Template talk:Cite web/Archive 6

From Wikipedia, the free encyclopedia
Jump to: navigation, search
← Archive 5 Archive 6 Archive 7 →


Translation parameter

I have added support for the translation parameter. Will add docs later. Revert if any problems are encountered. See also here. Crum375 (talk) 13:43, 23 June 2009 (UTC)

Should a url of a translated non-English language page be included in the template - something like |trans_url=  ?? Or would that mess up the template? (talk) 09:11, 17 September 2009 (UTC)
No, I see no reason to do that. Leave it to readers to decide which translation tool they prefer, rather than giving preferential treatment to any one. -- AnmaFinotera (talk · contribs) 14:24, 17 September 2009 (UTC)

Accessyear parameter

At the moment the interaction between accessdate and accessyear is defined by the following code

{{#if: {{{accessyear|}}}|{{{accessdate}}} {{{accessyear}}}|{{{accessdate}}}}}

If the accessdate parameter includes the year and an accessyear is defined, the result will have a double year. The solution would be to scrap this line entirely. But in that case, if the accessdate does not include the year and the accessyear is defined, we will not have the year at all.

In other words, the present code represent a judgement as to what is the less likely mistake. I would like to ask editors to reevaluate that judgement. Based on the documentation page, which calls for a full date in accessdate, and does not even mention the deprecated accessyear parameter, I would have chosen the opposite solution. Debresser (talk) 22:29, 10 September 2009 (UTC)

Discussion of second problem

Scrapping the accessyear parameter would solve it, but would that have consensus? Maintaining the status quo is another option, and changing the code is the other. Debresser (talk) 16:03, 14 September 2009 (UTC)

I think the best propositions are deprecating accessyear (in which case a bot would have to do some work to retain accessdates), or keep the status quo. Debresser (talk) 16:06, 14 September 2009 (UTC)

How many cases are there of |accessyear= being present but |accessdate= being absent? Basically, how many references would lose their "Retrieved on"? If there are a small number, I think a manual fixup might be best; several of the links may well turn out dead in any case, so might usefully be removed.
I'm not sure how to search for that combination (template A contains parameter B and not parameter C) without a bot. Please correct me if I'm making the wrong assumption there, but I've never managed to get the standard WP search tool to succeed with anything other than the most basic searches (word A is present, word B is present or both are present in the same article). The "advanced" search just seems to restrict to particular namespaces, rather than being like Google's advanced search facility. Now, if I had access to a UNIX shell prompt, read rights for the wiki sources and execute rights for egrep it would be a different matter... --Redrose64 (talk) 10:50, 15 September 2009 (UTC)
Obviously instances of accessyear without an accessdate would be dealt with by changing "accessyear" to "accessdate". Debresser (talk) 10:54, 15 September 2009 (UTC)
I can run a search on last week's database dump if that would help? Rjwilmsi 12:27, 15 September 2009 (UTC)
I don't think that will be necessary, thank you, because I didn't mean to deprecate any parameter without putting something else in its place. Debresser (talk) 12:44, 15 September 2009 (UTC)
It is impossible for a situation to arise where the "accessdate" of a reference can be anything other than a specific date. If you encounter a plain |accessyear= parameter, you can just click through the link, ensure that it still works, and then update the |accessdate= to the current date! There should never be a need for a parameter other than |accessdate= alone. Deprecated, replace, and remove, |accessyear=, |accessday=, |accessmonth=, IMO. Happymelon 14:14, 17 September 2009 (UTC)
Agree with User:Happy-melon. Personally, when editing a page and encountering there an existing {{cite web}} ref with either incomplete or missing access date, I've been back through page history to find out when it was added, and used the date of that revision as the value for |accessdate= - see, for example, this edit where I went back to this and this to get the access dates. Tedious, admittedly, but accurate. --Redrose64 (talk) 17:07, 17 September 2009 (UTC)
Why more accurate? Accessdates represent the most recent time when the link is known to have worked, to help us spot dead links. Really, we should be filtering old accessdates into a category and going round refreshing them! Happymelon 17:38, 17 September 2009 (UTC)
It's more accurate because it shows the date that the cited web page was found by the user who originally added the reference. There's nothing in the documentation stating that the date should be the "most recent time when the link is known to have worked" - it states, and I quote, "Full date when item was accessed ..." which I suppose is ambiguous. --Redrose64 (talk) 18:44, 17 September 2009 (UTC)
I agree it's ambiguous; my interpretation of it was always the "last time known to have worked" date. I don't see any value in recording what effectively amounts to the date the reference was added; that can be easily found through wikiblame if it's needed. Much more useful is, as I said, an indication of when the reference was last checked to be valid. (also)Happymelon 15:07, 18 September 2009 (UTC)
I'm sorry, but I shan't be updating |accessdate= whenever I visit an external link that succeeds. There are too many of them. When I discover any that fail, what I am doing is to pop in one of these:
--Redrose64 (talk) 15:49, 18 September 2009 (UTC)
Fortunately, this being a volunteer community and all that, there's no requirement for anyone to do anything. It's the difference between what happens in reality, and what would happen in an ideal world. In an ideal world, therefore, what would happen to |accessdate= parameters? Happymelon 11:28, 19 September 2009 (UTC)

Pre-filled "accessdate" parameter

Hi-I just noticed that the pre-filled "accessdate" parameter has been removed from the /doc page of this highly used template, without any discussion or consensus about the matter that I can find. The reason stated was that since the date format is used differently around the English-speaking world, the pre-filled date in the DD-MM-YYYY style was removed and left empty, because in some places, the MM-DD-YYYY format is more widely used. I'm concerned that editors (myself included) copying this format without the pre-filled access date will not remember to add any date at all. Wouldn't it be better to list the example with the date filled in, or if necessary, list two examples, each with the different date formats? That would make it much easier to use for copying & pasting the format, otherwise, the current form is inefficient, necessitating an extra (somewhat forgettable) step. --Funandtrvl (talk) 15:31, 27 September 2009 (UTC)

There have been several disconnected discussions. You can go over to Wikipedia talk:Manual of Style (dates and numbers) and pick one of the several current discussions. ---— Gadget850 (Ed) talk 16:21, 27 September 2009 (UTC)
Yes, thanks for the link. I realize that this is a long debated topic without any general consensus. It's too bad we can't all agree on something for the accessdate parameter, because with the "blank space for the date", it's only promoting the lack of efficiency for keystrokes. --Funandtrvl (talk) 16:33, 27 September 2009 (UTC)
We could give examples for all the date formats commonly used; would that be better? Eubulides (talk) 16:39, 27 September 2009 (UTC)
Yes, that would be better, then editors can just copy & paste the relevant one. Thanks! --Funandtrvl (talk) 16:56, 27 September 2009 (UTC)
OK, done. To avoid combinatorial explosion I trimmed down all the common alternatives to one, but with three commonly-used date formats. Eubulides (talk) 23:50, 27 September 2009 (UTC)
Thank you very much for fixing that, I appreciate it, it saves some steps!! --Funandtrvl (talk) 05:45, 28 September 2009 (UTC)

Author formatting

By trimming down all the common alternatives to one, you may be giving the impression that |author= is preferred to |last= |first=. An examination of the template source shows that several pairs of the latter (ie |last1= |first1= all the way up to |last9= |first9=) may be provided, and each pair may have an associated |authorlinkn=; however only one |author= is permitted, and where this is used for more than one author, |coauthors= is necessary and I don't think the latter will work with more than one |authorlink=. --Redrose64 (talk) 09:30, 28 September 2009 (UTC)
OK, I added examples of the |last1= etc. style. Eubulides (talk) 18:15, 28 September 2009 (UTC)
Made one chg to match instructions under opt. parameters, since "author" should only contain one author's name, and used 2 instead of 3, last, first, examples to prevent breaking to 2nd line. --Funandtrvl (talk) 18:44, 28 September 2009 (UTC)
No, there's no requirement that author must list only one author, and it's fairly common for it to list multiple authors. Furthermore, there's no requirement that author must contain last name first. The instructions should not imply that last-name-first is preferred. Eubulides (talk) 23:48, 28 September 2009 (UTC)
|authorlink= doesn't work as expected unless |author= has just one author (the whole string is linked to a single article, not just the first-named author); however, the order of names (first last vs last, first) is immaterial for |authorlink= to work --Redrose64 (talk) 23:58, 28 September 2009 (UTC)
Hi-User Redrose64 is correct about authorlink not working with more than one author. I thought I read somewhere that there was a discussion that the last & first parameters are preferred to the author & coauthor parameters, but I haven't searched lately for that discussion! Anyways, I think Last, First should be used as the example for a citation, according to WP:CITE#HOW, the last name is usually first in any bibliography, it also matches the "listas" parameter instructions from WP:BIOG. So I don't agree that the instructions shouldn't imply that last name comes first, because I sure am tired of fixing a lot of bibliographies in WP articles, for that very type of style error. --Funandtrvl (talk) 00:16, 29 September 2009 (UTC)
  • "I think Last, First should be used as the example for a citation, according to WP:CITE#HOW". No, WP:CITE#How does not say that "last, first" is preferred.
  • "authorlink not working with more than one author" It is quite common to use |author= without |authorlink=, and in that case it works just fine with multiple authors; individual authors in within |author= can be wikilinked as needed when |author= is used, and this is also common. It is also common to avoid wikilinking authors entirely, as per WP:OVERLINKING. In all these styles, |author= works just fine with multiple authors.
  • Many featured articles format author names themselves, using |author=, and there is nothing wrong with that. This is independent of whether last names are first; for example, the Vancouver system, a last-name-first style, is very commonly used in medical articles, and medical articles typically use |author= in order to employ that style, because the format used by |last= etc. is incompatible with Vancouver style. The {{cite web}} template is not about enforcing a particular format for authors; it allows all the common ones, and the documentation should not imply otherwise.
Eubulides (talk) 01:05, 29 September 2009 (UTC)
A clarification, I did not list all of the links in the see also section of WP:CITE#HOW. However, that is what I was refering to in that section. For example, the commonly used styles of APA, MLA, CMOS, Harvard, and Turabian use last, first. If the exception is the Vancouver style, couldn't the description be updated to indicate that? Isn't it the same issue as listing only three of the most commonly used date forms? If an editor has a format to follow, or copy and paste, isn't it easier that way? Also, the issue is not which format for the parameters to use, author or last, first, the issue is whether to show an example of using the author parameter with last, first. It is not a requirement, but if the five major citation style manuals show the author with the last name first, wouldn't you say that the example using last, first, would not be a problem? Like I requested before when the "today's date" parameter example was removed, it is much easier to copy and paste when an example is shown. The instructions can be updated to include the citation style differences and recommendations. --Funandtrvl (talk) 20:32, 29 September 2009 (UTC)
In practice |author= is used more often than |last1= and so should be listed first. Again, there seems to be a misperception here that |last1= etc. supports the major styles. This is incorrect. For example, the template mishandles multiple authors with the MLA style. So, if one really wants to use MLA style, one must often use |author= instead of |last1= etc. anyway. It is true that Vancouver is even more-poorly supported than MLA, because with the Vancouver style, one must use |author= even if there's just one author. But that doesn't mean MLA is well supported. Eubulides (talk) 20:51, 29 September 2009 (UTC)
I am aware of the fact that the |last=, etc. parameters do not support the MLA style, when using multiple authors, so that is not a misconception on my part. Nor do I disagree that |author= should be listed first. This is your quote that I'm referencing: "Furthermore, there's no requirement that author must contain last name first. The instructions should not imply that last-name-first is preferred.", in conjunction with: "Citations within a page should use consistent formats. However, there is no consensus about which format is best. The following examples are for citations where one author is listed as part of a single |author=Last, First parameter, along with |coauthors=Last, First; Last, First ... allowing for additional authors. Also shown below are three separate date formats that are commonly used in Wikipedia:". What I'm wondering is what your suggestion would be to improve this paragraph. It seems like we're talking about two different things here. --Funandtrvl (talk) 21:21, 29 September 2009 (UTC)
I'm sorry: I did misunderstand your comments. Let me try again (I hope I get it right this time). There are common styles where authors are listed first name first. For example, Turabian style uses first-name-first in footnotes, with the following as an example:
  • 10. Jeff Rybacki, “Neenah Creek Elementary School,” Wisconsin Dells School District, http://www.sdwd.k12.wi.us/neenahcreek.html [accessed March 8, 2008].
The blank form shouldn't specify any format for author names, as the blank form doesn't care about format if only |author= is specified. The first, simple form should not specify |coauthor=, as that parameter is too tricky: it requires a semicolon between author names, for example, and this is not at all obvious to a first-time user. Nor should the simple form give an example of |coauthor= with last-name-last, as that is less common (it contradicts MLA too, no?). Let's keep the first form simple; the more-complex stuff should be put later. I propose this diff (which I immediately reverted). Eubulides (talk) 23:53, 29 September 2009 (UTC)
Okay, now it makes sense. Go with it and be done with it! :) Funandtrvl (talk) 02:38, 30 September 2009 (UTC)

Odd formatting

Hello everyone! I'm not sure if this is the correct place to be asking, but I am having an issue with the cite web template in one of the articles I'm working on, so I thought I'd drop a question here. On the article Andalusian horse, two reference templates are for some reason putting some of the website url into the title. The link is working fine - it's just showing up wrong. Here's the formatting I'm using:

<ref>{{cite web|title=ANCCE|publisher=National Association of Purebred Horse Breeders of Spain|url=http://www.ancce.es/ver_wysiwyg.php?id=historia&seccion=¿Que es ANCEE?&subseccion=Historia|accessdate=2009-09-29}}</ref>

<ref>{{cite web|url=http://www.ancce.es/ver_wysiwyg.php?id=datos&seccion=El Caballo Español&subseccion=Datos del Caballo Español|title=Important Information about the PRE Horse|accessdate=2009-06-20|publisher=National Association of Purebred Spanish Horse Breeders of Spain}}</ref>

And here's how it's showing up:

es ANCEE?&subseccion=Historia "ANCCE". National Association of Purebred Horse Breeders of Spain. Retrieved 2009-09-29. 

Caballo Español&subseccion=Datos del Caballo Español "Important Information about the PRE Horse". National Association of Purebred Spanish Horse Breeders of Spain. Retrieved 2009-06-20. 

Does anyone have any ideas about how I can fix this? Dana boomer (talk) 16:49, 30 September 2009 (UTC)

The problem is that those URLS have. Replace the spaces with their proper URL encoding (%20 for spaces). "ANCCE". National Association of Purebred Horse Breeders of Spain. Retrieved 2009-09-29.  -- AnmaFinotera (talk · contribs) 16:53, 30 September 2009 (UTC)
Awesome! Thanks for the quick response. Guess you learn something new every day :) Dana boomer (talk) 17:04, 30 September 2009 (UTC)

series identifier

There is no way to specify a series identifier, like "Technical report TR23-45A" or a subtitle. Both of these would be helpful. (talk) 02:52, 1 October 2009 (UTC)

Italics for "Work" parameter

The "Work" parameter seems to automatically italicize the entry, but not all applicable entries should be italicized, such as websites. This seems like a problem to me. Drewcifer (talk) 06:50, 11 August 2009 (UTC)

Since there is no manual of style to indicate what output this template is to produce, how do you know websites should not be italicized? --Jc3s5h (talk) 10:52, 11 August 2009 (UTC)
This is correct. Use work when the publisher should be italicized (like citing the NY Times website or the like), otherwise use the publisher field which is not italicized. -- AnmaFinotera (talk · contribs) 12:42, 11 August 2009 (UTC)
Disagree. The work is always the published medium, the publisher is the company doing the publishing. Work is italicised, publisher is not. Huntster (t@c) 04:20, 12 August 2009 (UTC)
Publisher and work are two different things, at least in the case of many websites. Allmusic, for example, is published by Macrovision. And you're right, there is no MOS about this template specifically, but there is an MOS about websites and that they shouldn't be italicized. Where it's at eludes me at the moment, but if you need proof I can try and dig it up. So, that said, any template that is meant to facilitate websites should have the ability to facilitate the website in the appropriate style, ie, not italicized. Drewcifer (talk) 05:05, 12 August 2009 (UTC)
Oh, I don't disagree with regard to that, just pointing out to AnmaFinotera that the two fields cannot and must not be used interchangeably. Since Cite web shouldn't be used to cite news stories that were printed in physical form (for example), I see no problem with removing the italics altogether from this template, speaking in broad terms. Huntster (t@c) 05:09, 12 August 2009 (UTC)
I feel that the work parameter should be made devoid of the auto-italicising. Instead we can make the publisher as auto-bracketed like the way it is done for Cite news templates. Just a thought. --Legolas (talk2me) 05:44, 12 August 2009 (UTC)
The template does italicize the work entry automatically, but if you italicize the entry like this: work=Allmusic, it would appear in normal font in the reference section. For example reference #10 in the article "Live to Tell". Frcm1988 (talk) 06:12, 12 August 2009 (UTC)
Websites are not a "published medium" but generally a publisher of content. Makes sense to me. And not saying use them interchangeably, there are some web sources where you can and should use both. -- AnmaFinotera (talk · contribs) 11:15, 12 August 2009 (UTC)
Websites are most certainly published media..."published" does not necessarily mean the same thing as "printed". The website is never a publisher of content, merely the work that contains the content...there will always be an individual, company, or other entity behind that website. Two totally different things. If you can include both sets of data, then do so, but "work" is the only thing that really needed. Huntster (t@c) 22:50, 12 August 2009 (UTC)

The concept of what a website is, and a publisher is, is not made clear in the documentation (perhaps because there is no agreement on the meaning). In my mind, a website is a work; the medium is the World Wide Web. The publisher is a corporation, partnership, or individual. Unless you think Tron or The Matrix are non-fiction, publishers cannot exist in electronic form. --Jc3s5h (talk) 13:27, 12 August 2009 (UTC)

My only worry with re-doing the template is that it would fail to be backwards compatible. ie, all of the instances where the "work" parameter should be italicized. I can't necessarily think of any examples, but I'm sure they exist. So instead, what if we just added an extra parameter: "website". It's more straightforwardly language-wise (calling the website the "work" always was a bit of a stretch, IMO), it would be un-italicized, and it wouldn't mess up all of the millions of times the template's already been used. Drewcifer (talk) 17:01, 12 August 2009 (UTC)
Unnamed editor who posted at 17:01, 12 August 2009 UT, what is a website? --Jc3s5h (talk) 17:14, 12 August 2009 (UTC)
Sorry, that was me. One too many tildes, though I've fixed it now. I'm not sure what you mean by your question. Go to website to find out more I guess. Or is this question leading somewhere? I'm very confused. Drewcifer (talk) 00:08, 13 August 2009 (UTC)
The website article indicates, and I agree, that a website is a collection of digital content that is addressed with a single domain name or IP address. It could be a work if it is under the creative control of a single entity. On the other hand, if the only thing the different components share is the address, then it does not qualify as a work. For example, home.comcast.com wouldn't qualify as a work because the component pages are independently created by the many subscribers to that ISP. On the other hand, Wikipedia (English version) is a single work. If there is to be a website parameter, there should be clear instructions about when to use it. In some cases, it would be redundant to give both a work parameter and a website parameter. Also, since a cite web citation will usually include a link, the website is indicated by the address, so isn't usually necessary. --Jc3s5h (talk) 00:36, 13 August 2009 (UTC)
I think we're somewhat in agreement here. Adding a website parameter is not meant to replace the work parameter. And alternatively, they could of course be mis-used, like any parameter in any template. So of course clear instructions would be necessary to avoid misuse and redundancy. That said, the clear difference between what one would call a "work" and what one would call a "website" means we need to add or adjust something in the template. Drewcifer (talk) 00:55, 13 August 2009 (UTC)
Anyone mind pointing me to where it says that website names shouldn't be italicized? I can't find it. Goodraise 16:55, 5 September 2009 (UTC)
Wikipedia:Manual of Style (titles) doesn't specifically address website titles, however websites do not fall under the list of italicizing and in most citation styles, website names are not italicized nor put in quotes. -- AnmaFinotera (talk · contribs) 18:06, 5 September 2009 (UTC)
Okay, then it would seem that either (italicizing website names or not italicizing them) is acceptable under current guidelines. Personally, I don't care, but what I find disturbing is the upcoming practice of stuffing the website name into the publisher parameter of this template to avoid italicization. Is the only thing keeping us from adding a website parameter that we haven't decided on how the documentation should be updated? If so, how about this? Goodraise 04:01, 4 October 2009 (UTC)

Why is this outputting COinS?

