}}}|<span class=error>Error: {{para|acessdate}} is invalid</span>}}

but that will silently ignore other misspellings such as |accesdate= or |acesdate=

However, with Lua, it is possible to write code that reports on the use of parameters which have not been individually catered for. --Redrose64 (talk) 16:28, 17 April 2013 (UTC)

The code is in Module:Citation/CS1. -- Gadget850 (Ed)^talk 16:49, 17 April 2013 (UTC)

Essentially the process is that Lua provides a list of all parameters the template was given and you compare that with a list of all parameters that are supported (in this case provided by Module:Citation/CS1/Whitelist), and any parameter that was submitted but is not listed as supported generates an error warning. The code used here is somewhat specialized because Citation need to support endless lists of last1=, last2=, etc., but it wouldn't be that hard to make a general utility that does a whitelist check of parameter names. The only real issue is that people would have to keep and maintain such a whitelist of supported parameters for any template you wanted to check. Dragons flight (talk) 17:11, 17 April 2013 (UTC)

Thanks! That's a bit much for me to dive into right now, but I'll link this discussion from the template talk pages for future ref. — kwami (talk) 06:03, 21 April 2013 (UTC)

Section headings

I moved the comments in the headings to a new line, as they cause the section to not show in the edit summary. -- Gadget850 (Ed)^talk 12:38, 18 April 2013 (UTC)

I had placed the anchors at the front of the section title so that editors who clicked a citation error's help link would land on the help page at with a section header visible. I've now moved the anchors and their associated comments above the section header which accomplishes the same thing.

—Trappist the monk (talk) 13:07, 18 April 2013 (UTC)

When you edit a section, the section name wasn't being included in the edit summary. I thought it was the anchors at first, but it was the comments after the header markup. Nutshell: don't put any markup after the section header markup. -- Gadget850 (Ed)^talk 13:31, 18 April 2013 (UTC)

Seems like a bug that needs fixing. Anything inside  should be invisible.

—Trappist the monk (talk) 13:37, 18 April 2013 (UTC)

Breakage of the edit-summary pre-fill by trailing HTML comments has been a bug for at least a year now, and I'm pretty sure that there's a bugzilla. When I first noticed this problem, there was a related bug where a trailing space on the section header would also cause the pre-fill of the edit summary to fail; but that bug seems to have been fixed now. --Redrose64 (talk) 16:29, 18 April 2013 (UTC)

I checked before and just again and I did not see a bug logged. -- Gadget850 (Ed)^talk 18:20, 18 April 2013 (UTC)

displayauthors= suggested: how to keep 9?

I have a citation with nine author names, but I can check that there are more (in the source of course). Is there way to have the nine names published, and "et al." afterwards? As it is now, I'll have to drop the ninth to trigger the "et al." (by setting displayauthors=8. -DePiep (talk) 15:27, 21 April 2013 (UTC)

No. Add more authors. -- Gadget850 (Ed)^talk 15:32, 21 April 2013 (UTC)

I addedd |last10=N.N and |displayauthors=9. -DePiep (talk) 19:09, 21 April 2013 (UTC)

Forgive me for asking, but for what purpose? What does N.N mean? Will other editors understand what it means?

—Trappist the monk (talk) 19:28, 21 April 2013 (UTC)

We should have a parameter to disable the COinS output for when it is deliberately polluted. -- Gadget850 (Ed)^talk 19:35, 21 April 2013 (UTC)

Doesn't COinS need to know the "et al." situation? -DePiep (talk) 00:05, 22 April 2013 (UTC)

I found an example in this edit (I don't know if there are any others):

Fields, P.; Studier, M.; Diamond, H.; Mech, J.; Inghram, M.; Pyle, G.; Stevens, C.; Fried, S.; Manning, W.; N.N. (1956). "Transplutonium Elements in Thermonuclear Test Debris". Physical Review. 102 (1): 180. Bibcode:1956PhRv..102..180F. doi:10.1103/PhysRev.102.180. {{cite journal}}: Unknown parameter |displayauthors= ignored (|display-authors= suggested) (help)

Here is the same citation but omitting |last10=N.N.

Fields, P.; Studier, M.; Diamond, H.; Mech, J.; Inghram, M.; Pyle, G.; Stevens, C.; Fried, S.; Manning, W. (1956). "Transplutonium Elements in Thermonuclear Test Debris". Physical Review. 102 (1): 180. Bibcode:1956PhRv..102..180F. doi:10.1103/PhysRev.102.180. {{cite journal}}: Unknown parameter |displayauthors= ignored (|display-authors= suggested) (help)

It appears that the source has more than nine authors, but only nine are known, so DePiep is using |last10=N.N. to trigger the "et al." --Redrose64 (talk) 21:17, 21 April 2013 (UTC)

Hmm, follow the bibcode and doi links and you can find thirteen author names. So it would appear that the names are known and the editor has elected to omit them infavor of this mechanism. Might mean Nomen nescio but this use is sort of a misuse of the term, isn't it?

—Trappist the monk (talk) 21:46, 21 April 2013 (UTC)

As Redrose64 says, that was my hack. Actually I did not omit them, more like I maintained the omission (the earlier editor did). I do not see a need to add names above nine if they are not displayed; and I thought "N.N." comes close to correctness. Today I've done this about maybe two dozen times (in WP:ELEM).

Since I learn that it is a pollution, I'll stop it and I'll just add the known name #10 for the same effect: it is not shown and "et al." is added. -DePiep (talk) 23:46, 21 April 2013 (UTC)

Is there some reason not to just use all 13 authors?

Fields, P.; Studier, M.; Diamond, H.; Mech, J.; Inghram, M.; Pyle, G.; Stevens, C.; Fried, S.; Manning, W.; Ghiorso, A.; Thompson, S.; Higgins, G.; Seaborg, G. (1956). "Transplutonium Elements in Thermonuclear Test Debris". Physical Review. 102 (1): 180. Bibcode:1956PhRv..102..180F. doi:10.1103/PhysRev.102.180.

The new citations allow arbitrarily long lists of authors. The old citations were often truncated because they didn't support additional authors, and we can maintain that by using displayauthors=, but it seems preferable to add the remaining authors when possible. Dragons flight (talk) 00:29, 22 April 2013 (UTC)

The main reason for keeping just 9 names is that first I (we) want to clean out all "new" cite errors from some 300 WP:ELEMENT pages (task is tracked here). For now we only reproduce the pre-Lua cite situation without error message. Additionaly, the singular situation "nine authors, exactly, so don't add 'et al.'" now can be checked and corrected (done once so far). Apart from the |last10=N.N. hack, which is to be undone I agree, this process is not introducing new ambivalence, incorrectness or errors.

Adding another 1 to 20 names (and removing |displayauthors=) can be a next cite-improvement task. For this we can use the check for pages "|displayauthors=m where m < number of authors listed in the template". -DePiep (talk) 08:57, 22 April 2013 (UTC)

(edit conflict)

To your earlier question about COinS need[ing] to know the "et al." situation, I don't know if it needs to know. What it gets is not a function of |displayauthor=. If you set |display author=1, only the first author is rendered but all of the authors in the author list are included in the COinS data.

Have I answered your question?

—Trappist the monk (talk) 00:41, 22 April 2013 (UTC)

Dunno, but first let me repeat that my |last10=N.N. hack should be undone for the pulluting reason. Then, I don't know COinS, but indeed it should not bother with the displayed names, but directly with the list of entered names. When and if COinS needs the fact of "et al. list", it could be deducted from wikicode facts. For example, when |author=... et al., or when "|displayauthors=m where m < number of authors listed in the template". For this moment & place it seems enough to get the | message gone by adding correct info. coinS can follow. -DePiep (talk) 08:57, 22 April 2013 (UTC)

Behold the glorious new CS1: 33 authors listed (and this transcluded usefully 6 times). I have not limited the number of displays, but added a param to allow so individually, in the specific {{cite doi}}-link. -DePiep (talk) 21:34, 27 April 2013 (UTC)

New issue: how many authors to display? All 30 or 40 is reasonable? Project discussion here, general village pump question here, with example. -14:59, 28 April 2013 (UTC)

Fixed link above after archivingLeadSongDog come howl! 01:18, 3 June 2013 (UTC)

This error talkpage is hardly a suitable place for a MOS discussion, but there has never been a wide stylistic consensus on how many to display, and that is unlikely to change simply because new code now allows for the technical possibility of displaying more. The practice in different topic areas varies widely. In some disciplines, extensive lists of authors are wanted, while in others two or three is normal. No general rule is available. LeadSongDog come howl! 01:45, 3 June 2013 (UTC)

unusedurl= is useful and important, and should not be an error

Where should this be discussed, if not here? "unusedurl=" was an allowed formulation used to keep a record of previously working URLs which are now dead but which may return, or alternate URLs which could be used in the event of "url=" going dead. "unusedurl=" entities are "backroom" notes to other editors, without the ugly necessity of . They are not intended as "archive" urls, which is why they are not used in "archiveurl=". They are alternates. These are the result of sometimes very difficult spelunking work by editors, and should not be flagged as errors. If another name is preferred, I nominate "alternateurl=". It's great that Lua has allowed improved error handling, but the definition of "what is an error" should be discussed, and not decreed by the developers. --Lexein (talk) 03:15, 22 April 2013 (UTC)

When was this allowed? This parameter name doesn't appear in any of the documentation for Citation Style 1. unusedurl= is used in only a handful of pages, at least according to this search result.

It seems to me that the proper way to identify alternate urls is to place them in a section of the article's talk page rather than in the article itself.

—Trappist the monk (talk) 11:26, 22 April 2013 (UTC)

It was recommended by other editors for the reasons I list above. In Talk pages, such notes will disappear into the talk page archives, effectively lost. I'm against lost work. --Lexein (talk) 15:20, 22 April 2013 (UTC)

The older templates did little error checking, thus an undefined parameter like 'foo' would not give an error, but it was never parsed. 'unusedurl' was never defined, thus the Lua templates never supported it. -- Gadget850 (Ed)^talk 11:34, 22 April 2013 (UTC)

True, but unhelpful. Hardcoding against slack is similar to "that which is not specifically permitted is prohibited," which is not the Wikipedia way. This goes against the Five Pillars, which explicitly states that there are no hard and fast rules. That syntax can be rigidly checked and enforced doesn't mean it necessarily should be. I'm seeing developer zeal trumping the principles of the project. Slack is good. --Lexein (talk) 15:20, 22 April 2013 (UTC)

So, we should allow typos such as 'newpaper' to silently fail without showing the field or an error? Or, for editors to use 'day of week' copied from a different template without alerting them that it isn't supported in these templates?