Why are cite-web-generated References list items being output with COinS spans? This makes no sense unless the resource might exist as a book or article in a periodical, something which cannot be specified in the template. Worse, everything is being designated as a book. This is limiting the value of COinS because a large percentage of the spans in the citation lists are now nonsense (corresponding to web-only content, designated as books). Hence this is also harming applications that use coins by making them look like buggy time-wasting junk. (talk) 18:31, 9 October 2009 (UTC)—Preceding unsigned comment added by (talk) 18:05, 9 October 2009

Please sign your posts, by typing four tildes thus: ~~~~ --Redrose64 (talk) 18:16, 9 October 2009 (UTC)

Incorrect error message

On 2009 Philadelphia Phillies season, this error message is displayed about 50 times: "Error: If you specify |archivedate=, you must also specify |archiveurl=." The thing is, they all already have that parameter. What should I do to fix this? Coemgenus 15:21, 19 October 2009 (UTC)

It was a temporary error caused by a recent change that has since been reverted; try purging your cache and see if it works.  Skomorokh, barbarian  15:25, 19 October 2009 (UTC)
Thanks, that worked. Coemgenus 15:31, 19 October 2009 (UTC)

archiveurl error

In History of Louisville, Kentucky, in reference #12, I'm seeing the error "Error: If you specify |archivedate=, you must also specify |archiveurl=", even though archiveurl is indeed specified. Is there something I'm missing? Thanks. Stevie is the man! TalkWork 15:24, 19 October 2009 (UTC)

Someone obviously changed something because I'm getting this error on archived stuff I know was fine yesterday. I suspect it may be {{Citation/core}} but I'm looking into it and hopefully it will be fixed soon. Rambo's Revenge (talk) 15:26, 19 October 2009 (UTC)
I had the same problem. I posted it to the village pump, as well, to see if anyone there knows what's up. Coemgenus 15:27, 19 October 2009 (UTC)

See User talk:Amalthea#cite web etc.. Rettetast (talk) 15:40, 19 October 2009 (UTC)

It's fixed, see for yourself. Nothing broke, just that it took a while to change to all the new templates. In the mean time, we had these messages. Debresser (talk) 16:24, 19 October 2009 (UTC)

Thanks! Stevie is the man! TalkWork 17:53, 19 October 2009 (UTC)

Upper case form of template name

A recent edit by Debresser (talk · contribs) changed all instances of (lower case) {{cite web}} to (upper case) {{Cite web}} in the documentation, with the cryptic comment "m using AWB". When I reverted, Debresser reinstalled the edit (along with a typo) with the summary "Restore Cite web with capital. That is the name of the template, and what it should be. Common or not is another matter. We are giving the official documenttation here. Not how to do it wrong." This edit summary is incorrect. First, there's nothing "wrong" with {{cite web}}; it is perfectly well supported, and it's easier for many editors to read when editing. Second, that part of the documentation is giving common formats in widespread use, not the official name of the template. The official name of the template is in big letters at the start of Template:Cite web, and this official name is unaffected by what goes into the documentation.

The documentation has been using (lower case) {{cite web}} ever since it was created in 2006. This should not be changed to the upper case form simply because of a personal preference that the upper case is what it "should be".

Debresser went on to install the same edit in many other instances of the documentation for {{cite book}}, {{cite news}}, etc.[1][2][3][4][5][6][7][8][9][10] These edits, like this one, go against common usage of these macros. For now I am reverting these changes. I suggest that further discussion and consensus be done here before recommending such a change in common practice. Eubulides (talk) 05:25, 21 October 2009 (UTC)

Minor correction to my previous comment: another editor reverted one of the abovementioned changes before I got to it. Eubulides (talk) 05:38, 21 October 2009 (UTC)
Ok, let's discuss it. This is not a "big deal", since anyway the software recognises both.
The name of a template or article or any page for that matter is spelled with a capital in Wikipedia. I have seen many editors being carefull to keep this convention. I could mention User:SmackBot as one of the most well-known. I think it is only proper that we should strive to conformity in this. So while not being zealous about it, I think we should give the right example when drawing up documentation. Many documentation pages, and other pages including instructions concerning the use of templates indeed do so. I do not think it was proper for User:Eubulides to revert this minor improvement to Wikipedia on the grounds he mentioned. I'd therefore suggest that unless consensus here would indicate otherwise, we go back to the capitalised versions. Debresser (talk) 06:30, 21 October 2009 (UTC)
Since this question involves quite a few documentation pages, of several citation templates, I put up a link to this discussion on each of them. Debresser (talk) 06:39, 21 October 2009 (UTC)
I think Eubulides was right to revert in the first instance - if we're changing this it should be discussed first. As far as I'm concerned both the uppercase and lowercase versions of the template name work exactly the same. I don't know why it matters which is used. Rjwilmsi 06:43, 21 October 2009 (UTC)
  • Longstanding consensus on this page, and in related macros such as {{cite journal}}, is to use the template names without capitalization.
  • The convention for capitalization differs among macros; for some macros, capitalization is more common. But that's OK: there's no need to insist on capitalization for all macros, or on lowercase for all macros.. User:SmackBot may capitalize some other macros, but it rightly does not capitalize {{cite web}}.
  • I now see that Debresser made a similar change recently to the {{reflist}} documentation (here, which I just now reverted), but this (again) appears to be a personal preference rather than any consensus. Let's leave this documentation alone, please. There are some advantages to lowercase macro names here: for example, they make it easier for editors to focus on the text of the article (the more-important part) and to ignore the macro names (the less-important part). There is no need to insist on capitalizing the names when common practice is otherwise.
  • It is worrisome that the capitalization edits were installed with uninformative edit summaries like "minor", as these edits were not minor, and they deserved better summaries.
Eubulides (talk) 07:02, 21 October 2009 (UTC)
  • I don't think there is anything "major" about changing from lower to upper case! So stop making it look as though the "m" came to hide anything. Debresser (talk) 09:17, 21 October 2009 (UTC)
  • And they have been that way on reflist a whole month without anybody making a problem out of it. So I really object to you reverting these edits unless there would be a good reason. Which comes to exclude the claim of "consensus". This has never been discussed, so there is no consensus. Debresser (talk) 09:17, 21 October 2009 (UTC)
  • I have seen you yourself all over these documentation pages for the last few months, and making edits without much of a consensus also. With edit summaries like "reordering", "improving"... Whom did you ask before you made all those tens of edits? So now you absolutely have to revert my capitals? Debresser (talk) 09:18, 21 October 2009 (UTC)
The Wikipedia:BOLD, revert, discuss cycle is useful and is quite common, and that's what I was doing. It's less common to see the pattern that we've seen in {{cite web/doc}} and {{reflist/doc}}, where A does a bold edit to a longstanding version, B reverts, A reinstalls the edit, and then A complains that the edit is minor and that anyway there is no consensus so A's version should stay installed. Eubulides (talk) 09:27, 21 October 2009 (UTC)
I know, and agree. If there is a reason. But there is also something like not undoing a good faith edit without a good reason. That is what bothers me most here. You have made many edits that made me frown. And without much discussion, if at all. But I don't undo a fellow editor's edits without good reason. And this is what you do? To a perfectly acceptable minor edit? Debresser (talk) 09:32, 21 October 2009 (UTC)
Try not taking this personally. Your change, though it's only a single letter, potentially affects the actions of thousands of editors. So it's definitely worth of discussion and your edits worthy of reversion until we hash out a good solution here. KellenT 09:51, 21 October 2009 (UTC)
Also, I gave multiple good reasons (or, at least, reasons that I thought good :-) in the discussion above. First, the longstanding and stable tradition is lower case. Second, lower case allows editors to concentrate better on article text (the more-important stuff), and be less distracted by the template name (the less-important stuff). Eubulides (talk) 09:54, 21 October 2009 (UTC)
The template's name in the database begins with a capital letter, and therefore the "official name" begins with a capital letter. But this is due to a technical limitation in MediaWiki, so it's pretty clear that this was not a real choice by any editor, just a result of the environment in which we are working. It's also not wrong, as Debresser suggested in an edit summary, to use a lowercased template name for the same technical reason. There has been an implicit consensus regarding the names since the template documentation's creation; thousands of editors have used the templates and never felt it important/correct to change the names to uppercase. My personal preference is definitely for lowercased citation template names in their general usage, because I find this less distracting typographically. KellenT 09:50, 21 October 2009 (UTC)
If I used the word "wrong", that was wrong. :) Any user is free to use whatever style he pleases. I have great respect for that. I did not it appreciate when User:Eubulides started being possessive about these pages to the extent of following me around to revert minor changes. Debresser (talk) 16:34, 21 October 2009 (UTC)
The previous comment omits relevant chronology. Debresser installed the change here; I reverted with an edit summary saying why; in response, Debresser reinstalled the change and then went to ten other templates and installed similar changes there, marking them all minor.[11][12][13][14][15][16][17][18][19][20] Eubulides (talk) 20:58, 21 October 2009 (UTC)
And how are you so sure I noticed your revert? Please don't be childish. "You were first!" "No, you were first!"... Debresser (talk) 07:16, 22 October 2009 (UTC)

My initial thought on this is that it is not worth the amount of discussion already expended on it, much less the degree of passion displayed. The templates work exactly the same either way, regardless of what the examples show in the documentation. This is beyond "minor", it's trivial. It isn't worth the effort to systematically change it in either direction. That includes following Debresser around just to revert his (unnecessary but not harmful) changes. If you are making some other real improvement to the documentation, and while you've got the edit window open you want to waste time converting the capitalization, then that's your call. It affects nothing except possibly pleasing or offending your own subjective aesthetic preferences. --RL0919 (talk) 13:06, 21 October 2009 (UTC)

  • Trout slap to everyone for wasting so much time on a meaningless change. --ThaddeusB (talk) 14:51, 21 October 2009 (UTC)
    • I agree, but the problem is that Debresser has now taken RL0919's comment as a strategy to change things back the way Debresser prefers it, and has recently installed unrelated changes to many pages of documentation, including in these edits changes uppercase all the template names,[21][22][23][24][25][26][27] despite the obvious consensus here against the capitalized form and against this kind of change. This is continued imposition of a personal preference that differs from common usage, and it's quite counterproductive to make changes like this. Eubulides (talk) 20:58, 21 October 2009 (UTC)
It may differ from the more prevalent usage, but it is common as well. And conformity is also important. Better have them all capitalsied then part lower part upper case.
The edits I made to those pages were important updates to explanations regarding archive parameters, based on the formulation worked out (with my help) in {{Cite web/doc}} months ago, and very relevant in view of my rich experience in the error category of broken citations. I did not make edits to other docpages where my edits had been reverted previously. Debresser (talk) 21:06, 21 October 2009 (UTC)
These have all been in lowercase since the dawn of time and they are nearly always used in lowercase in articles. Keep them lowercase, otherwise it's just going to be even uglier in the edit windows. There's no reason to change them, so keep them as they were. Headbomb {ταλκκοντριβς – WP Physics} 01:45, 22 October 2009 (UTC)
At the dawn of time there was no Wikipedia either, so that argument is a "non-argument". But your opposition to upper case is duely noted. It's a matter of taste, probably. To me, I expect upper case after "{{" brackets, and find it esthetically satisfying. Debresser (talk) 07:18, 22 October 2009 (UTC)
And I expect to see a lowercase, as I imagine, most people would. And could you be less patronizing? It's pretty annoying. Headbomb {ταλκκοντριβς – WP Physics} 13:59, 22 October 2009 (UTC)
I'm sorry, but that is my style. I am not even aware of it. So you'll just have to put up with that. Debresser (talk) 18:06, 22 October 2009 (UTC)

No further comment, and there is consensus (with the notable exception of Debresser) that these should be left alone, so I changed them back to the longstanding state. WP:RETAIN seems appropriate here. Eubulides (talk) 17:27, 30 October 2009 (UTC)

Thank you for enforcing the consensus. You are precisely the type of non-involved editor who should be doing that. And what about the consensus that this is not worth an edit? It is very hypocrite, to pick from consensus just those parts that you like, and ignore those that you don't. Well, people are know by their deeds. Debresser (talk) 16:22, 31 October 2009 (UTC)

Double period bug

The cite template double period bug is among the most frequent typos on Wikipedia. I collected some examples here. It affects about 45,000 articles (1 1/2% of a random sample of all 3 million), and about 23% of the most frequently read articles (presumably because they are longer and better referenced). Since I can't get anyone to fix the template or write a bot, I hope to use my WP:AWB edits to fight this problem at the same time I do Manual of Style edits. 45,000 is too many articles to fix one at a time, even with AWB, but perhaps I can fix the most frequently encountered examples of the problem that way. The problem affects Cite Web, Cite Book, Cite News, Cite Press Release, Cite Journal, and Cite Encyclopedia, so it probably affects all the cite templates.

Sometimes the template adds an extra period; sometimes it doesn't. So I can't tell AWB to remove any period from the end of any cite template field, because that would remove correct single periods along with removing one of a pair of duplicate periods. It's easy to say I should use AWB preview for every cite template change, but that removes most of the reason to use AWB. So I collected my list of examples, experimented with it, and looked for a pattern. The pattern is that when the displayed version of the template adds a parenthesis or other punctuation, it doesn't add an extra period to go with it. Unfortunately, AWB reads the coded parameters, not the displayed finished product. I think the pattern goes like this, at least for the most frequently encountered versions of this bug:

publisher= should almost never end with a period. However, for cite press release and cite journal, which lack the first= parameter, the final period should not be removed if there is a (non-blank) date= parameter.

first= The final period shouldn't be removed if there is a non-blank date= or a year= parameter.

Similarly, author= and coauthors= should keep their final period (if any) if there is a date= or year= parameter.

title=, accessdate=, isbn=, page=, pages=, editor=, editor2-first=, and encyclopedia= should have any final period removed.

location=, last=, edition=, and editor1-first= should be left alone.

year= is unlikely to end with a period, but if it does, it will be duplicated if there is a month= parameter, and not otherwise. month= should be left alone, assuming there is a year= to go with it.

journal= should be left alone unless volume= and issue= are both missing.

So is that about right? Art LaPella (talk) 02:34, 16 November 2009 (UTC)

A discussion has been going on for some time at Template talk:Citation#Full stop at the end of the template 2 which might affect the above. --Redrose64 (talk) 10:46, 16 November 2009 (UTC)
That discussion has been stuck for a while now. Because of a few whiners. Debresser (talk) 16:09, 16 November 2009 (UTC)

Parameter "edition" needed

{{Editprotected}} An edition parameter is needed, for online works that change over time but hold static materials that are being cited, and probably for some other cases. The field should only appear if the work parameter is used, and it should be formatted like, and have the same syntax as, the same field in Template:Cite book. Example:

<ref>{{Cite web
|title=What is Occam's Razor?
|year=1997 article in May, 2009 compilation
|work=Usenet Physics FAQ
|editor=Don Koks
|accessdate=August 17, 2009

[a 1]

would be much more elegant as:

<ref>{{Cite web
|title=What is Occam's Razor?
|work=Usenet Physics FAQ
|edition=May, 2009
|accessdate=August 17, 2009

[a 2]

  1. ^ Gibbs, Phil (1997 article in May, 2009 compilation). "What is Occam's Razor?". Usenet Physics FAQ. Retrieved August 17, 2009. 
  2. ^ Gibbs, Phil (1997). "What is Occam's Razor?". Usenet Physics FAQ (May, 2009 ed.). Retrieved August 17, 2009.

Please note the differences:

  • accessdate: When the editor saw the cited page.
  • date, year, month: Relate to when the cited page or series of pages was/were published (i.e., usually proximal to the authorship date)
  • archivedate: relates only to Archive.org, etc.
  • edition: An open parameter in which any value may be inserted, e.g. "2009", "3rd", "revised", that relate to the larger work containing the cited page or series of pages. When used to provide a work-publication date, may be many years different from the date (or year / month) values.

SMcCandlish [talk] [cont] ‹(-¿-)› 21:11, 17 August 2009 (UTC)

This looks like it might be a good idea. But would you mind leaving it a few days to garner comments and obtain a consensus before placing the {{editprotected}} request? — Martin (MSGJ · talk) 08:58, 18 August 2009 (UTC)
An edition parameter is definitely needed. Do you have the code for the change so we can slap the editprotected banner back on it? --Philosopher Let us reason together. 13:49, 3 September 2009 (UTC)

{{Editprotected}} There's been no objection, over 2 months later. — SMcCandlish [talk] [cont] ‹(-¿-)› 04:51, 20 November 2009 (UTC)

Would you be able to place the required code on the sandbox version? — Martin (MSGJ · talk) 20:42, 20 November 2009 (UTC)
Disabled for now, as no response and it seems the code is not yet written. If you can't do this yourself, you'll need to find someone familiar with these templates who can code it for you. — Martin (MSGJ · talk) 07:27, 21 November 2009 (UTC)

Date formatting

This template documentation looks like it violates WP:MOSNUM, which says that YYYY-MM-DD date formats should not be used. Shouldn't it be updated to conform to the guideline? Timmeh 03:13, 12 November 2009 (UTC)

See Wikipedia:Mosnum/proposal on YYYY-MM-DD numerical dates. Good luck with that. ---— Gadget850 (Ed) talk 03:50, 12 November 2009 (UTC)
It probably should, and in practice instances generally are, but as noted, there is no consensus one way or the other, basically leaving it up to the individual editors of specific articles to decide, in the end. -- AnmaFinotera (talk · contribs) 03:54, 12 November 2009 (UTC)
WP:MOSNUM does not say that YYYY-MM-DD date formats should not be used in citations. It says only that the format shouldn't be used in prose. It's quite common for the format to be used in citations. Eubulides (talk) 04:26, 12 November 2009 (UTC)
I personally think it is an ambiguous and unclear format which should not be used anywhere. Debresser (talk) 05:51, 12 November 2009 (UTC)
Please can we take this debate back to Wikipedia:Mosnum/proposal on YYYY-MM-DD numerical dates and thus keep it in one place? --Redrose64 (talk) 11:58, 12 November 2009 (UTC)

Suggestion: Why not make this template format the date given in the accessdate parameter according to the user preferences, as is done for {{date}}? That might satisfy more people than the current situation. I looked at the Mosnum proposal but quickly ran back here. -84user (talk) 23:28, 14 November 2009 (UTC)

I think the best thing to do would be to encourage the use of one format for consistency, either the mdy or dmy format, unless there is a good reason not to. Using the iso format introduces ambiguity as to the date and month that should really be discouraged. Timmeh 23:49, 14 November 2009 (UTC)
Please take this back to Wikipedia:Mosnum/proposal on YYYY-MM-DD numerical dates and avoid WP:CFORK --Redrose64 (talk) 23:52, 14 November 2009 (UTC)
Following the massive RfC to abandon date-linking and autoformatting, this function -which existed within the template and which converted ISO dates to dmy or mdy according to users' preferences - was switched off. Having it autoformat again is a retrograde step. Ohconfucius ¡digame! 04:03, 18 November 2009 (UTC)
Agree with Oh. Or Redrose. Or Timmeh, so that the footnote format can be consistent w/the format in the body of the article. But definitely not with forcing an all-numerical format, as it increases ambiguity.--Epeefleche (talk) 18:40, 18 November 2009 (UTC)
The template does not violate MOSNUM, which states "YYYY-MM-DD style dates (1976-05-31) are uncommon in English prose, and should not be used within sentences". Clearly citation templates are not encompassed by that. As pointed out by others, a general proposal which would have included citations was recently discussed and dismissed. I see no value in having a repeat of that, and the discussions that preceded it, here. wjematherbigissue 18:54, 18 November 2009 (UTC)

"page=" parameter issue

Minor thing I've noticed with some references I've used, the "page=" parameter is only showing the number, not "p. #". To illustrate:

  • {{cite web |publisher=[[IGN]] |author=Shea, Cam |url=http://xbox360.ign.com/articles/954/954036p2.html |title=Street Fighter IV AU Review | page=2 |date=2009-02-12 |accessdate=2009-08-09}}

Happens with "pages=" as well. Sorry if this is an already known about issue.--Kung Fu Man (talk) 14:47, 9 August 2009 (UTC)

KieferSkunk (talk) — 18:37, 10 August 2009 (UTC)

Probably as a result of the above fix there are now thousands of Wikipedia articles that have a double pp. i.e. "pp. pp." in the citation e.g. Impossible differential cryptanalysis which has as 'Further reading' entry "Eli Biham, ... pp. pp.186–194. ..." Is there a way the template could ignore any "pp"s in the pages= field (and "p"s in the page= field)? Diverman (talk) 02:59, 26 November 2009 (UTC)
I checked the database dump and found 924 pages with this problem. I plan to fix them using AWB, so that no change to the template would be necessary. Svick (talk) 23:33, 26 November 2009 (UTC)
Is this the same problem? It doesn't look so resolved to me. I've been automatically "fixing" pages=pp. when I find it, but maybe that's what we need to make the template work. At least in this example from Henry Wells (general), "pages=pp." is necessary to get "pp." to show:
{{cite journal|first=General Sir James|last=Marshall-Cornwall|authorlink=James Marshall-Cornwall|title=Staff Officer 1914–1918|journal=War Monthly|issue=No. 42|pages=pp. 9–15|publisher=[[Marshall Cavendish]]|location=London|year=1977}}
Marshall-Cornwall, General Sir James (1977). "Staff Officer 1914–1918". War Monthly (London: Marshall Cavendish) (No. 42): pp. 9–15.  Art LaPella (talk) 05:00, 27 November 2009 (UTC)
I see what you mean Art, the "cite journal" appears to work differently to the "cite web". I couldn't get the {{citation needed|date={{subst:CURRENTMONTHNAME}} {{subst:CURRENTYEAR}}}} to auto-display the "pp." on the Henry Wells article, but {{cite web}} did work. I also got {{cite journal}} to auto-add the "pp." using an example based on the one at the start of this thread. In the three examples below I have used "|pages=9–15" in the template.
  cite web   -    Marshall-Cornwall, General Sir James (1977). "Staff Officer 1914–1918". War Monthly (No. 42). London: Marshall Cavendish. pp. 9–15. 
  cite journal - Marshall-Cornwall, General Sir James (1977). "Staff Officer 1914–1918". War Monthly (London: Marshall Cavendish): 9–15. 
  cite journal - Shea, Cam (2009-02-12). Street Fighter IV AU Review. IGN. pp. 9–15. 
Mabye a template guru can tell us why one {{cite journal}} auto-shows the "pp." and the other one doesn't. - Diverman (talk) 11:30, 27 November 2009 (UTC)
Have looked at source for {{cite journal}}. If any of |journal=, |periodical=, |magazine= or |work= is specified, the page number(s) are not preceded by "p." or "pp.". If all those four are absent, the "p." or "pp." is used. The way it's coded suggests a deliberate decision, rather than an accidental bug. --Redrose64 (talk) 12:01, 27 November 2009 (UTC)
Have looked at source for {{cite web}}. This, on the other hand, always uses the "p."/"pp." form regardless of the presence or absence of other parameters. --Redrose64 (talk) 20:14, 27 November 2009 (UTC)

(outdent) I am not sure why {{cite journal}} omits the "p." or "pp." prefix in some cases. I left a message on the talk page and proposed adding {{Page numbers}} to {{cite journal}}. {{Page numbers}} is designed for use in other templates. It attempts to detect whether the prefix has been supplied by the user. If not, it supplies the prefix. — John Cardinal (talk) 21:50, 27 November 2009 (UTC)


{{editprotected}} Can someone please update this to the current date? It is still reading December 2 on all date parameters. Thanks. Copana2002 (talk) 18:32, 7 December 2009 (UTC)

Purging the page was enough to do this, no need to change the template. You can do it yourself next time. Svick (talk) 18:49, 7 December 2009 (UTC)

Work and publisher

There's a rather long discussion below, but the dispute here seems to revolve around differing interpretations of the name for a couple of parameters. One possible solution is to either add a "Website" parameter, or to rename one of the existing "Work" or "Publisher" parameters to become "Website". Another solution is to (re)revert the documentation page back to it's original state, prior to the undiscussed and unadvertised change made to it in August. The other option is to do nothing, leaving the change made in August as "live".
V = I * R (talk to Ω) 05:51, 4 December 2009 (UTC)

I've just reverted a change which was made to the documentation this last August. The main reason for the reversion is that such a change is actively causing confusion among editors. This prescriptive advice is not something that I see as being constructive to the encyclopedia either, especially considering the predominance of existing dissimilar usage patterns. We should have a conversation about such a change before implementing it, at the very least.

Personally, I find the existence of the two parameters in question to be confusing and constraining. If I had my way I would pick one term and add multiple instances of it in the same manner that author name parameters do (for ex: work, work1, work2, work3, etc..., although I would likely use "publisher"). There have been cases where I wished for at least one more similar field, and I think that somewhat permanently diffusing this potential confusion/conflict would certainly be a good thing.
V = I * R (talk to Ω) 12:42, 19 November 2009 (UTC)

The thing about cite web is that just about anything can be put on the web, including journals, books, TV shows, and movies. Thus, the cite web template needs the flexibility of cite journal, cite book, cite video, and almost all the rest. The title parameter is the title of the smallest information unit that
  • has a title
  • contains all the information needed to support the claim.
That might be a journal article, a TV episode, a book in a series of books, etc. In order to locate the source or understand its context, it might be useful to give the title of the next-higher information unit. That is what the work parameter is for.
  • title=Transaction Security System| work=IBM Systems Journal
  • title=The Two Towers| work=The Lord of the Rings
  • title=Once More, with Feeling| work=Buffy the Vampire Slayer
Jc3s5h (talk) 20:53, 19 November 2009 (UTC)
That's exactly why I'd like to see the two parameters be more deeply structured, actually. Having a variable number of publisher/work parameters, and making there usage seem less constrained, seems very useful when it comes to web references. The web largely just isn't as structured as books or journals, and there are many instances of for example: XYZ.com domain, hosting FOO publication, which is re-publishing BAR content on their cite. More directly relevant here though is that perscriptive instructions that either the work/publisher parameter are specifically for or specifically not to be used for certain things (such as "don't put website.com in the publisher= parameter) are way to constrictive.
V = I * R (talk to Ω) 21:07, 19 November 2009 (UTC)

Citing sources is the epitome of an encyclopedia.

The word source, as used in Wikipedia, has three related meanings: the piece of work itself (an article, book, paper, document), the creator of the work (for example, the writer), and the publisher of the work (for example, The New York Times or Cambridge University Press) (WP:V#Notes)

From that quote work is a cited work piece, a referent freelance author's work, or a publisher's referent employment work.

In WP:CITE and WP:V and WP:LAYOUT and WP:WORKS, work consistently means human effort, whether written, or sounded or formed out, in arts or sciences, whether completed or ongoing, or whether by an individual or their employer.

The key to my attempt at understanding publisher V.S. work fields here is who is employing who. In order to need the publisher field, a freelance author employs a publisher, and we use the publisher field. In the confusing case where the publisher employs the author the publisher is the work, and the publisher field remains blank. Confusion arises when the employer is a publisher or the employee is a self-publisher. A self-publisher is not really published in the traditional and more reliable sense, but hosted. Still we use the publisher field as we would for a professional freelance author.

The template doc should read something like:
  • work: If this referent is part of a larger work, such as a book, periodical or website, write the name of that work. If the publisher is an employer of the author, use this work field instead of the publisher field. Consider the lookup tool whois to find the work for the URL.
  • publisher: If the publisher does not employ the author of the work, for example, if the website is hosted cite the host as a publisher. If the publisher is also the author's employer, use the work field instead and leave this blank. Consider the lookup tool whois to find the publisher for the URL.

CpiralCpiral 01:31, 20 November 2009 (UTC)

Cprial is putting the cart before the horse. The title is the title of the material that is being cited (because it supports the claim in the Wikipedia article). The publisher is the publisher of that material. If the material is part of a larger work, the title of the larger work goes in the work parameter. If you don't like that meaning of work, the solution would be to rename the parameter to something better, not to cloud the meaning of publisher. --Jc3s5h (talk) 01:45, 20 November 2009 (UTC)
Perhaps, I'm reversing things in the hope that it might be the eventual way out of the confusion the other voices express concerning the two fields work and publisher. You seem content to shoo the thing off and just say "The publisher is the publisher of that material", but per the (two) reversions explanation, the publisher field is not used for publishers who employ the author. — CpiralCpiral 03:49, 20 November 2009 (UTC)
I don't know what "per the (two) reversions explanation" means, but the author is the person (or group) who actually creates the information, and the publisher is responsible for getting out to the public. Who employs who is irrelevant. --Jc3s5h (talk) 03:52, 20 November 2009 (UTC)
This conversation actually highlights the main issue that I have with the proposed addition to the documentation. We can't even agree among the few of us here what exactly defines a "work" and what defines a "publisher". That distinction is quickly becoming less important in the world for the simple fact that either can be relevant when discussing a website. That's why I was actually suggesting a less structured approach, above.
V = I * R (talk to Ω) 12:27, 20 November 2009 (UTC)
If editors need to be directed at an example, consider the following:
  • {{cite web |url=http://en.wikipedia.org/wiki/Han_Dynasty |title=Han Dynasty|work=Wikipedia|publisher=Wikimedia Foundation, Inc}}
There now. --Redrose64 (talk) 10:44, 20 November 2009 (UTC)
humm... I tend to agree that using both parameters is some form of an ideal, but the reality of usage is that the vast majority of users don't use the parameters in that manner. Pick any random article that uses {{cite web}} and you're more likely then not to see editors preferring one parameter over another (I happen to prefer the "publisher" param in most cases, since it's meaning is clear, but many others seem to prefer "work"). If the 173 or so people here want to attempt to create a Wikipedia wide cassus belli to edit war over how this template is used then I can't stop you, but I don't think that this is a particularly constructive issue to argue over.
V = I * R (talk to Ω) 12:21, 20 November 2009 (UTC)
That is a great example. It really says it all concerning the use of the fields work and publisher when the work is named slightly differently from the publisher. The reality of trying to document, every situation is futile. One simply "gets it" when one arrives upon the scene of a situation and takes an intelligent action. Thank you for that example. — CpiralCpiral

This conversation started because two separate editors tried to make the same improvement to the text describing the fields work and publisher. These two fields are closely related. Ohm reverted them both citing a "don't use the publisher field" type of argument, but the weakness of that reasoning was brilliantly illuminated by the work/publisher = Wikipedia/WikiMedia example. The problem, as Ohm points out, is many editors misuse these two fields, and they make the wrong edits describing these two fields, saying "use whois to get information", (which is really, in my opinion a good idea&mdasl;info, you know), and "The publisher is not usually the name of the website (that is usually the work)." (which, in my opinion is not very lucid).

I propose we reach a consensus on some new wording for these two fields in an effort to solve field purpose issues. Certainly we should start by first considering the possibility that the existing field names make sense, albeit esoteric, when they are documented and explained well enough? There is no limit on the size of the words explaining either field. These fields are difficult to document. (See the deceptively simple field editor in Template:Cite_book/doc#Description.) The work and publish fields of template:cite web fields are poorly documented. Summarizing my earlier writing I think the word employ should be employed because since the word source can have three different meanings, all related to work, and work can also have different meanings, including a relationship to employ. — CpiralCpiral 18:51, 20 November 2009 (UTC)

WP:V states a source can mean the author, the work, or the publisher. These are captured in the Cite web template in the author, title, or publisher parameters respectively. In exceptional cases, there may be a work-within-a-work, in which case title is used to name the smaller work and work is used to name the containing work. While the word "work" can be used in the sense "As of 1996, T. R. Bednar worked for IBM", that would not affect how the article he coauthored would be cited:
  • {{cite web| url = http://domino.research.ibm.com/tchjr/journalindex.nsf/4ac37cf0bdc4dd6a85256547004d47e1/905766d989ef415d85256bfa0067fc27?OpenDocument |author = Bednar, T.R. |coauthors = Piro, R.A.; Stout, D.W.; Wissel, L.; and Zuchowski, P.S. | title=Technology-migratable ASIC library design |work=IBM Journal of Research and Development |publisher=IBM |date = June 1996 |url=http://domino.research.ibm.com/tchjr/journalindex.nsf/4ac37cf0bdc4dd6a85256547004d47e1/905766d989ef415d85256bfa0067fc27?OpenDocument}}
which renders as
If I were to cite another article in the same issue, where none of the authors worked for IBM, the citation would follow the same pattern. There would be no change to the work or publisher fields. So I just don't see how the employment relationship between the publisher and the author would change anything.
By the way, the publisher is not usually stated when citing journals, so the publisher field could be omitted. Also, {{Cite journal}} would be a better template to use in this case.

Jc3s5h (talk) 20:54, 20 November 2009 (UTC)

Jc3s5h, thank you for talking about my proposal to include the concept "employ" when trying to remedy the misunderstanding of the two fields "publisher" and "work". There is a problem with those two fields, as attested by veterans AnmaFinotera and Debresser, who spend most of there efforts wikignoming to improve Wikipedia. The WP:TPG wisely recommends keeping posts to 100 words or less. I will start another section "The employment concept in Publisher and Word fields". When I do, I hope to see you there. This discussion is approaching the 30[kB] limit. It's now at 27.5 [kB].— CpiralCpiral 21:28, 20 November 2009 (UTC)

Restoring part of the text

(Outdent) I'm restoring part of the text. It is very important that these citation templates all be as consistent as possible, or fewer and fewer editors will continue to use them (or ever learn to use them) as they continue to grow in complexity. — SMcCandlish [talk] [cont] ‹(-¿-)› 04:56, 20 November 2009 (UTC)

I re-reverted it back to the original text until we've finished this discussion. Let's not edit war over the text, especially since there seems to be some interest in discussing this. Now I need to catch up on the rest of this discussion...
V = I * R (talk to Ω) 12:09, 20 November 2009 (UTC)
All in good time. Please allow our discussion to produce enough steam for a bold brew. :-P Ohm. — CpiralCpiral 06:58, 20 November 2009 (UTC)
I think SMcCandlish did the right thing. "Hail" to this discussion, but that is no reason to remove that important and relevant part of information. Debresser (talk) 09:13, 20 November 2009 (UTC)

(Outdent again) Re: "Revert (again). How about we finish the discussion on the talk page before edit warring over this?" (edit summary from Ohms law (talk · contribs)):

A: Several months is more than enough time. Per WP:BOLD, WP:FILIBUSTER, WP:CONSENSUS, etc., productive, good faith editing is not to be held up by a "lone ranger" objector. There is no active controversy about the specific wording I restored, there is simply User:Ohms law reverting without fully understanding. This appears to be a knee-jerk "don't edit this!!!" reaction, which is a bit too WP:OWNish for my tastes...

B: There is nothing at all even potentially controversial about the part I restored. (I really don't care about the part that I did not restore, nor do I see how anyone could find it objectionable enough to have reversion fit about it, but oh well, whatever.) The part I did restore is plain and 100% factual. Some boneheads simply do not understand how the citation templates work, nor grok that the application of these templates must be consistent and sane, nor clue in to the fact that one cannot willy-nilly re-purpose p[arameters that already have particular, defined meanings across this entire family of templates. My one-sentence parenthetical addition helps fix that, at least with regard to one very frequently misused parameter in this particular template.

I really don't care to argue about this any further, and won't bother watching this discussion very closely. I generally trust that reason will prevail. If it didn't WP would already have collapsed under the weight of its own bullshit, yet it is going strong. If someone demands my further attention here, you know where to find me. :-)

SMcCandlish [talk] [cont] ‹(-¿-)› 14:35, 20 November 2009 (UTC)

This is starting to be really aggravating. I think that this is a great topic to have a discussion about, and it's something that I've wanted to bring up for quite a while. This simply does not seem like the best solution to what I think we both agree is a problem, is all. I'd love to come up with a solution that satisfies everyone.
V = I * R (talk to Ω) 14:43, 20 November 2009 (UTC)
There's nothing worth discussing. work (which is sometimes renamed, e.g. journal) and publisher (which isn't), have consistent definitions across this entire family of templates. To wit, work and its synomyms means a publication or presented work, in one medium or another, such as a book, magazine, TV show, blog, whatever. A publisher is the business entity that produced it, be that Hougton-Mifflin, New York Times Company, eBay Inc., self, or whatever. Very simple, very stable, very moot to argue about. — SMcCandlish [talk] [cont] ‹(-¿-)› 08:00, 23 November 2009 (UTC)

Good example

That is a good example. I would add that to the present text, without deleting any part of it. First comes the explanation in words, then the example. Debresser (talk) 11:01, 20 November 2009 (UTC)

Redrose64, why disable the internal link? We don't usually do that in documentation pages. Debresser (talk) 11:22, 20 November 2009 (UTC)
Have a look at my previous version - something has acted to take part of the |title= field and make that part of the URL; the arrowed box went in between "Han" and "Dynasty". If you click it, it throws a "Bad title" error.
You'll also notice that I left |location= out... that is because I'm unsure whether to put "San Francisco, CA" or "Los Angeles, CA". See foundation:Contact us - which address is relevant? --Redrose64 (talk) 11:29, 20 November 2009 (UTC)
I'd say that is less important and can be left out. Debresser (talk) 12:23, 20 November 2009 (UTC)
What is wrong with {{Cite web |url=http://en.wikipedia.org/wiki/Han_Dynasty%7Ctitle=Han Dynasty|work=Wikipedia|publisher=Wikimedia Foundation, Inc}}? Debresser (talk) 12:26, 20 November 2009 (UTC)
Very strange. I'll ask some people to have a look at this. Debresser (talk) 12:28, 20 November 2009 (UTC)
(ec) The MediaWiki parser happened. URL detection happens pretty late, after all templates are expanded, and it just includes everything it finds and accepts as URL characters. You'll have to trick it, by using e.g.
{{Tlx |Cite web |url{{=}}http://en.wikipedia.org/wiki/Han_Dynasty<i/> | title{{=}}Han Dynasty |work{{=}}Wikipedia |publisher{{=}}Wikimedia Foundation, Inc}}
{{Cite web |url=http://en.wikipedia.org/wiki/Han_Dynasty| title=Han Dynasty|work=Wikipedia|publisher=Wikimedia Foundation, Inc}}.
Amalthea 12:39, 20 November 2009 (UTC)
I kind of sectioned this off from the conversation above because, frankly, I'm somewhat confused about what you guys are talking about. I don't think that there's anything wrong with the example, but... it's also somewhat contrived to fit the desired mold of an ideal reference. There are many websites that decidedly do not fit well with this. Websites where they may be re-publishing others materiel is one good example. Regardless of that though, this seems to be going off on a slight tangent, talking about links and the location parameter.
V = I * R (talk to Ω) 12:35, 20 November 2009 (UTC)
I just wanted to demonstrate that a website could have both a |work= and a |publisher=, and that those could be different. My demonstration kind of fell over, so I amended it (diff here). To see the effect pre my amendment, try clicking the URL in Debresser's 12:26, 20 November 2009 edit.
The |location= is just another parameter which goes with the |publisher=: see {{cite book}} where it's more commonly found. All publishers must have an address; it's somewhere to serve libel writs, if nothing else. --Redrose64 (talk) 13:15, 20 November 2009 (UTC)
Ah, the diff makes things much clearer. Thanks. I do understand the point with location, but I actually think that it highlights well the reason that there are different templates and that the instructions are different. It's not even nearly as clear cut, in the case of websites, who the publisher/work is (if their different at all) and what location is really relevent, as compared to book/magazine/physical publishing. "Real" publishing is much more structured then web publishing is, so I don't see how attempting to shoehorn the less structured web publishing paradigm into the physical publishing paradigm is really helpful.
V = I * R (talk to Ω) 13:37, 20 November 2009 (UTC)
Anyway, more eyeballs could only help.
V = I * R (talk to Ω) 12:36, 20 November 2009 (UTC)

Yes it's probably template expansion or parser functions doing something nasty. Try:

Rich Farmbrough, 12:50, 20 November 2009 (UTC).