You asked why this parameter did not work, and we answered. In your long opening, was there a specific proposal to add a new parameter? -- Gadget850 (Ed)^talk 16:08, 22 April 2013 (UTC)

Checking for unknown parameters has caught tens of thousands of citations that were malformed or incomplete due to mistakes or typos in usage. It has also caught roughly several hundred cases where people, like you, were intentionally using undocumented parameters as a way to include hidden comments. Frankly, that's a trade off that I have no problem with. If you wanted to include hidden comments, we have a specific syntax for hidden comments, e.g . And having to convert one use of an intentionally blank parameter to an actual hidden comment in order to find and correct several dozen malformed citations, doesn't seem like a bad deal to me. Asking editors to use hidden comments for material like this also has the potential to be clearer than undocumented forms like |unusedurl=, since the person adding such a comment is hopefully more likely to explain why the url is unused (e.g. dead link, alternative source that is the same as |url=), rather than leaving people to guess based on the choice of parameter name. However, if you want to make a case for adding one (or even several) intentionally unused parameters for data storage then you are free to make that case, and consensus will prevail. I'd suggest choosing a more popular forum than this one (e.g. Help talk:Citation Style 1 or one of the village pumps), if you want to make that argument. Dragons flight (talk) 16:13, 22 April 2013 (UTC)

(edit conflict) Glad that true malformations or typos were caught: that's helpful. But booting "unofficial" but correctly spelled parameters goes too far, and as I said, goes against the Five Pillars. You understand, it's simply a kind of marshal law imposed from outside the usual process of discretion and discussion, and it's a way for developers who are editors to have their way (finally (whew!)), since it will all be hardcoded, done and dusted.

You invite me to a battle which I do not want, forced upon me. Now I'm suddenly expected to defend a position which never needed defending in the first place? Now, just because somebody else suddenly wants to play offense against the Five Pillars? Now, I have to petition to keep what has been working all along? To attempt to stem the tide of obsessive, and on this, pointless and destructive tidying? Coding obsessively for obsessive coding's sake is not the point of Wikipedia. It's not a playground for programmers. The performance improvements offered by Lua were not intended to immediately soaked up by nitpick features: that's the Microsoft (sorry, MS, you know it's true) way, not the Wikipedia way. And Lua isn't a free pass to mess with editors who were perking along quite happily, productively, and errorlessly, for seven years before people who like making rules decided, "Ah, now we can code more rules, and declare sensible editing to be errors."

What's happening here will result in a reduction of use of citation templates IMHO. People will see all that red, and just convert to plain text. Too bad for all the benefits offered by the template. So sorry.

As I said above, If another name is preferred, I nominate "alternateurl="

As for hidden comments: in my experience, they have very curious habit of being deleted by a variety of disrespecters of the work of others, who "don't see the need". Not a good solution, IMHO.

Finally, "move the discussion"? And be falsely accused (again) of forum shopping? Perhaps not. --Lexein (talk) 18:01, 22 April 2013 (UTC)

I don't see that I can be of help here. Sorry. -- Gadget850 (Ed)^talk 18:04, 22 April 2013 (UTC)

Further help

I would suggest adding a prominent mention of Wikipedia:Help desk and / or Wikipedia:Teahouse to give people somewhere else to ask for help if they aren't able to figure out what to do from this page alone. Dragons flight (talk) 17:50, 22 April 2013 (UTC)

Help desk

I've added links to the Help desk on the right margin at approximately the line that begins "To resolve ...". I decided against {{shortcut}} because the shortcut box seemed to large when viewed in the relatively small sections of the help page.

Thoughts?

—Trappist the monk (talk) 22:12, 22 April 2013 (UTC)

My thoughts? I'd like to note first that these "new" red cite text errors are a great improvement to WP, and applied in a smart way. I enjoy cleaning them. I write working on WP:ELEMENTS, this tracking (a serious topic in references I say). My secundary thought is, that it is difficult to connect with (to undestand) the Cite requirements and logic. My reading & understanding of the issue is: it is tough, but we can do it. A rewrite of Help_talk:CS1_errors I could be happy with -- but how and in what way? Best improvement could be to have someone ouside rewrite theis errortext. Really, I am smart but I cannot follow all of the texts. -DePiep (talk) 22:56, 22 April 2013 (UTC)

Whenever you find anything in Help:CS1 errors that is confusing or doesn't explain something completely, please point that out (or just fix it).

—Trappist the monk (talk) 12:54, 23 April 2013 (UTC)

I did just that to the Wikilink embedded in URL title section, and you modified my edits, making them less helpful. I have cleaned about 400 of these pages in the last few days and have some hard-earned tips to offer about what specific situations can cause this error and how to best resolve them. If my instructions are going to be removed from this page, is there a better place to post them? Thanks. Jonesey95 (talk) 22:04, 24 April 2013 (UTC)

I wasn't trying to step on your toes. The help text is more-or-less organized to list possible causes and then discuss solutions. It is also intended to be general in nature and focused on getting the editor to a reason for, and a solution to, the problem. I took the opportunity of your edit to revise that whole section so that it, I think, better serves the goal of getting an editor to the solution. It isn't just {{lang}} templates that cause the error, it's pretty much any template that outputs some sort of wikilinked text. In keeping with the general nature of the help text, I didn't refer to {{lang}} or any other template.

But you are making me wonder, is there a need for a related topics area? I have no idea how to get my brain around that nor how it might be implemented.

—Trappist the monk (talk) 22:52, 24 April 2013 (UTC)

Pseudoreplication appears in Category:Pages with citations lacking titles

Can anyone figure out why Pseudoreplication appears in Category:Pages with citations lacking titles, even though the article has not been edited in a while and all citations seems reasonable. The cat is not listed on the article page, but the article does appear in the cat. Thanks, Illia Connell (talk) 21:14, 24 April 2013 (UTC)

See Pseudoreplication#cite_note-2 - it shows Citation has no title --Redrose64 (talk) 21:26, 24 April 2013 (UTC)

Thanks for your quick reply. However, in my browser (FF 20.0.1, on Windows 7), it does not show the error message. Ref #2 is:

<ref>Millar, R.B., Anderson, M.R. 2004.  Remedies for pseudoreplication. Fisheries Research 397-407{{cite journal|doi = 10.1016/j.fishres.2004.08.016}}</ref>

, which is obviously a mix a citation styles. Regards, Illia Connell (talk) 21:35, 24 April 2013 (UTC)

As advised at Category:Pages with citations lacking titles: By default, Citation Style 1 and Citation Style 2 error messages are visible to all readers and maintenance messages are hidden from all readers.

To display maintenance messages in the rendered article, include the following text in your common CSS page (common.css) or your specific skin's CSS page and (skin.css).

(Note to new editors: those CSS pages are specific to you, and control your view of pages, by adding to your user account's CSS code. If you have not yet created such a page, then clicking one of the .css links above will yield a page that starts "Wikipedia does not have a user page with this exact name." Click the "Start the User:username/filename page" link, paste the text below, save the page, follow the instructions at the bottom of the new page on bypassing your browser's cache, and finally, in order to see the previously hidden maintenance messages, refresh the page you were editing earlier.)

:root .mw-parser-output .cs1-maint {display: inline;} /* display Citation Style 1 maintenance messages */

To display hidden-by-default error messages:

:root .mw-parser-output .cs1-hidden-error {display: inline;} /* display hidden Citation Style 1 error messages */

Even with this CSS installed, older pages in Wikipedia's cache may not have been updated to show these error messages even though the page is listed in one of the tracking categories. A null edit will resolve that issue.

After (error and/maintenance) messages are displayed, it might still not be easy to find them in a large article with a lot of citations. Messages can then be found by searching (with Ctrl-F) for "(help)" or "cs1".

To hide normally-displayed error messages:

:root .mw-parser-output .cs1-visible-error {display: none;} /* hide Citation Style 1 error messages */

You can personalize the display of these messages (such as changing the color), but you will need to ask someone who knows CSS or at the technical village pump if you do not understand how.

Nota bene: these CSS rules are not obeyed by Navigation popups. They also do not hide script warning messages in the Preview box that begin with "This is only a preview; your changes have not yet been saved".

--Redrose64 (talk) 21:59, 24 April 2013 (UTC)

And Category:Pages with citations lacking titles is a hidden category. Preferences → Appearance → Show hidden categories -- Gadget850^talk 22:04, 24 April 2013 (UTC)

Got it - updating my common.css fixed the problem - thanks to both for your speedy help. Regards, Illia Connell (talk) 22:14, 24 April 2013 (UTC)

Pages in the User, User talk, and Wikipedia talk namespaces are not included in the error tracking categories.

The note at the top of Category:Pages with citations lacking titles states: "Pages in the User, User talk, and Wikipedia talk namespaces are not included in the error tracking categories." However, these are in the cat: User talk:Pzrmd/Aagots vei, User talk:Pzrmd/Aasta Hansteens vei (Oslo), User talk:Pzrmd/Abbediengen terrasse, User talk:Pzrmd/Abbediengveien. Can anyone figure out why? Thanks, Illia Connell (talk) 21:22, 24 April 2013 (UTC)

Probably a bug in Module:Citation/CS1 or Module:Citation/CS1/Configuration. Suggest that you raise a thread at Module talk:Citation/CS1. --Redrose64 (talk) 21:32, 24 April 2013 (UTC)

Thanks, will do. Illia Connell (talk) 21:36, 24 April 2013 (UTC)

2009 NCAA Men's Basketball All-Americans appears in Category:Pages with citations having wikilinks embedded in URL titles

I'm having a similar problem to the one two sections above. 2009 NCAA Men's Basketball All-Americans appears (correctly) in Category:Pages with citations having wikilinks embedded in URL titles, but there is no red error message in the References. I believe that Reference number 40 is causing the problem (I have left it unchanged for now). I have fixed this error message in a few hundred pages over the last couple of days, and I have seen this problem in a small handful of the pages, maybe 3-5 out of the ones I have fixed. I have made the change to my CSS file recommended above, even though these wikilink errors show 99% of the time. Jonesey95 (talk) 22:11, 24 April 2013 (UTC)

Strike that. I re-read the info about the lack of error message and found the part about the null edit. Doing a null edit fixed this article. Jonesey95 (talk) 23:14, 24 April 2013 (UTC)

Wikilink embedded in URL title - overridable?

Is there any way to override the "Wikilink embedded in URL title" warning for cases where the title of the referenced webpage really contains pairs of brackets? I presume this would require encoding the brackets as the corresponding HTML character entities; am I understanding correctly? Should this be mentioned on the help page itself?

(Sorry I don't have a specific example handy; I will keep looking for one.) --SoledadKabocha (talk) 00:31, 30 April 2013 (UTC)

Yep, encode with [ ([) and ] (]).

{{cite web |url=http://example.com |title=[[Title in Brackets]]}}

→"[[Title in Brackets]]".

—Trappist the monk (talk) 01:13, 30 April 2013 (UTC)

Wrong template, right parameter

If {{cite}} is used by mistake when {{cite needed}} was intended, and there are no parameters, the page ends up in Category:Pages with empty citations which has an explanation at the top and describes a suitable fix. The same happens if there are only positional parameters, as in {{cite|May 2013}}. However, if an appropriate named parameter is used, i.e. {{cite|date=May 2013}}, the page is placed in Category:Pages with citations lacking titles which does not describe a suitable fix.

I think that either of two things should be done to cover such uses:

when the only named parameters are those respected by {{citation needed}}, i.e. |name= |reason= |date= categorise to Category:Pages with empty citations
add a suitable explanation and fix to Category:Pages with citations lacking titles

--Redrose64 (talk) 13:14, 17 May 2013 (UTC)

Citation without a title of any form and Empty citation tweaked. Is that better?