Thanks to users Rich Farmbrough and Amalthea for the two valid solutions.
I can tell you for a fact that using both a "work' and "publisher" is common enough to keep both parameters, and to be clear about the difference between them. Debresser (talk) 13:43, 20 November 2009 (UTC)
 :( This is getting to be a bit disapointing. Do you have statistics to back that statement up, at all? I at least qualified my observations as personal observations. One important issue to keep in mind here is that there are slightly different usage patterns in place between several topic areas.
V = I * R (talk to Ω) 13:56, 20 November 2009 (UTC)
I wrote in plain English "I can tell you for a fact". Those are my personal observations over half a year of intensive wikignoming with more than 10,000 edits in that area alone. For more precise statistics, consult a datadump. Debresser (talk) 14:07, 20 November 2009 (UTC)
Using a phrase that includes the term "fact" hardly qualifies your statement as a "personal observation". "I can tell you for a fact" is an assertion of absolute... well, fact, as contrasted by opinion. As I touched on above, this perception is likely due to the areas of Wikipedia which you have received the most exposure to, and it's always a good idea to question the general applicability of a self selected sample like that.
V = I * R (talk to Ω) 14:20, 20 November 2009 (UTC)
I scanned the dump from 26 October and found 1,724,255 instances of {{cite web}}. From those, 225,823 had only the work parameter specified, 900,711 only publisher and 232,195 had both. Svick (talk) 01:58, 21 November 2009 (UTC)
I second Debresser's obversations based on some 80,000+ edits, and my own editing style. I use both publisher and work and would be annoyed and disappointed if either was removed. Go take a look a few featured articles and featured lists for plenty of actual observations and usage patterns. IMHO - work should be used when the site is the site of a published work, such as New York Times, online book chapters, etc that would be italicized in a normal citation. Publisher is the publisher of the content and/or name of the website (as a plain website name, like Wikipedia), should not be italicized. -- AnmaFinotera (talk · contribs) 14:18, 20 November 2009 (UTC)
There you go, you stated the exact reason that I reverted back to the historical documentation text and (re)started this discussion: "Publisher is the publisher of the content and/or name of the website (as a plain website name, like Wikipedia), should not be italicized.". The way that the new text is written completely contradicts this, which in my view is a very standard viewpoint of editors.
V = I * R (talk to Ω) 14:23, 20 November 2009 (UTC)

New parameter?

{{editprotected}} What about leaving the work and publisher parameters completely alone, and adding a "website" parameter? Or something similar? It seems to me that would basically sidestep this whole problem and address everyone's concerns.
V = I * R (talk to Ω) 19:57, 20 November 2009 (UTC)

How would the website parameter differ from the url parameter? --Jc3s5h (talk) 21:07, 20 November 2009 (UTC)
If |website= were added, I propose that it be treated as the equivalent of |newspaper= in the {{cite news}} template, i.e., a specific case of |work=. I think that's what the proposer intended. Updating a previous example:
  • {{cite web |http://en.wikipedia.org/wiki/Han_Dynasty|title=Han Dynasty|website=Wikipedia|publisher=Wikimedia Foundation}}
When the work (website, newspaper, journal) is commonly known, we can omit the publisher:
  • {{cite web |url=http://en.wikipedia.org/wiki/Han_Dynasty|title=Han Dynasty|website=Wikipedia}}
Web sites are sometimes arbitrarily split into subsites and are sometimes informally named, but even so |website= seems likely to be used correctly more often than |work=. — John Cardinal (talk) 02:18, 21 November 2009 (UTC)
Exactly. The point about websites sometimes being arbitrarily split into sub-sites is something that I was attempting to point out originally as well, and is a large part of the basis for my objection to the documentation update. With all three fields available we should be able to accommodate all of the organizational situations and deal with the "the publisher is not the website!" criticisms.
V = I * R (talk to Ω) 10:48, 21 November 2009 (UTC)
Adding (stress this word) a |website= parameter would meet my full support. Debresser (talk) 17:24, 21 November 2009 (UTC)
Seems like a very reasonable solution to me as well. --RL0919 (talk) 17:52, 21 November 2009 (UTC)
Seems like we have a consensus then, so I added an edit protected request.
V = I * R (talk to Ω) 20:12, 21 November 2009 (UTC)

Undecided. Whether I would support this would depend on what the documentation said about it, and how it is handled with respect to |work=. If |website"= is intended to be a "a specific case of |work=" as John Cardinal suggests, what do we do if someone specifies both?

Suppose we put up a big red error message if both |website"= and |work= are specified. What, then, should the documentation say about it? How about something like this:

  • website: Specify |work= or |website"= but not both. Use |website"= when the website is a cohesive work of authorship, such as a dictionary, encyclopedia, or collection of related smaller works. If the domain name in the |url= parameter is an arbitrary name that provides no insight as to who is providing the information, consider using |website"= to provide context.

Jc3s5h (talk) 20:26, 21 November 2009 (UTC)

The design is grossly incomplete and a change request is premature. Jc3s5h (talk) 20:27, 21 November 2009 (UTC)

Agreed. Website and work are the same idea, so if somehow somebody would specify both (and people always manage to do the impossible), what should we do. I propose ignoring the work parameter. Debresser (talk) 20:45, 21 November 2009 (UTC)
If someone specifies both, we could ignore one or the other or concatenate the two values. My preference would be to ignore the |work = parameter.
I edited the sandbox to support |website=. If if both |website= and |work= are specified, |work= is ignored, no error message. (I am not opposed to a message, but I am not sure it's worth the trouble, either.) You can see |website= in action on the testcases page.
With regards to documentation, how about this, which is an edited version of what Jc3s5h wrote above:
  • website: Specify the title of the web site if the web site has a title and the web site is a cohesive work of authorship, such as a dictionary, encyclopedia, or collection of related smaller works. If you specify both |website= and |work=, |work= is ignored.
I chose to omit the bit about the domain name because I think it confuses the issue; even if the domain name in the URL clearly indicates the title of the web site, the URL is not visible. — John Cardinal (talk) 22:06, 21 November 2009 (UTC)
John Cardinal's documentation is good. I suppose that documenting that |work= is ignored whenever |website= is present will allow users to figure out what happened to their work parameter even without an error message. --Jc3s5h (talk) 22:16, 21 November 2009 (UTC)
Support the initiative and the way it is being worked out. Debresser (talk) 22:23, 21 November 2009 (UTC)
On further consideration, I suggest extending the documentation as follows:
  • website: Specify the title of the web site if the web site has a title and the website is a cohesive work of authorship, such as a dictionary, encyclopedia, or collection of related smaller works. If you are citing the entire website rather than some component of the website, give the title of the website with |title= and omit |website= . If you specify both |website= and |work=, |work= is ignored. Jc3s5h (talk) (sig added by John Cardinal)

(outdent) I'm fine with that documentation; specifying what to provide for a general reference to the website seems like a good idea. Alternately, we could adjust the template to do what the documentation suggests, i.e., if |title= is missing, use |website= as the title and do not use it as the work. In that case, we could adjust the documentation to describe that behavior. — John Cardinal (talk) 22:30, 21 November 2009 (UTC)

Update - I changed the sandbox to use |website= as the title if |title= is missing. See the testcases page.
Should we handle the "website but no title" case differently, formatting the website as a title (italics, no quotes) and make it a link? By way of comparison, if we cite part of a book, we would format it as "Genesis", The Bible. If we omit "Genesis", we don't change the way the title of the book is rendered. {{cite web}} currently needs a title to make a link, but we could change that behavior and use the website value as the link text. — John Cardinal (talk) 23:02, 21 November 2009 (UTC)
I think the documentation should encourage to use the title parameter when referencing the entire web site. If the title parameter is missing, I think the best guess is to treat the website parameter as if it were the title parameter, but that behavior should be documented. --Jc3s5h (talk) 23:12, 21 November 2009 (UTC)
This already seems like a bad idea to me. A website name is no title. The title parameter has its own error detection if unspecified. Don't carry on too far with what started as a good idea. Debresser (talk) 23:13, 21 November 2009 (UTC)
Debresser, suppose the website parameter existed I was going to cite a dictionary definition from the website http://www.merriam-webster.com. The title of that website is Dictionary and Thesaurus - Merriam-Webster Online. What parameters would you suggest using? --Jc3s5h (talk) 23:22, 21 November 2009 (UTC)
|title=Dictionary and Thesaurus - Merriam-Webster Online |website=www.merriam-webster.com. Or |title=Dictionary and Thesaurus |website=www.merriam-webster.com |work=Merriam-Webster Online. So as you see the website is not the title, ever. Debresser (talk) 23:36, 21 November 2009 (UTC)
I'd go with the second of these, with the addition of |publisher=Merriam-Webster, Inc and possibly |location=Springfield, MA --Redrose64 (talk) 23:41, 21 November 2009 (UTC)
I have withdrawn the above statement because I believe that it's been misinterpreted. I'm not in favour of adding |website=, just saying that if such a parameter were added, then |work= should still be used, and |publisher= ought also to be given. --Redrose64 (talk) 11:06, 23 November 2009 (UTC)

(outdent) Debresser and Redrose64's approach would, in most cases, just make the domain name of the url parameter visible. This is a departure from Wikipedia precedent. If that's what we wanted to do, we could just make the url parameter visible and clickable, and not make the title clickable.

If there is a consensus to make this change, it should be automated; the domain name should be extracted from the url automatically. Criteria should developed for cases where this is not appropriate, the criteria should be documented, and the website parameter should only be used when the criteria apply. If the website parameter is present, the automatic extraction and display of the domain name should be supressed. Finally, such a change would be equally applicable to all the other cite templates, so consensus should be sought in some unified place, and all the templates should be changed at the same time. --Jc3s5h (talk) 00:00, 22 November 2009 (UTC)

This is not completely true. Many web addresses use subdomains. But you are raising a strong argument against the use of this website parameter. Debresser (talk) 00:04, 22 November 2009 (UTC)
I fail to see why my suggestion would "make the domain name of the url parameter visible". The |title= field is there; that is what would be clickable, surely? --Redrose64 (talk) 00:17, 22 November 2009 (UTC)
What he means is that the website parameter would itself be the visible domain name. Debresser (talk) 00:41, 22 November 2009 (UTC)
Yes, Debresser, exactly. It might be worth noting that all the outside style guides such as APA Style call for the web address to be visible, which of course is the only alternative for paper publications. Wikipedia departs from all the paper-oriented guides by not making the address visible. I pity the poor scholar who has to type in certain web addresses from a piece of paper; I imagine it could take hours to get it right. --Jc3s5h (talk) 02:10, 22 November 2009 (UTC)

(outdent) Debresser, I disagree with your reaction to using the web site title as the title for the link. Many web sites are structured like books, albeit with pages that are roughly equivalent to chapters. Continuing with the analogy of a book, we allow editors to cite entire books now, with no reference to a page or chapter. I personally don't approve of such citations because it dramatically increases the difficulty for someone trying to verify the information. Still, if {{cite book}} allows us to cite a whole book, why doesn't {{cite web}} allow us to cite a whole web site? As it turns out, editors sometimes do this by using a URL that points to the main page of the site.

Using the URL of the site, or even just the domain name, is not equivalent to the concept of "web site" being discussed here. We should not equate the two. That's like saying that "135 Morrissey Boulevard Boston, MA 02125" is the equivalent of "The Boston Globe". Similarly, the domain name is not the name of the site. While some web sites adopt their domain name as their web site title, many don't, and in most of those cases, the name of the site is not a valid domain name.

I don't think we have to deploy either of the alternatives for using the name of the web site in place of the title parameter/page name, but if we do reject it, it ought to be for the right reasons.

One more thing: making the URL visible when viewing online is unnecessary. Web browsers insulate readers from that information for a set of good reasons. There is some utility in showing URLs when printing an article, but that can be controlled via CSS. — John Cardinal (talk) 02:40, 22 November 2009 (UTC)

P.S. Commenting on an example above:
No: |title=Dictionary and Thesaurus - Merriam-Webster Online |website=www.merriam-webster.com |url=...
That's not how |website= should be used. It's not for the domain name or URL. It's for the name of the site. In the example above, the "Title" would usually be for a specific page, so this is more appropriate:
Yes: |title=Definition of "example" |website=Dictionary and Thesaurus - Merriam-Webster Online |url=...
What I was also suggesting is to allow this:
Maybe: |website=Dictionary and Thesaurus - Merriam-Webster Online |url=...
In both my examples, Dictionary and Thesaurus - Merriam-Webster Online" would be formatted as if it were the title of a book, and it would also be a link. — John Cardinal (talk) 02:54, 22 November 2009 (UTC)
Cardinal, your examples are not intelligible without a full, working example, including url. Go ahead and use the sandbox version. --Jc3s5h (talk) 03:42, 22 November 2009 (UTC)


OK. Start with the basics: |website= shown with both bad and good examples. These examples use the sandbox version to produce the result even though the code shows {{cite web}}.

Incorrect use of website= Code {{cite web |title=Dictionary and Thesaurus - Merriam-Webster Online |website=www.merriam-webster.com |url=http://www.merriam-webster.com/dictionary}}
Result "Dictionary and Thesaurus - Merriam-Webster Online". http://www.merriam-webster.com/dictionary/. 
  1. |title= should usually be more specific than the name of the web site, although if citing the entire site that would be OK,
  2. but even when citing an entire site, there's no value in supplying the hostname/domain name in the |website= parameter.
Better use of website= Code {{cite web |title=Dictionary and Thesaurus |website=Merriam-Webster Online |url=http://www.merriam-webster.com/dictionary}}
Result "Dictionary and Thesaurus". http://www.merriam-webster.com/dictionary/example. 
  1. |title= is used for the title of a section of the web site,
  2. |website= is used for the name of the overall site.
Intended use of website= Code {{cite web |title=Definition of the word "example" |website=Merriam-Webster Online Dictionary and Thesaurus |url=http://www.merriam-webster.com/dictionary/example}}
Result "Definition of the word "example"". http://www.merriam-webster.com/dictionary/example. 
  1. |title= is used for the title of the specific page,
  2. |website= is used for the name of the site.

We can replicate all the examples above using the current production version if we use |work= in place of |website=. I'll omit the comments because they're the same as above, except this time it's |work= that's used incorrectly in the first instance.

Incorrect use of work= Code {{cite web |title=Dictionary and Thesaurus - Merriam-Webster Online |work=www.merriam-webster.com |url=http://www.merriam-webster.com/dictionary}}
Result "Dictionary and Thesaurus - Merriam-Webster Online". www.merriam-webster.com. 
Better use of work= Code {{cite web |title=Dictionary and Thesaurus |work=Merriam-Webster Online |url=http://www.merriam-webster.com/dictionary}}
Result "Dictionary and Thesaurus". Merriam-Webster Online. 
Intended use of work= Code {{cite web |title=Definition of the word "example" |work=Merriam-Webster Online Dictionary and Thesaurus |url=http://www.merriam-webster.com/dictionary/example}}
Result "Definition of the word "example"". Merriam-Webster Online Dictionary and Thesaurus. 

Let's use a traditional media organization that also has a web site and use the |website= parameter.

Correct use of website= Code {{cite web |title=Health overhaul narrowly advances |website=The Boston Globe |url=http://www.boston.com/news/nation/washington/articles/2009/11/22/health_overhaul_narrowly_advances/}}
Result "Health overhaul narrowly advances". http://www.boston.com/news/nation/washington/articles/2009/11/22/health_overhaul_narrowly_advances/. 

So far, the examples have been uncontroversial, we're just allowing |website= as a synonym for |work=. I think we agreed that editors would probably use |website= more readily than they use |work=. I may be wrong that we were in agreement, however, because I did not expect editors to suggest using |website= to show a domain name or any other part of the URL.

When Jc3s5h suggested an addition to the documentation, "If you are citing the entire website rather than some component of the website, give the title of the website with |title= and omit |website=", my reaction was that if someone wants to cite an entire web site, they shouldn't have to move the name of the web site into a different parameter. The natural thing is to omit the more specific parameter (|title=), and so my suggestion was to allow |title= to be omitted if |website= was provided. If that were implemented without any other changes, it would look like this:

Using value of website= as link text (sub-optimal) Code {{cite web |website=The Boston Globe |url=http://www.boston.com/BostonGlobe}}
Result "The Boston Globe". 

I think that's sub-optimal because The Boston Globe should not be quoted plain text, it should be unquoted italic text:

Using value of website= as link text (better) Code {{cite web |website=The Boston Globe |url=http://www.boston.com/BostonGlobe}}
Result The Boston Globe

My opinion as of now:

  1. We should add the |website= parameter at least as a synonym of |work=.
  2. We should decide if we want to allow |title= to be empty if |website= is not empty.
  3. If we allow it, we should adjust the template to format |website= as the "work" and not as the "title", i.e., the web site name would be an external link, unquoted, and italic.

John Cardinal (talk) 06:22, 22 November 2009 (UTC)

My opinion at the moment:

  1. We could add the website parameter, but be prepared that many editors will use it precisely the way I did, www.sitename.com, and no documentation will help against that.
  2. I agree that the website parameter should be on a pair with the work parameter.
  3. I hold that the title parameter should not be involved with the website parameter at all.