—Trappist the monk (talk) 13:03, 19 May 2013 (UTC)

Yes, I think so --Redrose64 (talk) 14:22, 19 May 2013 (UTC)

Cite episode; suggestion related to transcripturl= and transcript=

Regarding Cite episode: if transcripturl= is used without transcript=, maybe display a red-colored error. Without transcript=, transcripturl= will not be shown, and the software does not yet warn the editor about this. --82.170.113.123 (talk) 00:03, 25 May 2013 (UTC)

Correct ISBN gives error

Forensic Psychology and Law. ISBN 8387425313. {{cite book}}: Check |isbn= value: checksum (help)

Something's wrong here, because this ISBN is valid. --bender235 (talk) 20:05, 13 June 2013 (UTC)

Both Abebooks and isbn.org agree that this ISBN is malformed. According to the documentation, you can suppress the error message by adding |ignore-isbn-error=true. -- John of Reading (talk) 20:39, 13 June 2013 (UTC)

When confronted with an invalid ISBN, I assume that it's the last digit that is in error. By my calculations, ISBN 8-38742531-1 or ISBN 978-8-38742531-9 should be correct; but neither of those turn up any search results. --Redrose64 (talk) 20:57, 13 June 2013 (UTC)

(edit conflict)Yep, something is wrong. The isbn number is wrong. Try these: [1], [2] or any of the links at Special:BookSources/8387425313. Only Worldcat seems to think that 8387425313 is a valid isbn.

Can you confirm that 8387425313 matches the printed copy of your source?

Assuming that all of the other digits are correct, the checkdigit should be a 1, not a 3. You should only use |ignore-isbn-error=true if you can show that the flawed ISBN is truly correct. If you do use it, for the benefit of editors who follow you, leave a hidden note explaining why you did.

—Trappist the monk (talk) 21:01, 13 June 2013 (UTC)

I'm sorry, I do not have a printed copy in front of me. I only checked the source via Worldcat and came accross this error. I'll add the ignore-isbn-error paramter plus a note. --bender235 (talk) 23:48, 13 June 2013 (UTC)

Some pages I've found:

Google Books has an entry for the right book, but there is nothing resembling an ISBN;
Libris (which is in Swedish) also has an entry, containing the item "Ogiltigt nummer / annan version:85-87425-31-3", which resembles the WorldCat number (except for the second digit), but again as an ISBN it's not valid: ISBN 85-87425-31-3 "Ogiltigt nummer" means "Invalid number" or "Void number", which might be significant.
An entry at Wirtualny Gabinet Rycin-Rekord MARC has an item "020 \z 8387425313"
There are plenty of academic papers which have references containing text like "[in:] Forensic Psychology and Law: Traditional Quetions and New Ideas, A. Czerederecka, T. Jaśkiewicz - Obydzińska, J. Wójcikiewicz [eds], Institute of Forensic Research Publishers, Kraków 2000", but none of these give an ISBN.

The common theme is that the acronym ISBN is never used, not even by WorldCat. I rather suspect that 8387425313 (or 85-87425-31-3) is not an ISBN at all, but some other identifier which just happens to have ten digits. --Redrose64 (talk) 08:09, 14 June 2013 (UTC)

In which article(s) this book is mentioned? We need to see who introduced the number in WP. Even OP Bender235 says: I only checked the source via Worldcat, so did not find it originally. -DePiep (talk) 10:00, 14 June 2013 (UTC)

Facial composite is the one where this "ISBN" is given, but the publication is also used in E-FIT. --Redrose64 (talk) 10:29, 14 June 2013 (UTC)

The number was entered by ... Bender235 [3] OP !? Since the number is nowhere defined as an ISBN, as Redrose noted, I we can conclude that it is not an ISBN (especially since it would be a wrong ISBN). I propose deleting the |ISBN=8387425313. -DePiep (talk) 19:56, 14 June 2013 (UTC)

Done [4] -DePiep (talk) 08:53, 17 June 2013 (UTC)

Of course I was the one introducing the numnber. I was doing WP:WCC work, and WorldCat gave me this number. Anyhow, I will check whether the actual book has this (or any) ISBN this week at Berlin State Library, and then I'll reply here. --bender235 (talk) 12:15, 17 June 2013 (UTC)

I do not doubt your intention. I was just tying to backtrack the origin of the number entered. I understand you concur with the conclusion (not an isbn) here? -DePiep (talk) 13:50, 17 June 2013 (UTC)

Been at State Library. ISBN is printed as such in the book. And the same number is on the back: ISBN 83-87425-31-3. So we keep it now? --bender235 (talk) 15:28, 18 June 2013 (UTC)

Yes. I reinstalled the isbn and the error-suppression [5]. -DePiep (talk) 15:43, 18 June 2013 (UTC)

Cite episode error notes

I've noticed a few things and would like to leave it to someone more knowledgeable and skilled to address them:

1. There is no section on this Help page for "Error: |url= requires |title= when using {{cite episode}}: Empty citation (help)" and similar (but not identical) errors related to the Cite episode template.

2. Articles with these Cite episode errors are currently located in Category:Articles with incorrect citation syntax. They should probably be put in their own subcategory.

3. There is some confusion described at Help_talk:Citation_Style_1#Confusing about these Cite episode errors. This confusion could be addressed in text created to resolve items 1 and 2 above.

Let me know if further clarification is required. Jonesey95 (talk) 21:05, 16 July 2013 (UTC)

Errors listed at Help:CS1 errors are associated with errors detected by the Lua module version of the CS1 citation templates. {{cite episode}} has not been converted from {{Citation/core}} so any error messages produced by {{cite episode}} aren't (and shouldn't) be part of Help:CS1 errors. The error messages that you see arose from a rather long discussion related to the deprecation of certain parameters.