Debresser (talk) 09:22, 22 November 2009 (UTC)

Regarding your #1, are you in favor of adding the parameter or not? I agree that some editors will misuse the parameter, but that's true of every parameter, including |title=, |work=, |publisher=: they are all misused now.
Regarding your #3, what is the rationale behind your opinion? How do you propose that someone should cite an entire web site? — John Cardinal (talk) 17:50, 22 November 2009 (UTC)
  1. I am neutral. I see something good in it, but also forsee much incorrect use.
  2. If I understand your question correctly that would be title=website. Debresser (talk) 21:39, 22 November 2009 (UTC)
A new |website= is completely pointless; that is what |work= is for. Argh. This makes as much sense as created a new |company= field to serve the same purpose as |publisher= and adding in a |writer= parameter to reduplicate the function of |author=. Yeesh. — SMcCandlish [talk] [cont] ‹(-¿-)› 08:00, 23 November 2009 (UTC)
Is |newspaper= completely pointless, too? — John Cardinal (talk) 23:34, 23 November 2009 (UTC)
It certainly would be at Template:Cite news if that template already had implemented |work=, as Template:Cite web has. If you simply want to implement |website= and/or |site= here as alternative synonyms of |work=, then what exactly is the point? It wouldn't actually serve any needed purpose or address any demonstrable problem, and would be crufty and unnecessarily confusing given how many parameters these templates already have. Regardless, if |website= were implemented, it absolutely positively should not be a new parameter, with new code, simply an "if" that points to the same code as |work=. Cf. KISS principle. — SMcCandlish [talk] [cont] ‹(-¿-)› 22:58, 26 November 2009 (UTC)
The point is, many editors misuse the current parameters, so it is broke. For a template named "cite web", |work= is not as clear as |website=. Will people use it improperly, for example, will people use it to specify the domain name or some other part of the URL? Perhaps, but I believe it will be less than the number that now use |publisher= when they should use |work=.
Regarding how {{para}|website}} would work, there are two proposals, the second of which includes the first and extends it. First, make it a synonym of |work=. Second, if |website= is provided and |title= is not, use the value of |website= (or |work=, it's synonym) as the link text. The second proposal was intended to cover the case where someone cite an entire website.
You've linked a lot of WP pages above, but they don't help establish why adding |website= is a bad idea; they just make some of your opinions into links. — John Cardinal (talk) 03:22, 27 November 2009 (UTC)
Have thought on this further. I'll support adding |website= as, and only as, synonymous with |work= (and even making it the main term for that parameter, as |newspaper= and |journal= are for their respective templates). It should not be used as a replacement for |title= under any circumstances, since citing an entire website isn't any more appropriate than citing an entire newspaper. If the average reader/editor cannot find and verify the citation, then the citation is simply worthless gibberish. Last I looked |title= is mandatory for every single cite-family template (in many cases, the sole required parameter). {{Cite web}} doesn't deserve some super-special exception. As an aside, if someone really cannot follow how the concept of not introducing unnecessary complexity relates directly to the concept of the KISS principle, or how adding duplicative "solutions", to "problems" that haven't clearly been shown to need addressing,m relates to the idea that we don't "fix" things that aren't actually broken, I'm not inclined to explain it to them, and perhaps they should go do something else. Or maybe they understand just fine and could be more productive by dropping the sarcasm and taking what I wrote at face value instead of looking for ways to be offended by it so they can respond with defensive umbrage. Heh. :-) Seriously, as long as the template ends up working like the rest of them, I don't have any real issue. If people want |website=, fine, it just needs to work exactly like |journal= and |newspaper=, or people will do worse-than-useless things with it. And |work= needs to continue to work, even if someone were willing to bot away all deployed uses of it, since frankly I doubt that most editors are ever going to bother to remember a growing number of template-specific alternatives to {[para|work}}. — SMcCandlish [talk] [cont] ‹(-¿-)› 05:26, 27 November 2009 (UTC)
I am fine with not allowing |website= to substitute for |title=. Another editor raised that point. I would rather dissuade editors from citing an entire web site.
Regarding your other comments, I didn't respond with "defensive umbrage". Your links to WP:BROKE, WP:CREEP, and WP:KISS did not help your argument and were unnecessary at face-value: I wasn't going to follow the links and read those pages and I doubt anyone else participating in the discussion would either. All those links did was convert your opinions into blue-text. I also wasn't being sarcastic; your comments didn't make a persuasive case for why |website= should not be added.
If we stick to the meat of the argument, the addition of |website= should depend on whether or not doing so will improve citations to web pages. If we think it will, then the objection that adding the parameter adds unnecessary complexity doesn't hold water; if it improves the citations, then it's not unnecessary: the primary goal of the citation templates is to improve citations. Similarly, while I'm all for consistency, if being consistent leads many editors astray, it's a foolish consistency, the type that is the hobgoblin of little minds, and meanwhile, other templates have parameters that are similar to |website=.
Lastly, wording your reply as a response to an unnamed "someone" so you can include less-than-civil comments like "perhaps they should go do something else" reflects poorly on your consensus-building skills and is unlikely to help convince me or anyone else to agree with you. — John Cardinal (talk) 17:04, 27 November 2009 (UTC)
Why not simply improve the documentation? Maybe to make things clearer it should explain what the various parameters do. One creates a hyperlink and encloses in quotation marks the other italicizes etc. It would also be handy to have direct links to pages that describe the rules of citation formatting used by the template in the documentation itself. What formats are compatible with the template? Is it using APA, MLA, Chicago, etc. I've also found the need to switch from the documentation of {{cite web}} to {{cite book}} to {{cite journal}} looking at the examples to compare, see how they differ, and determine the best parameters and template to use. Having a direct link or index to the documentation of each instead of having to go through the category page would help facilitate the process when overhauling important articles with many cited sources. I've tried to make citations for a country article with over 150 sources more consistent and uniform for example and it is a pain just to get 20 fixed. Lambanog (talk) 05:15, 9 December 2009 (UTC)

Error message

There is a check to ensure that |title= is defined. It is now showing rather oddly:

{{cite web |url=http://example.org}}

http://example.org.  Missing or empty |title= (help)

---— Gadget850 (Ed) talk 23:17, 26 November 2009 (UTC)

This needs to be fixed: 1. the external link in front of the word "Cite web" is completely out of place. 2. it is also simply redundant, because the link follows at the end. Debresser (talk) 15:26, 9 December 2009 (UTC)
I fixed this in the sandbox. Can somebody make sure I did it right (the testcases look good to me)? The change I did is that whenever the error message would show up, it won't pass an url to {{Citation/core}}. Svick (talk) 00:50, 10 December 2009 (UTC)
That change doesn't look right, since it doesn't output the link at all. The link should be output, so that the poor reader can see what's being linked to (albeit without a title, and with a diagnostic). Eubulides (talk) 00:55, 10 December 2009 (UTC)
I am of the same opinion. The title should be linked, but only once, and in the right place. Debresser (talk) 01:23, 10 December 2009 (UTC)
What about current sandbox version? I'm not sure that adding the url to PS parameter of {{Citation/core}} is the best solution, but I couldn't think of anything better. Svick (talk) 10:26, 10 December 2009 (UTC)

Quotes in wrong position bug

It may be pedantic but I noticed that the way WP displays the results of this template looks clumsy and wrong to me.
Using the example on Waterbeach:

* {{cite web|author=Oliver Merrington|title=Waterbeach|url=http://homepage.ntlworld.com/oliver.merrington/waterbeach/|accessdate=10 December 2009}}
Oliver Merrington. "Waterbeach <icon>". Retrieved 10 December 2009.

To my mind the second quote (quotation mark) should be before the icon. What do other people think?--Lidos (talk) 09:51, 10 December 2009 (UTC)

see also archived discussion (July 2008)--Lidos (talk) 09:56, 10 December 2009 (UTC)
I think that it's nothing to do with {{Cite web}} but is in fact the way that {{Citation/core}} uses {{Citation/make link}}. View the source for {{Citation/core}}, look for section "Title of included work". You will see:
{{#if:{{{Periodical|}}}||"}}{{Citation/make link

and later in the same section:


so the quotes are added outside of the {{Citation/make link}} (provided that nothing had filled in |Periodical=). This is probably best discussed at Template talk:Citation/core. --Redrose64 (talk) 10:29, 10 December 2009 (UTC)

All discussion (and a solution) is now at Template talk:Citation/core .--Lidos (talk) 09:13, 11 December 2009 (UTC)

Possible error in placement of Date when Author is not used.

Maybe this has been discussed before or there is a reason for it, if so please just direct me to the proper discussion or explain it, but I think the date is improperly placed when an author is not given.

The usual case is

Author (Date). "Title". Work. 

however if an author is not given what happens is

"Title". Work. Date. 

shouldn't it be

→ "Title". (Date). Work.

and if we are to follow some style websites I've seen they would recommend

→ "Title". (Date). In Work.

Is there a way to add that In there without any formatting? — Lambanog (talk) 11:15, 19 December 2009 (UTC)

The order that the fields are output in is defined by {{Citation/core}}; and there have indeed been previous discussions. The most recent is at Template talk:Citation/core#Escalated render error: items lacking author.; but I suspect that'll go the same way as the others - something along the lines of "this is the Wikipedia style, agreed by consensus; and further consensus is required to change it".
An "In" is sometimes shown; again, that's down to {{Citation/core}} and is only done in certain circumstances - I think primarily when a cited work has both author and editor (or both last and editor-last). It's also used when the language is shown (ie "in French", etc.). --Redrose64 (talk) 12:32, 19 December 2009 (UTC)

→ "Title". (Date) Work.

I would say. The stop after (Date) is only good if (Date) is parenthetical to Title. Or even

→ "Title" (Date), in Work.

Rich Farmbrough, 05:37, 23 December 2009 (UTC).

Dateformat parameter

Am I right in thinking that the "dateformat" parameter doesn't do anything any more? My memory is vague, but I seem to recall that it had something to do with autoformatting of dates, which is now deprecated, right? If I come upon it in a citation that I'm editing, can I just remove it, or does it still serve some purpose? —Josiah Rowe (talkcontribs) 18:13, 16 December 2009 (UTC)

Yes. |dateformat= is passed on to {{Citation/core}} via the latter's |DateFormat= parameter - which is ignored. Removing it from uses of {{cite web}} will not change the behaviour of that template in any article; but should it ever be reinstated, results might not be predictable. --Redrose64 (talk) 20:04, 16 December 2009 (UTC)
Just remove it. I have deprecated it, but not yet removed the dummy passing. If it is reinstated in core there is no guarantee it will work the same way. Rich Farmbrough, 14:48, 24 December 2009 (UTC).

Page title for dynamic content?

Do we have a style convention for the following situation: the http://www.olympic.org/en/content/All-Olympic-results-since-1896/ page can be used to query a database with a specific set of results, but the page name is always "All the Medallists since 1896", so that's what I use for the title parameter to this template. Is there a suggested way to display the URL parameters used to create the dynamic URL? For example, if I link to http://www.olympic.org/en/content/All-Olympic-results-since-1896/?AthleteName=&Games=1333952&Country=347137&TargetResults=true to show all the Jamaican medallists from 2008, the page title doesn't show "Jamaica" or "Beijing 2008", but it would certainly be desirable to see that in the display output from this template. See Template:IOC medals/doc#Examples for more examples of what I mean. The output of cite web is the same, but the four links show very different results. Thanks — Andrwsc (talk · contribs) 22:50, 7 January 2010 (UTC)

I think you should use descriptive title, and it doesn't have to follow the real title of the linked page. Svick (talk) 23:15, 7 January 2010 (UTC)

Proposal: Add website= as synonym for work=

Simplifying a somewhat-stale proposal: I propose that we add |website= as a synonym for |work=. Using a more specific term will be more intuitive to editors who don't know the intent of |work=. If an editor specifies both |website= and |work=, |website= would have precedence. — John Cardinal (talk) 14:34, 9 December 2009 (UTC)

Agree. Just be warned that a lot of people will write "www.website.com" instead of the name of the website. Debresser (talk) 15:21, 9 December 2009 (UTC)
Support. It's a useful move in the right direction, even if people do just write "www.website.com" quite often. Rjwilmsi 17:57, 16 January 2010 (UTC)
Suggestion: Why not call the parameter something else. Instead of "website=" why not "sitename=" which might be differentiated later on if other parameters need to be included from a future "siteaddress=" "siteurl=" "sitehome=" "sitemain=" or maybe "sitewww=". Of course an alternative is simply to make the documentation page VERY VERY clear. Lambanog (talk) 11:33, 19 December 2009 (UTC)
I could support |sitename=, but I think |website= is more consistent with other cite-family parameters like |newspaper= in {{cite news}}, i.e., that parameter isn't |newspapername=. — John Cardinal (talk) 16:03, 17 January 2010 (UTC)
Support, per Rjwilmsi. Goodraise 19:11, 16 January 2010 (UTC)

Italics for web site name and use of work= parameter

I am involved in or aware of some discussions where editors assert that some web site names should not be italicized in web citations produced by cite web. See, for example, this post where an editor is trying to avoid italics. Another editor argued that whether the text appears in italics "should derive from the underlying publisher of the web site".

I do not agree with that logic. The purpose of the italics is not to reflect the nature of a related organization, its parent, or someone's estimation of the main activity of the organization. The purpose of the italics is to make it clear that the source material is part of a larger work and to identify the name of that larger work. When an organization maintains a web site, and that web site has articles, it doesn't matter whether the main activity of the organization is to maintain a database, broadcast television news, or make widgets. The name of the web site should be specified via |work=*, and it's proper for the template to italicize that text given the citation style favored by the cite family of templates.

* I'd prefer to use a parameter with a more specific name, but we've failed to reach consensus about that.

We should encourage users to specify the name of the web site using |work= and not with |publisher=. Currently, |publisher= is misused in the great majority of uses I've seen. When editors think italics are necessary, they often specify the markup manually as in |publisher=''Time'', rather than the correct usage, |work=Time.

I think we should begin by clarifying the documentation for {{cite web}}. We should stress that while the minimum parameters are |url= and |title=, typical usage should include |work= and either |date= (if dated) or |accessdate= (if not dated). We should also stress that |work= is for the name of the web site, and give good examples (what to do) and bad examples (what not to do).

Further, I'd explain (A) that the site name should not be specified via |publisher=, (B) that the publisher is usually the name of a firm in the publishing business and not the name of a web site, newspaper, or journal that firm publishes, and (C) that the name of the publisher is usually not needed for web citations.

What do other editors think about this? — John Cardinal (talk) 05:12, 15 January 2010 (UTC)

Wholeheartedly agree. I've been trying to get people to understand that "work" is the website or publication title, and "publisher" is the parent company's name, for ages. So far, lots of people either simply don't care or refuse to acknowledge this, even though it seems like straightforward common sense. Agreed also with point C, that "publisher" has virtually no value wrt website citations. Really, the only time I ever use "publisher" is with books and press releases, and perhaps assorted unusual cases. Huntster (t @ c) 07:05, 15 January 2010 (UTC)
I disagree. A website name is not italicized anymore than a book chapter would be. The only time work should be used is in the case of an online version of a published work, such as Time or New York Times. For something like Facebook, however, the website name is the publisher name. -- AnmaFinotera (talk · contribs) 07:31, 15 January 2010 (UTC)
Agree with John Cardinal (though I think we still want |accessdate= for dated pages). I thought all this was already explained in the documentation, so perhaps it has to be made clearer. I'm also looking at a cleanup exercise on misuse of the |publisher= field. Rjwilmsi 08:31, 15 January 2010 (UTC)
Regarding |publisher=: I hunt around for a copyright notice, which might be at page foot or reached via an "About Us" or "Contact Us" link. If it says "© Xxxxx 2010", I use |publisher=Xxxxx. If I cannot find a copyright note, I leave |publisher= blank. --Redrose64 (talk) 09:36, 15 January 2010 (UTC)
Collectionian, why do you equate a web site with a book chapter? Is your position that because we call them "web pages", a collection of web pages, i.e., a web site, is akin to a chapter? If so, I think that analogy is flawed.
The natural analogy seems to be as follows: the first named element of the source material is the title of the web page, and web pages are often called articles (with WP itself being a good example). The web page/article is part of a web site, the larger work, the way a printed article is part of a journal or a newspaper, and the way a chapter is part of a book. The larger work is often published by a firm in the publishing business. — John Cardinal (talk) 14:02, 15 January 2010 (UTC)
According to the Wikipedia:Manual of Style (text formatting), italics should be used in works of art and artifice, books, feature-length films and documentaries, musical albums, paintings, periodicals (newspapers, journals, and magazines), sculptures, etc. The titles of websites are not listed, for example Allmusic is an online database, with no printed work like The New York Times. In this case Allmusic is the work and Macrovision is the publisher, but the work should not be in italics. I think this might encourage some users to wrongly put the website name into the publisher parameter of the template to avoid italicization. Frcm1988 (talk) 15:37, 15 January 2010 (UTC)
The MOS is primarily addressing the prose of the article. Wikipedia (unfortunately!) allows editors to choose from among various popular citation styles, and those styles include text formatting that supports the specific purpose of a citation. The decision of whether to italicize or not does not depend on the primary activity of the organization; the italics are an element of the citation style and indicate that the source is part of a larger work. — — John Cardinal (talk) 18:43, 15 January 2010 (UTC)
I mostly agree with the topic opener. This has been bugging me for a while as well. As correctly pointed out above, WP:CITE gives editors a lot of leeway, including the freedom to use a citation style that requires website names to be italicized. The fact that Wikipedia:Manual of Style (text formatting) does not list website names as items to be italicized does not mean that website names should never be. Personally, I don't care whether website names are italicized or not (though I prefer the former). What I can't stand are code abominations like |publisher=[[Allmusic]]. [[Macrovision]].
I have to disagree with the opener on two points though. I find an accessdate always to be desirable. If nothing else, it can be of help when the website goes offline. The other thing is the importance of the publisher. The question of "Who stands behind this website I'm citing?" is way more important than the question "What do they call it?". Goodraise 19:09, 16 January 2010 (UTC)

Relevant discussion

There is a discussion occurring at Wikipedia:Centralized discussion/Wikipedia Citation Style. Your participation would be appreciated.
V = I * R (talk to Ohms law) 23:30, 18 January 2010 (UTC)

Proposal 2 for machine-translation link

I find machine-translation links very helpful despite their significant limitations. I propose two possiblities for a machine-translation link parameter: direct links with a warning or a geohack-style bouncer. I have checked both prototypes below in Firefox 3.5 and in lynx.

1. (User:Cxw/Cite_web_trans)

2. (User:Cxw/Cite_web_trans2)

Either option would clearly mark the translation as not the original reference, in accordance with say where you got it. Option 2 would additionally avoid implying endorsement by WikiMedia, in the same way GeoHack does now.

An example of when machine-translation links are useful: Kanagawa prefectural demographics (Japanese) (Translate: Google, Babelfish) (cited in Kanagawa). The statistics are the same in either language, but are much more accessible to English readers with the machine translation. While users certainly can copy links into their favorite translation sites, in my opinion, easier access to the translations would be a valuable feature. What say you?

Cxw (talk) 17:42, 22 January 2010 (UTC)

The File:High-contrast-dialog-warning.svg icons in your examples look horrible. I am strongly against any proposal that would add such eye-catching, distracting and unprofessional logos in all non-english instances of {{cite web}}. — Miym (talk) 17:55, 22 January 2010 (UTC)

I agree the icon isn't great, but it was the least space-consuming way I could think of to warn people not to rely on the machine translation (as discussed in the previous proposal). Any alternative suggestions for the logo? I looked at Nuvola apps important yellow.svg High-contrast-dialog-warning.svg Human-dialog-warning.svg Warning icon.svg and picked the one I did as the lowest-contrast of the three (since it's uncolored).

For that matter, Option 2 wouldn't require any icon at all, since the warning would be on the toolserver page. In that case, the only visible text added to {{cite web}} would be "(Translate)" Cxw (talk) 18:12, 22 January 2010 (UTC)

Hidden access dates

WP:CITEHOW now says "Citations for World Wide Web articles typically include:...the date you retrieved it (invisible to the reader if the article has a date of publication: <!--accessed: date-->). However the typical web page which should be cited with {{Cite web}} does not have a publication date, and so the models for cut&paste use should not comment out the access date parameter.

Frankly i think it is better displayed even when there is a pub date -- Indeed I will normally display it even when using {{cite news}}, but that is a different debate. But it is IMO a bad mistake to make the efective default hidden for typical uses of this cite web. This change has not been discused here that I know of. DES (talk) 03:11, 24 January 2010 (UTC)

  • Sorry, I missed the part about commenting out accessdate only if the publication date is absent; I reinstalled the part of my change that commented out accessdate only in that case. This has been discussed at some length, on multiple occasions, at Wikipedia talk:Citing sources; please see Wikipedia talk:Citing sources/Archive 25 #Retrieval dates for online versions of old printed sources, again for what I think is the most recent version.
  • The current version of {{cite web}}'s documentation is both too complicated and incoherent. Too complicated, because it has six "common forms". Incoherent, because it suggests both |date= and an uncommented |accessdate=, contrary to WP:CITEHOW. I propose that we simplify this by removing |accessdate= from the common forms; this will shrink the number of common forms down to two. Another possibility is to also add 3 more common forms, one for each accessdate= style, but in each of these 3 common forms, omit |date=; that would keep the 5 "common forms" coherent, though it would still be pretty complicated.
Eubulides (talk) 06:17, 24 January 2010 (UTC)
  • Accessdate is widely used, so removing it does not make sense. Neither am I in favor of commenting it out. Debresser (talk) 06:24, 24 January 2010 (UTC)
    • I also dislike |accessdate= being a comment. However, the documentation here should not contradict the guideline for citing sources. If removing |accessdate= from the common forms is not in the cards, then the 5-form solution is the simplest I can think of that does not contradict the guideline. Eubulides (talk) 06:42, 24 January 2010 (UTC)
      • Access dates should never be hidden, nor is there any guideline nor policy for supporting that. I have reverted it. Consensus by usage clearly shows that it should be included and not hidden. Accessdate should not be removed, it is both widely used and has large consensus for use from recent discussions. Some people may dislike it, however it has been again and again that it is both in keeping with normal citation styels and highly useful to many readers and editors alike. -- AnmaFinotera (talk · contribs) 07:36, 24 January 2010 (UTC)
        • The previous comment is contradicted by the guideline: WP:CITEHOW says to comment out accessdate for web pages that have publication dates. Consensus for this guideline was duly established. If the guideline is wrong, it should be fixed. It is not good for this documentation to contradict the guideline. Eubulides (talk) 08:26, 24 January 2010 (UTC)
That discussion showed that there was NO consensus for this change at all, despite its being snuck into the guideline anyway. Some CSS was added to give readers the option to hide it, but there was no consensus to change CITEHOW, which should be corrected. -- AnmaFinotera (talk · contribs) 08:32, 24 January 2010 (UTC)
OK, then the guideline should be fixed. I will follow up at Wikipedia talk:Citing sources #invisible to the reader?. Eubulides (talk) 08:47, 24 January 2010 (UTC)
Thank you, AnmaFinotera. The only reason I didn't do the same is because Eubulides is such an aggressive editor. Debresser (talk) 12:52, 24 January 2010 (UTC)

interwiki links

I just want to check - there's no way of using interwii links with this template, is there? so to give a link even for a site that's in the wp:interwiki map, we still have to use, say, http://en.wikibooks.org/wiki/Pagename rather than wikibooks:Pagename.

I guess it's rare that a wiki would be cited as a reference, but just wanted to clarify. Thanks.. --Chriswaterguy talk 01:18, 20 February 2010 (UTC)

You don't have to use the url parameter, so something like {{cite web | title = [[wikibooks:Pagename|Pagename]]}} works: "Pagename".  Also, it may be better to use {{cite book}} if the cited work is a book. Svick (talk) 01:29, 20 February 2010 (UTC)
Chris, to be brief, another wiki, even/especially other mediawiki sites, should never be used as a source in an article. Huntster (t @ c) 03:52, 20 February 2010 (UTC)
Never is too strong a word. Some Wikipedia articles made the news, thus could be cited to reflect the status of the article at that time, and diffs could be provided to as examples of reversions made by external agencies and whatnot. It is also completely fathomable that you are giving a link to the wikipedia article on the reference you used, such as C. Amsler et al. (Particle Data Group) (2008). "Review of Particle Physics", Physics Letters B, 667 1. Headbomb {ταλκκοντριβς – WP Physics} 04:07, 20 February 2010 (UTC)
Valid point, I suppose. I was recalling WP:CIRCULAR, which I don't remember as allowing circular links in any circumstance in the past. Personally speaking, though, I would never cite Wikipedia in any situation, even regarding internal meta-issues. If it's worth referencing, someone else will have referenced it first. Huntster (t @ c) 04:59, 20 February 2010 (UTC)

Archive parameters

When using the archiveurl=, etc (I try to archive all refs in articles I work on, before they go off line, with WebCite. Because it can take months before archive.com caches it and it just makes logical sense, especially when trying to get the said article to FA), could a parameter be added to not show it's been archived until after the link goes offline? For example, deadurl=no so it shows the regular format and then deadurl=yes, it shows the archived? Thanks. —Mike Allen 22:06, 1 February 2010 (UTC)

One could simply comment out the URLs if they're not yet needed. Eubulides (talk) 03:07, 2 February 2010 (UTC)
With this <!-- -->? I guess that's simpler than including it in the template.. —Mike Allen 03:28, 2 February 2010 (UTC)

Preferred link option

I generally archive pages I cite, which gives two links for the citation. Currently, the title link goes to the archived version, rather than the original. Could you please add a parameter |preflink= which permits overriding this behavior? It puts unnecessary load on the archive, and when the original site goes offline, it's a matter of seconds to update. Heck, a bot could do it. ;) Paradoctor (talk) 12:48, 22 February 2010 (UTC)

Just saw the above message. I still prefer having a parameter, but I'll use the comment bomb approach meanwhile. Paradoctor (talk) 12:51, 22 February 2010 (UTC)

Cite blog

Shouldn't there be a separate template for citing blogs? {{Cite blog}} redirects here.Smallman12q (talk) 14:10, 28 February 2010 (UTC)

Setting aside the issues of blogs as reliable sources, what fields would a blog need that are not supplied by this template? ---— Gadget850 (Ed) talk 14:22, 28 February 2010 (UTC)

::Not sure really, I wanted a parameter/template for blogs so that it could be easier to sort out whats a blog/whats not...more of a technical matter.Smallman12q (talk) 02:43, 4 March 2010 (UTC)

Type parameter

Actually, I'd like to have a "type" parameter so that you could identify a "blog", "article", "review"...etc. Smallman12q (talk) 13:55, 4 March 2010 (UTC)
What purpose would such parameter serve? Especially when such categorization is ambiguous (e.g. “This article on my blog is a review of …”). Svick (talk) 14:06, 4 March 2010 (UTC)
Second that question. If the blog is actually a reliable source, then its a web source period and there is no need to differentiate it. Ditto a review. -- AnmaFinotera (talk · contribs) 14:24, 4 March 2010 (UTC)
Well...when you put it that way, I guess there's no point to it...Smallman12q (talk) 23:59, 4 March 2010 (UTC)

ref and /ref codings

I think it would be handy if the ref and /ref codings were put at the beginning and the end of the examples. Then they could be copied and pasted directly when editing. One less click in the fiddly job of doing citations properly would be a small relief. Michael Glass (talk) 22:32, 14 March 2010 (UTC)

Not everybody uses {{cite web}} in this way. If using shortened footnotes, all the {{cite web}} are gathered together at the bottom, and the <ref></ref> merely contains the shortened footnote, which might be a {{harvnb}} (see, for example, Charwelton railway station). I therefore consider that the provision of <ref></ref> might lead editors to believe that that was the only permitted method. --Redrose64 (talk) 20:04, 15 March 2010 (UTC)

url parameter not creating link

I'm trying to use this url in a {{cite web}}: http://www.jamestown.org/programs/chinabrief/single/?tx_ttnews[tt_news]=28482&tx_ttnews[backPid]=191&no_cache=1, but it is not creating a reference with a link to the url. Here's the full cite I have:
{{cite web|url=http://www.jamestown.org/programs/chinabrief/single/?tx_ttnews[tt_news]=28482&tx_ttnews[backPid]=191&no_cache=1 |title=Hu Jintao: The Bird that Keeps its Head Down|last=Yao|first=Jin (pen name)|date=November 21, 2001|work=China Brief (Volume: 1 Issue: 10)|publisher=The Jamestown Foundation|accessdate=28 March 2010}}
which produces:
Yao, Jin (pen name) (November 21, 2001). [tt_news=28482&tx_ttnews[backPid]=191&no_cache=1 "Hu Jintao: The Bird that Keeps its Head Down"]. China Brief (Volume: 1 Issue: 10). The Jamestown Foundation. Retrieved 28 March 2010. 
Any ideas on a workaround or solution? Thanks. —Joshua Scott (LiberalFascist) talk 19:42, 28 March 2010 (UTC)

Probably the brackets. See Help:URL for encoding. ---— Gadget850 (Ed) talk 20:03, 28 March 2010 (UTC)
Per Gadget850, encode the brackets in the URL i.e. use http://www.jamestown.org/programs/chinabrief/single/?tx_ttnews%5Btt_news%5D%3D28482%26tx_ttnews%5BbackPid%5D=191&no_cache=1 instead. Rjwilmsi
It might be an idea to similarly encode the equals signs, ampersands and question marks as well. See Percent-encoding. --Redrose64 (talk) 20:41, 28 March 2010 (UTC)
m:Help:URL is actually the better reference. ---— Gadget850 (Ed) talk 00:20, 29 March 2010 (UTC)


What do you do if you are citing a whole website. The name of the website is the work so what is the title (the page says it's required)? What do you do for date if you reference a website that is constantly updated and provides no date on when it was last updated? McLerristarr (talk) 05:09, 14 April 2010 (UTC)

The |title= parameter is mandatory, but |work= is optional, so if you only have one of them, put it in |title=. The value that you use for |title= should be related to the text that appears in the title bar of your browser when viewing the page; part of it might be more suitable for the |work=. So, to take a purely hypothetical case, at this very moment, the title bar of my browser states
 Editing Template talk:Cite web (section) – Wikipedia, the free encyclopedia – Mozilla Firefox 
so I would discard the browser's name, and split the rest at the dash, to give
|title=Editing Template talk:Cite web (section) |work=Wikipedia, the free encyclopedia
If it's undated, then the |date= should be left blank (it's optional); but if the page is "constantly updated", it's possibly not a WP:RS because if you come back to it after a period, it's not necessarily going to be possible to WP:VERIFY the reference. --Redrose64 (talk) 09:25, 14 April 2010 (UTC)
Well, that referencing example is more obvious. What would you do if you were (hypothetically) referencing the whole of Wikipedia? Would you put "Wikipedia" as the title?
It's not unreliable, it's not a wiki, it's edited by one person. "Constantly updated" isn't really true but it could be updated at any time and occasionally is if spelling mistakes are found for example, which isn't all that rare. There's no date of last update, so I'll just put 2010. McLerristarr (talk) 10:14, 14 April 2010 (UTC)
Yes, |title=Wikipedia with |work= either left blank or (better) omitted.
If all you have is a year, it should go in |year= - |date= is for full dates. --Redrose64 (talk) 11:08, 14 April 2010 (UTC)
I'd have to question why you would ever "cite" an entire website in a single citation? 99% of the time, if you are citing a reliable website, the cite should point to a specific page in that site. -- AnmaFinotera (talk · contribs) 13:00, 14 April 2010 (UTC)
Not true at all. For example, if you were writing about The Beatles, then a website about The Beatles would have many useful pages. It's silly to have to name all the ones you used, much easier to just cite the entire website.—Preceding unsigned comment added by Mclay1 (talkcontribs) 14 April 22:34 UTC
If the material that backs up a statement is really scattered all through a website, it might be appropriate to cite the whole website. But if the material backing up a statement is on a specific page, that page should be cited, just as a page in a book should be cited whenever possible. Jc3s5h (talk) 23:41, 14 April 2010 (UTC)
No, it would not be silly. Same as you cite the pages of a magazine or book for a printed source, you should be specific as possible. If the website has over 200 pages, expecting someone to go through all 200 to find where a single statement is made is silly. -- AnmaFinotera (talk · contribs) 01:09, 15 April 2010 (UTC)
It's not a matter of ease of citing, it's a matter of verifiability, so if a reader cannot verify what is written, it may be challenged or removed. I would encourage you to cite the specific page, especially if the information is in any way controversial or unexpected. —Joshua Scott (LiberalFascist) 04:12, 15 April 2010 (UTC)
I'm not talking about just a statement, I mean if you're using the website as a major source. It would be pointless to reference after every sentence or paragraph if it was the same source over and over. Especially if it was just a source that confirmed what another source said. I've seen whole websites referenced many times, they always put the title of the website as the title. I want to know if that's correct? I think there should be a better way of doing it. McLerristarr (talk) 09:35, 15 April 2010 (UTC)
It doesn't matter if it is used as a "major" source or a "minor" one. Same as a book. Reference each page on use. It is not pointless, it is proper and giving a correct reference. If an article is so totally dependent on and using a single website on paragraph after paragraph, that would indicate a major problem with the article and bring to question issues about validity of the article's contents, and possibly notability if the claim is that it is the only source. You may have seen it done, but that does not mean it should be done. The better way to do it that you are asking for is to actually properly reference the article with specific references to each page. -- AnmaFinotera (talk · contribs) 13:06, 15 April 2010 (UTC)

Short footnote to sub-pages of a site?

For instance, I include as a reference a Glossary which has about 20 sub-pages (corresponding roughly to letters of the alphabet). Throughout the article I reference material on various sub-pages. Currently I have a choice: I can either do a {{cite web}} in the References section for each and every page I use, or do a {{cite web}} in every <ref></ref>. Is there a technique where I can specify the sub-page in a Wikilink? Dmforcier (talk) 17:11, 15 April 2010 (UTC)

I don't know of anyway to do that with web cites like you can with book cites, though someone more familiar with the whole shortened form thing may have more ideas. I don't really see anything wrong with creating an individual reference for each page, however, same as noted above re the question on "citing a whole website". You can used named refs if you need to reference it more than once. -- AnmaFinotera (talk · contribs) 17:33, 15 April 2010 (UTC)
Something along these lines has been done on Mumbai, where the |p= parameter of the {{harvnb}} template has been used to create an external link for the specific page. The general idea is:
<ref>{{harvnb|Doe|2010|p=[http://www.example.com/example.asp?page=123 123]}}</ref>
*{{cite web |url=http://www.example.com/ |title=John Doe's example |last=Doe |first=John |year=2010 |ref=harv }}
Something very similar could be done with {{sfn}}. --Redrose64 (talk) 20:18, 15 April 2010 (UTC)
Thanks, I'll try that.
I'm not sure I get harvnb links in references to web sites. The last and year params connect the ref to the citation instead of a cite name. But when is a website published? Does the publication date change whenever the site changes? What would be the year for e.g. a Wikipedia article? In this case there is a copyright statement, so I'll use the latest c. date, but there are a lot of sites out there without such info. Dmforcier (talk) 18:09, 16 April 2010 (UTC)
Yes, in the {{cite web}} it's normally the |last=, or |last1= to |last4=, together with either |year= or |date=, that are used to construct the link for harvard referencing; but you can construct custom links. Many websites do give the current year for copyright purposes, even though any given page might not itself have changed in years. The year/date is, however, optional for harvard referencing; the following are valid:
<ref>{{harvnb|Doe|p=[http://www.example.com/example.asp?page=123 123]}}</ref>
*{{cite web |url=http://www.example.com/ |title=John Doe's example |last=Doe |first=John |ref=harv }}
Here are the above as working examples:
Statement A (year in ref).[1] Statement B (no year in ref).[2]
  1. ^ Doe 2010, p. 123
  2. ^ Doe, p. 124
--Redrose64 (talk) 19:09, 16 April 2010 (UTC)
year/date is optional? Cool. I'll remember that. Thanks so much for your help, guys. I'll be back if I have trouble. (In other words, "I'll be back.") Dmforcier (talk) 01:20, 28 April 2010 (UTC)


language=es generates (in es)

language={{es}} generates (in (Spanish))

A couple of parenthesis should be suppressed. HaŋaRoa (talk) 22:25, 11 May 2010 (UTC)

The parameter does not use country codes or templates, just plain text. Set |language=Spanish. ---— Gadget850 (Ed) talk 22:36, 11 May 2010 (UTC)
What about a new parameter as in "language template={{es}}"? The reason the {{xx}} templates exist is because it's tiresome to type [xx language|xx]. HaŋaRoa (talk) 22:42, 11 May 2010 (UTC)
This and the other {{cite xxx}} templates are based on {{Citation/core}}, so you can take it up there. I really don't think it is worth it to add yet another sub-template to save a few characters, especially when the bulk of references are in English and should not use the parameter. Why would you want to link the language in a citation? ---— Gadget850 (Ed) talk 22:59, 11 May 2010 (UTC)
And templates like {{es}} don't link either. Svick (talk) 23:25, 11 May 2010 (UTC)

Templates that use cite web as a backend

Hello all. I have authored a number of specific source citation templates such as {{cite gnis}}. They all transclude cite web. I experimented using Citation/core for a awhile but quite that approach about the time of the great date wars. I would like to get the advice of those with more insight than myself about which template to use (cite web or Citation/core). Is server load a significant factor. I seems like this issue is related to the double redirect thing. –droll [chat] 22:21, 15 May 2010 (UTC)

Well, {{cite web}} transcludes {{citation/core}} in any case, so if your template transcludes {{cite web}}, it will still use {{citation/core}} at a deeper level, and will be slightly slower than going directly for {{citation/core}}. Don't worry about double-redirects: it's a different mechanism - and in fact, if you examine some other templates, such as {{convert}}, they go through quite a few levels where different templates are transcluded before you get to the ultimate Wikicode or HTML.
If it's working using {{cite web}}, leave it as it is until somebody complains. --Redrose64 (talk) 15:04, 16 May 2010 (UTC)

documentation should mention template:rp for second use of ref but at a different page or section

E.g. Near the "page" documentation entry:

For subsequent references to the same webpage, but on a different page or section, use Template:rp. For example:

 <ref name=somewebpage>{{rp=page 56}}</ref>


 <ref name=somewebpage>{{rp=section 3}}</ref>

—Preceding unsigned comment added by Facts707 (talkcontribs)

I tried this, and there is no visual indication of the page number for the second reference. Please post a full working example and explain the value of this. Jc3s5h (talk) 16:34, 17 May 2010 (UTC)
(edit conflict) {{rp}} doesn't work with cite web. I just tested this here  --Joshua Scott (LiberalFascist) 16:35, 17 May 2010 (UTC)
The thing is, when you use name=somewebpage for a second time, its content is ignored and it is merged into the first. I use another method; for the first ref which cites a given url, make sure that the {{cite web}} has a |ref=, give it a unique value such as |ref=example.com and build the second ref like this:
Statement 1.<ref>{{cite web |url=http://www.example.com/ |title=Example.com home page |last=Doe |first=John |date=17 May 2010 |at=section 1 |ref=example.com }}</ref>
Statement 2.<ref>[[#example.com|Example.com home page]], section 3</ref>
which produces
Statement 1.[1]
Statement 2.[2]
  1. ^ Doe, John (17 May 2010). "Example.com home page". section 1. 
  2. ^ Example.com home page, section 3
If you click the second one (in Firefox or Chrome, not in IE), it takes you to the first. How's that? --Redrose64 (talk) 17:21, 17 May 2010 (UTC)
In that case, maybe you should use Harvard referencing:
Statement 1.[1]
Statement 2.[2]
  1. ^ Doe 2010, section 1
  2. ^ Doe 2010, section 3
It works similarly to your example, except that the name of the ref is generated automatically (using ref=harv in {{cite web}} and similar templates), you use {{harvnb}} to reference it and the reference to the whole site is among general references (not generated by {{reflist}}). Svick (talk) 11:22, 18 May 2010 (UTC)
Personally I almost always do use {{harvnb}}, or the even more versatile form provided by {{sfn}}. I was trying to explain short-note links in terms that expose, rather than hide, the mechanics, particularly since not every web page bears an author. --Redrose64 (talk) 12:01, 18 May 2010 (UTC)

{{rp}} doesn't work with cite web.

Your example is wrong. I don't know where you got the =. And you don't stuff it inside the <ref>...</ref> tags.

Example reference.<ref name=ex1>{{cite web |title=Example |url=http://www.example.org}}</ref>{{rp|5}} Another example.<ref name=ex1/>{{rp|7}} Yet another example.<ref name=ex1/>{{rp|42}}

Example reference.[1]:5 Another example.[1]:7 Yet another example.[1]:42

  1. ^ a b c "Example". 

use name=somewebpage for a second time, its content is ignored and it is merged into the first

It isn't ignored— the content is reused. If you are going to go to the trouble of defining the reference again, then you need to define it in full. Your example only works if there is no other references in between the two, else the second reference is missing information. ---— Gadget850 (Ed) talk 17:46, 17 May 2010 (UTC)

The content of the second is ignored, and the content of the first is indeed reused; but they appear in the same row in the {{reflist}}, so in that way, they are merged. My point was that whenever the content within the <ref></ref> is to differ from any previous use, then either the name=value must also differ, or be omitted entirely. --Redrose64 (talk) 12:01, 18 May 2010 (UTC)
If you define references with the same name, then yes, the second is ignored. This is by design, as allowing both would be ambiguous. RefToolbar has an error check for this.
Bottom line— to add page numbers:
Use <ref>...</ref> tags and define separate cites
Use <ref>...</ref> tags with a name and use {{rp}} or use {{r}}
Use shortened references
Use parenthetical references (Harvard)
---— Gadget850 (Ed) talk 16:02, 18 May 2010 (UTC)

URLs with square brackets

While trying to cite http://www.gmp-architekten.de/index.php?id=4&L=1&tx_mimpdb_pi1[showUid]=338&tx_mimpdb_pi1[filter_typology]=53&tx_mimpdb_pi1[typological]=1&cHash=004bd04c39 in the Haidian Christian Church article, I was left with no choice but to find one of the few URL shortening services not already blocked by our spam filters. Is there any supported mechanism for citing URLs that contain square brackets?   — C M B J   23:53, 25 May 2010 (UTC)

Yes, it is possible, but you have to use hex codes to accomplish the linking. See m:Help:URL#URLs in external links for help with figuring out what codes to use. In your case, the url would be http://www.gmp-architekten.de/index.php?id=4&L=1&tx_mimpdb_pi1%5bshowUid%5d=338&tx_mimpdb_pi1%5bfilter_typology%5d=53&tx_mimpdb_pi1%5btypological%5d=1&cHash=004bd04c39, and you can see where I replaced the brackets with hex codes. Link works fine now :) Huntster (t @ c) 01:27, 26 May 2010 (UTC)

Online articles spanning multiple pages

{{Cite web}} works great for "one page" web articles—but what if an article spans multiple pages, each page having its own URL? Should I technically be using one {{cite web}} template per URL, and therefore for each page? In which parameter would I specify the page number to avoid confusion to the reader? To add another layer to this question, the reason I'm asking is because I will be archiving a 6-page online article, and each page (already having its own original URL) will now have its own archive URL. – Kerαunoςcopiagalaxies 23:48, 10 June 2010 (UTC)

If there is not a printable version that has all six pages together, then yes, you could use a separate instance for each page. -- AnmaFinotera (talk · contribs) 23:55, 10 June 2010 (UTC)
Hmm, interesting idea. Reminds me of mobile pages that remove all the ads and jazz and just contain text. I'll check for that. – Kerαunoςcopiagalaxies 00:32, 11 June 2010 (UTC)
Re "In which parameter would I specify the page number to avoid confusion to the reader?" - there is a |page= parameter which may be used. There are also |pages= and |at=; but only one of the three can be used. --Redrose64 (talk) 10:22, 11 June 2010 (UTC)
Hey, that's fantastic! Didn't even occur to me to look through the parameters. I'll give "at=" or "page=" a try, thanks so much! – Kerαunoςcopiagalaxies 17:50, 11 June 2010 (UTC)

Could we have a vertical layout for the common forms, please?