Within the code for {{cite episode}} is a call to {{Citation error}} which specifies the categories.

So, Help:CS1 errors is not the place to address the {{cite episode}} error messages until it is converted to use Module:Citation/CS1. Until such time, the best place to address these error messages may be in the template's documentation page. When converted to Lua, {{cite episode}} will likely categorize errors in a more appropriate way.

Trappist the monk (talk) 21:55, 16 July 2013 (UTC)

Your first sentence is news to me. The lead on the Help page did not state that the page described only Lua-related errors. Cite episode appears to be a CS1 template, so I would have thought that an error generated by this CS1 template would be documented on this page, which is called "Help:CS1 errors". I added a clarification to the lead sentence. If I missed some nuance, feel free to correct it. Thanks. Jonesey95 (talk) 04:43, 17 July 2013 (UTC)

|format requires |url generated on cite journal templates with valid doi or pmid links

The "|format requires |url" error is being generated by cite journal citations with valid doi links.

Example: Template:cite doi/10.1093.2Fjhered.2Fesm035, which renders as:

Template:Cite doi/10.1093.2Fjhered.2Fesm035

The above example has a valid url embedded in the doi link, so |format should be allowed.

Yes, you can use "type" as a workaround, but (1) the "cite journal" documentation does not list |url as a prerequisite for |format, and (2) the doi link is a url.

The above problem also applies to cite journal citations with valid pmid links, e.g. Template:Cite pmid/2593988, which renders as:

Template:Cite pmid/2593988

Can this be fixed in the Lua code? Jonesey95 (talk) 18:49, 19 July 2013 (UTC)

|format= is being misused in these citations. Its purpose is to identify the file format of the digital resource provided in |url=: pdf, xls, etc (except html). The error message is correct because the citations underlying the doi and pmid templates don't have values assigned to |url=.

These shortcut citations can be fixed. In the first case simply edit the doi template to include the url of the full text (available at the doi link) and remove |format= or change it to |type=. In the second case, edit the underlying pmid template: |format= should be removed or changed to |type= – though the full article is not available for free so the addition of |subscription=yes is appropriate.

–Trappist the monk (talk) 19:30, 19 July 2013 (UTC)

Understood. Thanks for the clarification.

FWIW, my understanding is that one benefit of the doi is that it is supposed to be an improvement over using the |url parameter, since the doi is "permanent", while the url can change (that said, I have been cleaning up a bunch of citations with dois that appear to have once been valid but are now dead....). Putting in a linkrot-prone url when the doi serves as a better url seems like a step backwards. Jonesey95 (talk) 20:59, 19 July 2013 (UTC)

accessdate= requires url=? What if it has pmid= or doi=?

See Help:CS1_errors#accessdate_missing_url.

Schüffler A, Sterner O, Anke H (2007). "Cytotoxic alpha-pyrones from Xylaria hypoxylon". Z. Naturforsch., C, J. Biosci. 62 (3–4): 169–72. PMID 17542480. {{cite journal}}: |access-date= requires |url= (help)CS1 maint: multiple names: authors list (link)
Liu Q, Wang H, Ng TB (2006). "First report of a xylose-specific lectin with potent hemagglutinating, antiproliferative and anti-mitogenic activities from a wild ascomycete mushroom". Biochim. Biophys. Acta. 1760 (12): 1914–9. doi:10.1016/j.bbagen.2006.07.010. PMID 16952421. {{cite journal}}: |access-date= requires |url= (help); Unknown parameter |month= ignored (help)CS1 maint: multiple names: authors list (link)

Some references have links in the pmid= or doi= parameters, but the url= is empty. This means that accessdate= is appropriate, but it still generates an error?? I have sometimes seen broken DOI links (in an old Nature News source), so there is still a purpose for accessdate=.

This error shouldn't be displayed when doi= or pmid= are being used to generate the external link.

--Enric Naval (talk) 11:27, 29 August 2013 (UTC)

|accessdate= is used to identify the point in time that an editor consulted an on-line source, specifically those sources that are not dated and in the way of the constantly changing web, often ephemeral. Journals, newspapers, books, magazines, etc exist outside the web and are dated, serialized, cataloged. An |accessdate= does not add any value to the citation of these types of sources.

—Trappist the monk (talk) 12:47, 29 August 2013 (UTC)

I concur- PMID and doi sources have a publication date, thus an access date is not needed. Per the documentation:

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 or linked documents that do not change; mainly of use for web pages that change frequently or have no publication date.

-- Gadget850^talk 17:16, 29 August 2013 (UTC)

This has utterly screwed quite a few hundred references I have added. I understand the reasons above, makes sense, but how the heck is that going to get fixed? Will there be a bot? As it stood I was following the citation template to the letter (albeit apparently putting in more information than was needed) and now everything is screwed. Panyd^{The muffin is not subtle} 21:39, 4 September 2013 (UTC)

We've been saying for well over a year on this page and elsewhere (including the doc pages for the various templates) that accessdates are redundant and unnecessary when there is no url. It's not like it's new news. --Redrose64 (talk) 21:56, 4 September 2013 (UTC)

Going to every article I've added references to and seeing a giant red error message is news. It wasn't there before and I've just seen Category:Pages using citations with accessdate and no URL - you cannot possibly be telling me that it's acceptable to have that many pages with glaring error messages without some sort of bot intervention. Panyd^{The muffin is not subtle} 13:49, 5 September 2013 (UTC)

Category:Pages using citations with accessdate and no URL was created five months ago, and it began to be populated straight away. That there are so many pages in it shows the lack of awareness, not necessarily a lack of desire to clean them up. Personally I have been cleaning up some of the other subcategories of Category:Articles with incorrect citation syntax - for example, when I started on Category:Pages with empty citations it was well over 1000. Sending a bot in will not mean that all instances are handled correctly.

I'm certain that some of the pages in Category:Pages using citations with accessdate and no URL are there because the magic characters url= are missing (in which case the page will also be in Category:Pages with citations using unnamed parameters). I'm also willing to bet that some of the pages in Category:Pages using citations with accessdate and no URL are there because some vandal has blanked or removed the |url= parameter, or has altered it in some way so that it's no longer recognised for what it should be. Should the vandalism be reverted, or should the vandalism be compounded by a bot making what it thinks is the correct fix but which actually removes even more good info? --Redrose64 (talk) 16:59, 5 September 2013 (UTC)

A bot could comment out the accessdate, or comment it only if the url parameter has a value. If the url parameter is empty or missing, I believe that the article will show up in Category:Pages using web citations with no URL, so an editor working on that category will be able to look in the article's history to see if a vandal has wiped the URL. – Jonesey95 (talk)

They only show in Category:Pages using web citations with no URL if they use {{cite web}}, for which a missing URL is always an error (what's the good of citing a web page if you don't indicate which web page?). If a template where the URL is not mandatory but optional (e.g. {{cite news}}, {{cite journal}}, {{cite book}}, etc.) is used with exactly the same parameters, it still goes into Category:Pages using citations with accessdate and no URL but does not go into Category:Pages using web citations with no URL. --Redrose64 (talk) 19:59, 5 September 2013 (UTC)

I still think a bot should be put into place before the error messages. You're right though, does show an extraordinary lack of knowledge that this was coming. How are we to fix this if it happens again? And shouldn't {{cite book}} be allowed to have an access date as well as a url? If a new edition of the book comes out that drastically changes the original content, that may well be a very important piece of information when it comes to verifying what's in the article. Panyd^{The muffin is not subtle} 15:11, 6 September 2013 (UTC)

Books have publication dates, edition numbers, and isbns to distinguish one version from another. Even if you consulted the book at google books, you are still consulting the book, not an on-line source – yeah, I know, that seems counter-intuitive, but you really are consulting the book. So, that new version with radically different content, is clearly identifiable by using those parameters. |accessdate= should not be used with {{cite book}}.

—Trappist the monk (talk) 15:42, 6 September 2013 (UTC)

An |accessdate= may be used with {{cite book}} provided that there is also a |url=. In this respect, {{cite book}} is no different from any of the other templates. --Redrose64 (talk) 16:47, 6 September 2013 (UTC)

Error display with newline

See oldid: this error message seem to be picking up newlines as part of the invalid text, hence leading to a display error. It seems to me any leading/trailing whitespace should not be shown in the error text, in the same way that it's not part of a regular parameter value. Thanks Rjwilmsi 16:16, 1 September 2013 (UTC)

Ah, but the text in question was part of a positional parameter, where leading and trailing whitespace is significant. It's only named parameters that do not treat leading/trailing whitespace as part of a parameter value. --Redrose64 (talk) 18:21, 1 September 2013 (UTC)

The {{sic}} template has an extra parameter "nolink=y" so that it can be used in text that is going to be used as a link, for example in the "title" parameter of {{cite web}}. There is also a "?" option that outputs a "?" and adds a tracking category, Category:Articles containing possible transcription errors. I've used this very many times to flag up typographic errors in book/webpage titles, in cases where I can't verify the original title.

This is no longer working. The category is displayed as plain text, with the error message "Wikilink embedded in URL title" - Example. How is this to be made to work again?

The solution adopted by Gilo1969 (talk · contribs) in this edit can't be the best one, since it lost the "?", the category, and the "reason" parameter.

One possibility is to modify the {{sic}} template so that it no longer outputs a category. Another is to modify the LUA code so that category links are treated differently from other links. We can then begin to find and revert all the edits that have inappropriately removed a {{sic|?}} template. -- John of Reading (talk) 08:00, 17 September 2013 (UTC)

It's complicated. {{sic}} isn't simply a visual indicator of typos and other anomalies; it adds text and characters to the title. The text and characters are added to the title before Module:Citation/CS1 sees it. As a result, CS1 can't know that the additional text and characters aren't supposed to be part of the title.

I suspect that it's possible to detect and move category wikilinks embedded in a title; the rest of whatever a template adds to the title will remain. The template-modified title is then made part of the COinS metadata which is used by referencing tools outside of Wikipedia. It seems inappropriate to me for a title in COinS metadata to be different from the original with the possible exception of silent corrections – like the silent removal of the second the from your example article (taking MOS:QUOTE as a guide).

I don't know if templates embedded in |title= can be made to work.

—Trappist the monk (talk) 12:47, 17 September 2013 (UTC)

Titles of works or articles should be given exactly as shown in the source. There should be no need to use {{sic}} at all. --Redrose64 (talk) 13:18, 17 September 2013 (UTC)

The variant I've used more often in titles is [sic?], which is for the case where the source is offline or is a dead link. The title might be incorrect in the original, or the problem might just be a copying error. [sic?] tells the reader that there may be a problem with the stated title, and it tells potential editors to check the source if they have an accessible copy. I believe this is useful, both for the readers and the potential editors. -- John of Reading (talk) 13:43, 17 September 2013 (UTC)

I think the nolink=y option is working for me in title parameters to fix the "wikilink in title" error. I've applied it a couple of times since learning about it. I browsed through my recent Contributions and couldn't find an example to verify it, though. I checked John of Reading's Example link above, and I don't see the Wikilink category.