Could we have a vertical layout for the common forms, please? Michael Glass (talk) 02:46, 17 June 2010 (UTC)


The parameter URL does not work, there seems to be no choice other than to intentionally misspell it as "url" (not "link", not "address", or any other proper words). Book citations does allow you to use the correct uppercase for ISBN and I had expected URL to work. Could someone please fix this so that URL also works? I'd be much obliged if "access date" worked too and I again I was not forced to misspell the words and unnaturally omit a space. Thanks in advance. -- Horkana (talk) 15:45, 12 May 2010 (UTC)

Please see Template:Cite web#Usage, which states
Note: None of the parameters should be capitalised, in order to avoid the article being tagged as having a broken citation. For example, use "url", "title", etc. - not "URL", "Title", etc.
The normal behaviour of Wikipedia templates is that the parameter names are case-sensitive, and further, parameter names are normally all-lowercase. The form shown in the documentation should always be followed; in this case it is described as
{{cite web |url=
in every instance. To allow case-insensitivity for parameter names would require the templates to be made much more complex, which would significantly increase load time. For a template as complex as {{cite web}}, load time is already at a point where reduction, not increase, of functionality is being requested. --Redrose64 (talk) 19:24, 12 May 2010 (UTC)
I agree. However, {{cite book}} does have several aliases, including isbn and ISBN, separator and seperator. Aliases also make maintenance tools and bots more complex as well. ---— Gadget850 (Ed) talk 19:43, 12 May 2010 (UTC)

I understand how it works, and have read the usage section. I am suggesting how it should work and saying it would be preferable to use proper words and not be forced to misspell things.
I did not say upper and lowercase should be allowed for all parameters just that editors should not be required to deliberately misspell. I make an extra effort to be careful and try to spell correctly and often spell check my edits too. It would seem better to avoid requiring editors to deliberate misspell things when (as demonstrated by ISBN) it is possible to do better. If I had the choice to use a more meaningful and simple real word such as address, or link, instead of a deliberate misspelling of what is a technical acronym such as "url" that too would be an acceptable solution, but the simplest solution would seem to be to allow URL.
When there are guidelines WP:PERF/WP:DWAP which specifically say do not worry about performance it seems strange when you say to worry about performance.
Similarly is it not the right thing to do to make things a little more complicated for maintenance tools and robots if it makes things easier for human editors or readers? -- Horkana (talk) 21:21, 12 May 2010 (UTC)

Ignoring performance is how you get articles like List of allied military operations of the Vietnam War, and it doesn't even use citation templates— imagine if it did. There has to be a balance. ---— Gadget850 (Ed) talk 22:40, 12 May 2010 (UTC)
Yes; "Served by srv274 in 59.547 secs." - not good. --Redrose64 (talk) 12:09, 13 May 2010 (UTC)
Readability and coherence would be the better reason to split that full A-Z list into other groupings, to serve the reader first and then worry about performance later. It isn't surprising that the guidelines hand down simplistic absolutes and lack subtlety, most guidelines seem to be decided by a very small consensus of whoever happened to be there on a particular day. I don't think anyone could disagree that making an article good for readers is a far more important priority before worrying about performance.
Anyhow that is beside the point, I reiterate my main issue that it is not a good thing to require deliberate misspellings when it could be easily avoided. -- Horkana (talk) 13:21, 13 May 2010 (UTC)
So you have a full list of aliases for all of the {[cite xxx}} templates? ---— Gadget850 (Ed) talk 14:05, 13 May 2010 (UTC)
No I'm not asking for full list of aliases, not at all. Cite book doesn't have a full list of aliases either but it does have allow both isbn and ISBN in keeping with the correct capitalisation. Similarly I'm asking that cite web specifically allow URL and possibly also "access date" so that correct spelling can be allowed. (It is a shame whoever wrote the first template didn't have the sense to use a normal word like "link" - much easier to understand than an acronym - and this would not have even been an issue.)
From the lack of understanding why this is even a bad thing it doesn't seem as if I'm going to get any help or advice on how to set up an alias to allow URL as well as url but I would appreciate if an editor would point me to the relevant documentation, maybe I can figure it out myself. -- Horkana (talk) 17:20, 17 June 2010 (UTC)
The "correct" spelling of parameter names with regards to English language usage is something that is unimportant: this is because the parameter names don't show in the final page, only in edit boxes. What is important is the correct spelling of the text in the articles as displayed.
Aliasing a parameter would mean editing the template source, searching for every instance of the existing parameter, and changing it to a form which allows the alternative form. Consider the existing |url= parameter. This is found in the template source as {{{url|}}}; this would need to be altered to {{{url|{{{URL|}}}}}} - there are two instances. |accessdate= is more difficult; there are three instances, and {{{accessdate|}}} would become either {{{accessdate|{{{access date|}}}}}} or {{{accessdate|}}}{{{access date|}}} depending on context and intent.
Such changes need to be carried out with great care: there is plenty of potential for error, since there are thousands of pages which use this template: we cannot break them by accident.
It's not so long ago (30 December 2008 to be precise) that we managed to get rid of |access-date=, in the interests of harmonisation with the other cite templates, so it's not likely that we'd want to bring back something that (a) causes disharmony and (b) is very similar to something removed with consensus. As for |isbn=: that doesn't occur in {{cite web}} at all. If your citation has an ISBN, you should probably be using {{cite book}}. --Redrose64 (talk) 17:59, 17 June 2010 (UTC)

"Work" is broken!

for some reason the "work parameter" is throwing out errors. see here for example.

[Expression error: Unexpected <  operator ...]

...could someone check on this? --emerson7 22:15, 7 June 2010 (UTC)

I don't think it's the |work= parameter, nor even {{cite web}}. I've copied the entire page to one of my sandboxes, and no errors are shown. Have tried WP:PURGEing the offending page, no luck. --Redrose64 (talk) 22:32, 7 June 2010 (UTC)
I don't think it is the template either. If you look at just one section with the <references /> tag, they all work fine in the same page. Looking at the article history, this edit seems to be the one that broke the page[28]. It was fine before then.[29] Not entirely sure what is broken in the new bit added, but when I tested undoing it, it fixed the errors, so may want to check that code more carefully. I tried removing the refs there themselves, and that didn't fix it.-- AnmaFinotera (talk · contribs) 22:51, 7 June 2010 (UTC)
I can't figure this problem out either. When I isolate that new material in a sandbox, it works fine. I didn't immediately notice a problem with the other citations with errors, but I only took a random sampling. Very curious. Huntster (t @ c) 00:44, 8 June 2010 (UTC)
I expect that there are too many templates on the page. Open the whole page for editing an look at the template list at the end. ---— Gadget850 (Ed) talk 01:12, 8 June 2010 (UTC)
Have seen some with more, though I suspect the article would benefit either way from dropping all the flagicons, which don't seem to meet WP:MOSFLAG are seem a bit gratuitous. If it is too many templates, that would certainly clear things out a bit. -- AnmaFinotera (talk · contribs) 01:23, 8 June 2010 (UTC)
i hardly think that's the problem. i should think the wiki software is far more robust than that. --emerson7 01:47, 8 June 2010 (UTC)
I agree, does not appear to be the problem. The limit report is no where near maxed out, and I've seen articles work fine with far more templates and more citations than that one. This could be anything from a random code problem to a server-side issue. Huntster (t @ c) 02:48, 8 June 2010 (UTC)
Fixed. Apparently you can't use {{dts}} twice in the same cell. Which makes sense since it is a sortkey. ---— Gadget850 (Ed) talk 15:53, 8 June 2010 (UTC)
Mmm. In which case, why didn't this edit throw a wobbly? --Redrose64 (talk) 16:48, 8 June 2010 (UTC)
Good question. I was sandboxing this, and can't reproduce it. ---— Gadget850 (Ed) talk 16:57, 8 June 2010 (UTC)
I'm working on MakeRef but not getting any help at en.WP or meta.wikimedia. In case of ecs, I hope {{inuse}} works at meta.wikimedia. If it matters, MakeRef is my top priority as I use it a dozens a week and want to recommend it as part of my standard toolbox. --Philcha (talk) 11:21, 17 June 2010 (UTC)
  • Fixed! for those interested, this problem has been resolved. apparently it was caused by syntax errors when entering the {{dts}} template. kudos to Johnson487682 (talk · contribs) who fixed it here! --emerson7 22:45, 18 June 2010 (UTC)

Square brackets in title are mangled with url

Example: *{{cite web |author=Leung K |url=http://www.ncbi.nlm.nih.gov/bookshelf/br.fcgi?book=micad&part=AV-45-18F |title=(E)-4-(2-(6-(2-(2-(2-(<sup>18</sup>F-fluoroethoxy)ethoxy)ethoxy)pyridin-3-yl)vinyl)-N-methyl benzenamine [[<sup>18</sup>F]AV-45] |work=Molecular Imaging and Contrast Agent Database |date=April 8, 2010 |accessdate=2010-06-24}} renders as:

I suspect this is partly a cite web issue and partly a Tidy issue, but I'm not sure how to escape the title to get it to render correctly. LeadSongDog come howl! 19:52, 24 June 2010 (UTC)

Replace the [ with %5B and replace ] with %5D. See Percent-encoding for more information. – Kerαunoςcopiagalaxies 20:23, 24 June 2010 (UTC)
so that

*{{cite web |author=Leung K |url=http://www.ncbi.nlm.nih.gov/bookshelf/br.fcgi?book=micad&part=AV-45-18F |title=(E)-4-(2-(6-(2-(2-(2-(<sup>18</sup>F-fluoroethoxy)ethoxy)ethoxy)pyridin-3-yl)vinyl)-N-methyl benzenamine %5B%5B<sup>18</sup>F%5DAV-45%5D |work=Molecular Imaging and Contrast Agent Database |date=April 8, 2010 |accessdate=2010-06-24}} renders as:

Thank you, but this doesn't seem like much of an improvement. LeadSongDog come howl! 21:05, 24 June 2010 (UTC)

Add two pairs of <nowiki></nowiki> to your original version:
*{{cite web |author=Leung K |url=http://www.ncbi.nlm.nih.gov/bookshelf/br.fcgi?book=micad&part=AV-45-18F |title=(E)-4-(2-(6-(2-(2-(2-(<sup>18</sup>F-fluoroethoxy)ethoxy)ethoxy)pyridin-3-yl)vinyl)-N-methyl benzenamine <nowiki>[[</nowiki><sup>18</sup>F<nowiki>]AV-45]</nowiki> |work=Molecular Imaging and Contrast Agent Database |date=April 8, 2010 |accessdate=2010-06-24}}
renders as:
--Redrose64 (talk) 21:38, 24 June 2010 (UTC)
Thank you. That is, at least, legible. Unfortunately the emitted html still puts the nowiki tags in the title field of the metadata as follows:

<ul> <li><span class="citation web">Leung K (April 8, 2010). <a href="http://www.ncbi.nlm.nih.gov/bookshelf/br.fcgi?book=micad&part=AV-45-18F" class="external text" rel="nofollow">"(E)-4-(2-(6-(2-(2-(2-(<sup>18</sup>F-fluoroethoxy)ethoxy)ethoxy)pyridin-3-yl)vinyl)-N-methyl benzenamine [[<sup>18</sup>F]AV-45]"</a>. <i>Molecular Imaging and Contrast Agent Database</i><span class="printonly">. <a href="http://www.ncbi.nlm.nih.gov/bookshelf/br.fcgi?book=micad&part=AV-45-18F" class="external free" rel="nofollow">http://www.ncbi.nlm.nih.gov/bookshelf/br.fcgi?book=micad&part=AV-45-18F</a></span><span class="reference-accessdate">. Retrieved 2010-06-24</span>.</span><span class="Z3988" title="ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=bookitem&rft.btitle=%28E%29-4-%282-%286-%282-%282-%282-%28%3Csup%3E18%3C%2Fsup%3EF-fluoroethoxy%29ethoxy%29ethoxy%29pyridin-3-yl%29vinyl%29-N-methyl+benzenamine+%7FUNIQ577ac5861f92d462-nowiki-00000010-QINU%7F%3Csup%3E18%3C%2Fsup%3EF%7FUNIQ577ac5861f92d462-nowiki-00000011-QINU%7F&rft.atitle=Molecular+Imaging+and+Contrast+Agent+Database&rft.aulast=Leung+K&rft.au=Leung+K&rft.date=April+8%2C+2010&rft_id=http%3A%2F%2Fwww.ncbi.nlm.nih.gov%2Fbookshelf%2Fbr.fcgi%3Fbook%3Dmicad%26part%3DAV-45-18F&rfr_id=info:sid/en.wikipedia.org:Template_talk:Cite_web"><span style="display: none;"> </span></span></li> I'm not sure how big an issue that is, though. LeadSongDog come howl! 22:19, 24 June 2010 (UTC)

You might notice that it does the same with the <sup></sup> tags: I spotted "%3Csup%3E18%3C%2Fsup%3E" in there. --Redrose64 (talk) 22:43, 24 June 2010 (UTC)
Yes, but according to HTML_element#Other inline elements, sup tags are valid HTML. We aren't supposed to render invalid HTML.LeadSongDog come howl! 02:00, 25 June 2010 (UTC)
Technically, it's not invalid HTML, because it's in the title parameter and all “dangerous” characters are Urlencoded. It just says that the value of btitle is (E)-4-(2-(6-(2-(2-(2-(<sup>18</sup>F-fluoroethoxy)ethoxy)ethoxy)pyridin-3-yl)vinyl)-N-methyl benzenamine UNIQ577ac5861f92d462-nowiki-00000010-QINU<sup>18</sup>FUNIQ577ac5861f92d462-nowiki-00000011-QINU. If you want to interpret that as HTML (I don't know whether COinS says you should), then it's valid HTML fragment. If you want to interpret it in any way, then it doesn't make much sense, because of the UNIQ-QINU markers, that are used inside the MediaWiki parser, and I'm pretty sure it's a bug that they are present here. Svick (talk) 19:55, 25 June 2010 (UTC)

Minor conflict between {{Cite web}} and {{Citation}}.

During a Wikipedia:Featured list candidates review a very minor conflict between {{Cite web}} and {{Citation}}, was found. When using "accessdate = April 5, 2010", {{Cite web}} created a capital "Retrieved April 5, 2010, while {{Citation}} creates a lower case "retrieved April 5, 2010". I am not 100% sure which is correct, but I believe that {{Cite web}} is the one that needs to be changed. If I am wrong, let me know and I will post this to the {{Citation}} discussion page, as either way, as a matter of consistency, one needs to be changed, but I not being an administrator, I can't do it myself.

For an example: See Ref 4:(used {{Cite web}}) vs Ref 14 (used {{Citation}}) --ARTEST4ECHO talk 12:39, 29 June 2010 (UTC)

You aren't supposed to mix {{Citation}} with the other Cite series of templates— they have deliberately different styles. You will also note that {{Citation}} uses periods as separators and {{Cite web}} uses commas. There used be a note on this somewhere. ---— Gadget850 (Ed) talk 12:53, 29 June 2010 (UTC)
As Gadget850 notes, there is a "conflict" because they are two different styles of citations. I also remember it being noted somewhere that you should not mix {{citation}} with the individual citation templates, i.e. {{cite web}}, {{cite book}}, etc, but can't remember where at the moment. The list should be fixed to use one or the other, whichever is the prevailing style overall. -- AnmaFinotera (talk · contribs) 13:06, 29 June 2010 (UTC)
If a page does bear both styles, sooner or later a bot (usually Citation bot 1 (talk · contribs)) will come along and *try* to decide which is in the majority, and adjust the minority to match. It doesn't always do it correctly (see the complaints page), so it's best to sort it out yourself before the bot notices. --Redrose64 (talk) 16:22, 29 June 2010 (UTC)
Good point. I posted a bug a few weeks ago and it looks like others are seeing similar issues. ---— Gadget850 (Ed) talk 16:31, 29 June 2010 (UTC)

Detailing required parameters better

On my downtime, I like to clean up the AWBC category. I am finding that about 90% of the errors generated stem from a lack of a "title" element in the {{cite web}} template used. Would it be possible to include a <!-- Required parameter --> comment in the examples so that new users copy/pasting the template would better see that if URL or TITLE are not entered, it will result in an error? I see there is a section that lists them under "Required parameters", but these errors could be further prevented by showing it directly in the coding (as other HTML comments are included in examples within template documentations). The above category gets about 75-100 articles added every week from this error. Any chance of this happening to lighten up the cite errors? ~ [ Scott M. Howard ] ~ [ Talk ]:[ Contribs ] ~ 19:24, 26 July 2010 (UTC)

Done, mostly. I added <!-- required --> to the vertical list of parameters alone, since I don't think adding to the lists of horizontal parameters would be very noticeable. Huntster (t @ c) 19:50, 26 July 2010 (UTC)
Thanks, Hunster. That's exactly what I was looking for. =D ~ [ Scott M. Howard ] ~ [ Talk ]:[ Contribs ] ~ 19:51, 26 July 2010 (UTC)

If website is not capitalized on its page, should it be capitalized in the "Work" field?

This is in reference to the following use of the template:

  • Tito, Greg (2005-10-04). "Alpha Centauri". The Escapist. Durham, NC, USA 27713: Themis Group Inc. Retrieved 2010-08-06. 

In text form

{{Cite web | url=http://www.escapistmagazine.com/articles/view/issues/issue_13/86-Alpha-Centauri | title=Alpha Centauri | first=Greg | last=Tito | date=2005-10-04 | work=The Escapist | publisher=Themis Group Inc | location=Durham, NC, USA 27713 | accessdate=2010-08-06 | ref=Tito2005}}

The Escapist (magazine) is type set as "the escapist." Should it be capitalized in the template? Thank you. Vyeh (talk) 15:47, 6 August 2010 (UTC)

The article title follows the rules at Wikipedia:Manual of Style (trademarks) by using title case, thus the citation should follow the same rule. ---— Gadget850 (Ed) talk 15:58, 6 August 2010 (UTC)
Thank you. Vyeh (talk) 16:22, 6 August 2010 (UTC)

Citation confusing with respect to archive

In the Cantenna article, the following citation appears:

^ "PC FAQ's: Build a 'Cantenna' Aerial". Archived from the original on 2008-04-11. 

This wording makes it appear that clicking on "PC FAQ's..." will lead to the archived copy, and clicking on "original" will lead to the original, from which the copy was made. In fact, neither is true. The first link leads to the current version of the website (which is a dead link) and the second leads to the archived copy. Jc3s5h (talk) 17:35, 10 September 2010 (UTC)

Not what I am seeing. Archive first, then dead original. ---— Gadget850 (Ed) talk 18:02, 10 September 2010 (UTC)
Now I see what Gadget850 sees. I'm not sure what I did to get the other result. Never mind. Jc3s5h (talk) 18:24, 10 September 2010 (UTC)

Template:Cite database

User:Droll created this template recently; I think it is a duplicate to this template, inconsistent with the other cite x templates. In any case, the template was not discussed before its creation, so I asked for opinions at Wikipedia talk:Citation templates, but a bit of discussion is at Template talk:Cite database instead. —innotata 16:15, 16 September 2010 (UTC)

"Work" vs "Publisher" parameters

I'm somewhat confused about these two parameters. Publisher, is "the website... hosted by a government service, educational institution, or company. (The publisher is not usually the name of the website, that is usually the work)." Work can be a "book, periodical or website", however the work is automatically italicized. Why on earth would a website that isn't an actual published journal be italicized? To avoid using contradictory apostrophe marks (to cancel out the italics), should a website that is not italicized be placed in the publisher parameter? I'm specifically confused about citing, for example, Blabbermouth.net for album/song articles. The way I understood it was this: Blabbermouth.net was the "larger work", and Roadrunner Records, being the hosting site, was the publisher. So should I de-italicize Blabbermouth.net in the work parameter? Should I remove Roadrunner Records from the publisher parameter altogether? – Kerαunoςcopiagalaxies 03:07, 26 May 2010 (UTC)

Personally, to me Blabbermouth.net is the publisher and there is no work. Work isn't a required field, and should only be used (IMHO) where the site is the online version of a newspaper, magazine, book, etc that would normally be italicized. -- AnmaFinotera (talk · contribs) 03:11, 26 May 2010 (UTC)
The only issue here is that |work= is always italicised, and that is an exceedingly minor issue. It would have been better if such an automatic formatting setup had not been implemented, but its been there far too long to get rid of now. Ultimately, |work= should always be used to describe the website or name of the website, and |publisher= should be used to provide the name of the company that publishes the website. We have other templates to handle news and journal citations, so that isn't an issue. Quite frankly, I never include the publisher as it just clutters the citation line, but including the website name or simplified address in the work field is important for immediate visual identification of where the citation came from. These templates are just useful for maintaining visual cohesion amongst citations, but if you're looking for absolute accuracy in form, I'd recommend writing the citation out manually.
As for the question asked above, I would leave Blabbermouth.net in the |work= field, and left italicised, since that is their stated name. I'd just leave out Roadrunner Records...it just serves no real purpose in the citation. Huntster (t @ c) 05:35, 26 May 2010 (UTC)
Thanks for the replies. I guess I'm making a bigger deal out of this than is necessary, but I'm leaning toward what AnmaFinotera wrote; I'm thinking the publisher parameter should be the standard parameter used, simply because it doesn't italicize. Blabbermouth.net and most websites shouldn't, in general, be italicized, simply because they're not actual publications. If this is the case, then the template instructions need to be changed to clarify this, only because of the italics issue. Websites should be placed in the publisher parameter, unless it is the website of a published journal, which should then be placed in the work parameter. – Kerαunoςcopiagalaxies 08:43, 26 May 2010 (UTC)
Wouldn't that be counterproductive? When I see "publisher", I never think of the website itself. The term publisher, by its very definition, means "one who publishes", usually meaning the company that is doing the publishing, as I wrote above. It would be very misleading to tell people to use publisher in any other circumstance. I thought there was talk about adding a "website" (or similar) parameter, which could be designated as the proper place to put the website name, but I can't find it now. Huntster (t @ c) 10:07, 26 May 2010 (UTC)
I suppose it could be misleading; I didn't realize there was an important distinction between the two. I was boiling work/publisher down to italicized/non-italicized. I could keep the website in the work parameter, but I would add apostrophes to cancel out the italics, then, unless it was USA Today or Flying or similar. Would that work? Btw, a website parameter would be very nice. – Kerαunoςcopiagalaxies 20:03, 26 May 2010 (UTC)
See archive 6, where on 20 November 2009 I illustrated the distinction between |work= and |publisher= by means of a hypothetical citation of Wikipedia itself (which is of course forbidden by WP:CIRCULAR). --Redrose64 (talk) 20:50, 26 May 2010 (UTC)
I was going to use Wikipedia as an example in my response above, actually, but decided against it. My example would have looked like yours: I understand the work/publisher intricacy perfectly; what I'm questioning is the italicizing and what I'm implementing is laziness. "Wikipedia", if placed in the work parameter, should not be italicized. Therefore, I would most likely throw it in the publisher parameter out of sheer laziness. But to be technically correct, I would have to use the template like so:
  • {{cite web |url=http://en.wikipedia.org/wiki/Han_Dynasty |title=Han Dynasty|work=Wikipedia|publisher=Wikimedia Foundation, Inc}}
Note that the italicized "Wikipedia" would return as a non-italicized "Wikipedia" in the references section. So I guess what I'm asking is, should I be canceling out the italics in the work section, instead of moving the "work" into the publisher parameter? Or is {{cite web}} condoning the italicizing of all websites? – Kerαunoςcopiagalaxies 21:48, 26 May 2010 (UTC)
Unfortunately, it is condoning the italicising, simply because it is doing it automatically. I see nothing wrong with cancelling out the italics, so long as the work and publisher are in their technically correct locations. Huntster (t @ c) 19:52, 27 May 2010 (UTC)
Great, canceling the italics it is, then (for non-published works). I'll be more conscious of the work/publisher fields and I'll correct the templates I've added in the past as I come across them. Thanks for your responses. – Kerαunoςcopiagalaxies 20:27, 27 May 2010 (UTC)

Evidently, SmackBot reverts the anti-italicizing marks; this is somewhat frustrating. Websites simply should not be italicized unless they are the websites of a published journal. – Kerαunoςcopiagalaxies 19:41, 8 June 2010 (UTC)

Another reason I said just use publisher. This is a more accurate usage, IMHO, anyway (as already noted above :-) ). -- AnmaFinotera (talk · contribs) 19:46, 8 June 2010 (UTC)
I think AWB is smart enough to know that the parameter has italics added so removes the "redundant" italics. Cancelling is perhaps dubious, as either the field should be italicised by default or it shouldn't. italicising Wikipdedia would seem fine, as Britannica, but not "Google" or "next.com". If necessary we should add a website parameter, keep things as simple as possible, but no simpler! Rich Farmbrough, 21:36, 8 June 2010 (UTC).
I agree, I would love to have a website parameter. It would clear up a lot of confusion this work= thing is causing. – Kerαunoςcopiagalaxies 00:34, 11 June 2010 (UTC)
Agree as well, |website= would be fantastic, but I'm not good enough with the citation meta-syntax to add it. Huntster (t @ c) 02:54, 11 June 2010 (UTC)
Do we just throw up the {{editprotected}} template? Otherwise, we could be waiting forever for it to be added. I also wonder if it could be included in the refTools "cite" form. Not sure where the headquarters for that option is. – Kerαunoςcopiagalaxies 04:18, 11 June 2010 (UTC)
No, editprotected is only used when the "fix" has already been developed and just needs to be deployed. Just need to find a competent editor who can write the appropriate bit. You might try asking at Template talk:Citation or Template talk:Citation/core, since it will likely involve a modification of some kind to the Citation/core meta template (oh, how I hate thee, meta templates). Huntster (t @ c) 04:33, 11 June 2010 (UTC)
Okay: Template_talk:Citation#Website_parameter. – Kerαunoςcopiagalaxies 06:31, 11 June 2010 (UTC)
Actually, the alias would be added here. Do I understand that you want website as an alias for publisher? ---— Gadget850 (Ed) talk 09:54, 11 June 2010 (UTC)
Well, it would only be similar to "publisher" in that it would not use italics. I don't know how the back-end is coded, but it would be most similar to "work", just, again, with no italics. Either "work" or "website" should be allowed in a citation, but not both, so I suppose a determination must be made as to which would take precedence in the case both are filled in (for some bizarre reason). Someone may want to use both "website" and "publisher" together (since the two are totally separate entities, in terms of the website being the thing that is published by the publisher), so I'd not make "website" an alias of "publisher". Huntster (t @ c) 10:02, 11 June 2010 (UTC)
Agreed with Huntster. The publisher= parameter would be used in conjunction with either work= or website=. Website= would supersede work=. But where work= automatically italicizes, website= would not. – Kerαunoςcopiagalaxies 10:07, 11 June 2010 (UTC)

Work vs publisher arbitrary break

Any hope of a website= parameter being added any time soon? Or should there be more of a discussion on it? – Kerαunoςcopiagalaxies 06:41, 17 June 2010 (UTC)

I am confused on what is desired. The website parameter was added to the sandbox back in November 2009:

{{cite web/sandbox |url=http://www.example.org/ |title=My Favorite Things, Part II |author=Doe, John |publisher=Open Publishing |date=30 April 2005 |website=Encyclopedia of Things |accessdate=6 July 2005 }}

Doe, John (30 April 2005). "My Favorite Things, Part II". Open Publishing. http://www.example.org/. Retrieved 6 July 2005. 

However, see discussions at Template talk:Cite web/Archive 6. ---— Gadget850 (Ed) talk 21:38, 18 June 2010 (UTC)

I'm confused by the archive; discussion sort of stopped. To me, it looks like there was consensus for the website= parameter, but I don't understand why it was never implemented into the actual template. But in your example above, the website= parameter is italicizing, which defeats its whole purpose. Work= automatically italicizes. Website= should not.
{{cite web/sandbox |url=http://www.example.org/ |title=My Favorite Things, Part II |author=Doe, John |publisher=Open Publishing |date=30 April 2005 |website=Encyclopedia of Things |accessdate=6 July 2005 }}
should look like:
Doe, John (30 April 2005). "My Favorite Things, Part II". Encyclopedia of Things. Open Publishing. Retrieved 6 July 2005.
Kerαunoςcopiagalaxies 23:01, 18 June 2010 (UTC)
To Gadget850, you didn't read the above discussion. The sandbox version simply makes "website" an alias of "work", whereas what is desired is for "website" to be an entirely different parameter, without italics, where either "work" or "website" could be used in a citation, but not both. Huntster (t @ c) 23:52, 18 June 2010 (UTC)
I didn't do the sandbox version. Per the APA guide[30] and the MHRA guide[31] p. 61 it looks like the website name would be italicized. ---— Gadget850 (Ed) talk 12:48, 19 June 2010 (UTC)
I was not able to retrieve the MHRA guide, but the APA guide shows on page 203 that the name of the publication is in italics, and the website is presented as a visible URL in roman type. It seems evident that in cases like example 21, the name of the website, Google Books, is different from the name of the publication, The standard edition of the complete of Sigmund Freud, and the name of the website is not stated except as part of the URL. Jc3s5h (talk) 15:19, 19 June 2010 (UTC)
Ah, I see. Maybe this is more complex than I originally thought. The MLA guidelines for electronic sources show, for example, a page on a Website should be cited like so: "How to Make Vegetarian Chili." eHow.com. eHow, n.d. Web. 24 Feb. 2009. (Note, the "n.d." means publishing date information was not available.) In which case, the website (eHow.com) is italicized, the publisher (eHow) is not. (MLA specifies the publisher as the "name of institution/organization affiliated with the site." So... maybe websites should be italicized after all? If this is the case, then should the website= alias be incorporated into the template? And I would love to have some sort of consensus reached here so this discussion can be used to support any changes made (it turns out, the un-italicizing of website names has been catching on on a few pages, and now this may need to be reversed.) (Aside: Huntster, thanks for the arbitrary break; even though Wikipedia:Talk_page_guidelines#Technical_and_format_standards doesn't mention to make the break specific to the discussion, I thought I would change it to help edit summaries on editors' watchlists stay on target.) – Kerαunoςcopiagalaxies 23:50, 19 June 2010 (UTC)
As best I see, this template is a hybrid of APA and MHRA with some variants. None of the style guides seem to cover online publications such as Wikipedia. The citation templates do not show the URL when viewing, but do when printing or print previewing; instead, they are linked in the title. It does make sense to me that the website name would generally be treated as the work. A few examples:
I think a case could be made that where the work and publisher are the same, then the publisher field could be dropped. I believe the same rule has been applied similarly to location where the work includes the location, such as New York Times.
This discussion really covers all of the templates. Whatever is decided should be added to Wikipedia:Citing sources.
---— Gadget850 (Ed) talk 20:10, 20 June 2010 (UTC)

Work vs publisher (arbitrary July break)

I come here after seeing Rjwilmsi (talk · contribs) make use of his/her tool CiteCompletion on an article. That tool institutionalizes a convention for work/publisher that has existed for at least a year — since this documentation change to cite web, if not longer.

My work/publisher issue is different than the bottom-up implementation issue about italics that started this long thread; that one an artifact of the dependency of {{cite web}} on {{Citation/core}} (and apparently, the core's dependency on a particular style guide). My issue is a top-down one, a requirement I think {{cite web}} should meet, even if requires {{Citation/core}} to change. Take the following two cases:

Items 1 and 2 are obviously the url and the title. No issue with either. But what about items 3 and 4? In spite of the current documentation, I would use work for item 3 and publisher for item 4. In the case of a website, I think the publisher is either

Anyone have a better approach? Thanks, (talk) 21:40, 19 July 2010 (UTC).

My understanding is that #3 (in your examples) wouldn't be used at all. #4 is the work, and you wouldn't use a publisher. Hollywood Reporter didn't publish what you read on Hibberd's site—I mean this "literally". The URL, therefore, must change. What you read was a tertiary source (not unlike Wikipedia itself) repeating news from the secondary source—in this case, New York Times and Hollywood Reporter. (Btw, just a clarification, the italicizing is really no longer an issue; it was determined above that Websites are, in fact, italicized.) – Kerαunoςcopiagalaxies 02:22, 20 July 2010 (UTC)
Thanks for the reply. In the case of Hibberd's live feed, you used to be correct...his blog was at thrfeed.com or something similar. But if you try the old URL now, it redirects to the subdomain of hollywoodreporter.com as listed above, so the two examples are the same situation. As to your secondary/tertiary distinction, it sounds like you're saying that the NYT's website is, for reference purposes, to be considered distinct from the dead-tree version of the NYT. But if the template is called {{cite web}}, I would think its clear that the source being cited is what your called the "tertiary" work; otherwise, editors who find sources online can only cite online sources for which they are also able to verify the dead-tree version of the source.
I still think that for websites, blogs with specific identities and often multiple contributors posting warrant identification, and using work is a natural fit. The analogy is to columnists in a newspaper, or podcasters associated with an organization such as BBC or NPR: both the column/podcast and the newspaper or network have names worth identifying. Thanks. (talk) 21:03, 22 July 2010 (UTC)
Well, I didn't mean the dead-tree version as a source; journals usually print additional or different information on their websites anyway. Somewhere on the citation guidelines, it says that editors should try to use the original source. If an online journal prints a Reuters article, I will usually search Reuters for that original article and cite that instead. It's only a bit more effort, and definitely not necessary or required by any means. But it does help skirt any possible "reliable source" issues, especially if you would otherwise be citing a blog (for example). Anyway, the {{cite web}} template should be filled out as correctly as possible to the best of your judgment. But the publisher parameter is definitely not for the "original source" website. It is meant simply for the actual publisher company name, most likely a limited liability company or similar. Glad I could help! – Kerαunoςcopiagalaxies 04:40, 23 July 2010 (UTC)
If web sites should be italicized (e.g. amazon.com) should MOS:ITALICS be updated to reflect this? GoingBatty (talk) 15:02, 7 August 2010 (UTC)

Im sorry but I think everyone seems to be missing the point here. If everyone accepts that websites in references should appear in Italics then websites in the main prose should also appear in Italics which would be in complete odds and conflict with the rest of MOS:TEXT. This is not a clear cut thing and we should stop trying to standardise. --Lil-unique1 (talk) 16:23, 11 August 2010 (UTC)

Frankly I find it poor taste that questions about what a published 'work' is, have led to a series discussion about the changing MOS rather than an intelligent discussion about why the template needs to auto-format or actually looking in depth at what 'work' is. -- Lil_℧niquℇ №1 | talk2me 22:37, 21 September 2010 (UTC)

Parameter spelling

This template spells the "Ref" parameter with a lower-case "r", whereas {{Citation/core}} and {{Harvard citation no brackets}} spell them with a capital "R". I suggest the template be changed so that they all use the same spelling. SharkD  Talk  22:17, 25 September 2010 (UTC)

{{Citation/core}} deliberately uses many parameters with an initial capital. It is a meta-template and not designed to be used directly. Templates that use {{Citation/core}} can use the same names, but lower case, making them different parameters. ---— Gadget850 (Ed) talk 05:49, 26 September 2010 (UTC)


Was wondering if it would not be a good idea to have a parameter for "Subscription required".A s this is generally a required part of information in an FA/GA review when a link needs this to access the information. Currently there is no standardized format of this info. So we see things like "Subscription required" or "Requires Subscription" etc.. Moxy (talk) 15:56, 26 September 2010 (UTC)

See {{subscription required}} and {{registration required}} ---— Gadget850 (Ed) talk 17:22, 26 September 2010 (UTC)
yes i understand this are here...but are not the only way used to say this info and plus its just another templet that needs to load. Most of the time during a GA review i see |format=subscription required and not the templates....So does it not make sense to consolidate this and have a standard that is in one template and save the load time.Moxy (talk) 17:45, 26 September 2010 (UTC)
Well, ideally, if the subscription is required, the site should not be in the link. Putting it in amounts to an endorsement of the pay site. WP should make a ruling clarifying that pay sites, even academic ones, are not allowed. But, if WP is going to allow such links, it does seem to me also the reader should be tipped off by a parameter in the specification.Dave (talk) 20:22, 26 September 2010 (UTC)
How is linking to paywalled pages different to using books as references? You have to pay for them too. See also WP:PAYWALL. Svick (talk) 21:55, 26 September 2010 (UTC)
Well, the books are a red herring. Libraries are free for most people. But linking to a paywalled page isn't an endorsement of the distributor, it's just a link. Readers make their own choices of whether or not to pay for access, and many already have institutional subscription access. Wikiproject Resource Exchange can help those that don't. See the lengthy discussions at Wikipedia_talk:Verifiability#WP:PAYWALL and freely accessible sources. LeadSongDog come howl! 03:41, 27 September 2010 (UTC)

annotation field

There should be an annotation field. -- Tomdo08 (talk) 17:19, 29 September 2010 (UTC)

See Template talk:Cite journal#Annotation field. ---— Gadget850 (Ed) talk 05:44, 30 September 2010 (UTC)


I know pdfs automatically generate the PDF symbol e.g.

I was wondering if it is possible to code the format parameter to perform a similar function for XLS, DOC, etc. I have tried coded this manually before e.g.

This gets very close except the URL link diagonal arrow is present.

  • Currently there isn't a work around within cite web
Putting everything in |format=, for example, {{cite web|url=http://www.tvcriticsassociation.com/tca/files/July2007/TCA%20complete%20NOMINEES%20list.xls|title=Complete list of nominees|format={{XLSlink}} [[Microsoft Excel|XLS]]|work=[[Television Critics Association]]|accessdate=July 17, 2008}} gives a bracketed icon
Trying to append it to |title=, for example, {{cite web|url=http://www.tvcriticsassociation.com/tca/files/July2007/TCA%20complete%20NOMINEES%20list.xls|title=Complete list of nominees|format={{XLSlink}} [[Microsoft Excel|XLS]]|work=[[Television Critics Association]]|accessdate=July 17, 2008}} gives a stray link sign and end of quote

Any suggestions? Rambo's Revenge (talk) 12:02, 30 September 2010 (UTC)

The PDF icon is not produced by setting |format=PDF, nor is it generated by this template (or the underlying {{citation/core}}). It's generated at low level, possibly as deep as the MediaWiki software itself, and occurs whenever a URL ending ".pdf" is encountered. This can be demonstrated by using a URL such as http