To respond to Trappist the monk's comment about templates in title parameters, they work fine for me as long as they do not generate categories. I have been able to use {{sic}} with nolink=y and also {{!}} with no troubles. – Jonesey95 (talk) 18:47, 17 September 2013 (UTC)

Template:Sic has been edited so that it does not output a category for the time being. This fixes the immediate problem, in that the editors working on Category:Pages with citations having wikilinks embedded in URL titles won't remove any more [sic?]'s in the short term.

If the meta-data is not supposed to contain "[sic]" or "[sic?]", then could that be handled by the CS1 module somehow? There must be a large number of titles out there containing these, not necessarily using Template:sic. -- John of Reading (talk) 19:19, 17 September 2013 (UTC)

The metadata doesn't just contain "[sic]" or "[sic?]", it contains Wikicode, albeit in percent-encoded form: %26%2332%3B%26%2391%3B%27%27sic%3F%27%27%26%2393%3B which decodes to  [''sic?''] The   is a space; the [ and ] are the square brackets; but note also the four apostrophes. --Redrose64 (talk) 19:59, 17 September 2013 (UTC)

I'm sure someone could come up with a regex that would match all the likely variants. -- John of Reading (talk) 20:07, 17 September 2013 (UTC)

I guess I'm not sure that editing {{sic}} is beneficial. While it's true that editors working in Category:Pages with citations having wikilinks embedded in URL titles won't be seeing and removing the template from citations that include it, similarly, editors working in Category:Articles containing possible transcription errors will soon be seeing the contents of that category disappearing. My sense is that {{sic}} should be restored and that some discussion should take place to determine just what should be done about the issue.

First step in that, I think, is to determine if titles in citations should be 1) copied verbatim from the source without editorial markup; 2) copied verbatim from the source with editorial markup where appropriate; 3) copied from the source but silently edited to fix spellings, remove duplicate words, etc. I don't think that this talk page is the correct venue for that discussion. When that discussion has taken place, then this conversation should resume but at Help talk:Citation Style 1 rather than here.

—Trappist the monk (talk) 21:00, 17 September 2013 (UTC)

Q

If my browser can read Google Books URLs without the "http://" part in front of it, why can't we? These little red url notifications are irritating. Thank you. Drmies (talk) 17:07, 27 September 2013 (UTC)

I don't know the answer to your question, but can you please provide an example of a page where you are seeing url errors? After considerable effort by a small number of editors, there are currently zero Article-space articles in Category:Pages with URL errors (a figure that is down from many thousands when the category was created).

You might take your question to Wikipedia:Village pump (technical). There are lots of folks there who know answers to questions like yours. – Jonesey95 (talk) 17:38, 27 September 2013 (UTC)

Look at this version of an article. Drmies (talk) 18:56, 27 September 2013 (UTC)

Thanks for the link. I completely understand your frustration. There should be a way to make it so that URLs just work. The folks at the Village Pump may be able to provide you with an answer to the "why" question.

And for what it's worth, yes, the red errors are irritating, but irritations sometimes produce beautiful things. This particular red error led you to fix a non-working link. Thank you for being a conscientious editor and actually looking at the results of your edit (not everyone does). – Jonesey95 (talk) 23:53, 27 September 2013 (UTC)

Because what Module:Citation/CS1 does is convert the values provided by |url= and |title= into an external link in wiki markup that looks like this:

[<url> <title>]

which is then passed on to the wiki parser.

An external link without the url scheme, regardless of whether it is hand made or created with a template, is not seen by the parser as a valid external link. The parser won't tell you that the link is broken, but CS1 will:

[www.example.com title] → [www.example.com title]

{{cite web |url=www.example.com |title=Title}} → [www.example.com "Title"]. {{cite web}}: Check |url= value (help)

—Trappist the monk (talk) 17:42, 27 September 2013 (UTC)

Which brings me to the real question: why? Or, the realer question, when can you have this fixed? Thanks, Drmies (talk) 18:56, 27 September 2013 (UTC)

I think you'll have to seek the answers to those questions elsewhere. Issues related to the wiki parser are outside the scope of CS1 and this talk page. Perhaps WP:VPT is a good place to start.

—Trappist the monk (talk) 19:17, 27 September 2013 (UTC)

Originally, the http:// or https:// was mandatory for all URLs, whether those be bare (as in http://www.example.com) or enclosed in single square brackets (as in [http://www.example.com/ Example] → Example). About two years ago, the http: or https: became optional for the single square bracket form (as in [//www.example.com/ Example] → Example) but the double slash remained mandatory for all URLs: the [// sequence is one of the methods by which the MediaWiki parser recognises a URL as compared to plain text. Note that the http: or https: remains mandatory for bare URLs. Templates like {{cite book}} are essentially the single square bracket form in disguise; you can verify that by taking a {{cite book}} that has been filled in, pasting it into the "Input text" box at Special:ExpandTemplates and clicking OK. The box that appears below that will contain the equivalent Wiki markup. --Redrose64 (talk) 21:16, 27 September 2013 (UTC)

Error message anchors

It occurred to me during an unrelated conversation that the mechanism used to link between a CS1 error message and the related help text in Help:CS1 errors is unnecessarily cumbersome. As it is right now, there is an {{anchor}} template above each error message section heading in Help:CS1 errors. There is a matching anchor in Module:Citation/CS1/Configuration. It occurred to me that, because each error message is given its own section heading, the anchors aren't necessary and that Module:Citation/CS1 should simply link to the section heading. To test this I've changed the anchors in Module:Citation/CS1/Configuration/sandbox to link to the section headings. The citations below are all created with the CS1/sandbox version to prove (or not) that the idea works: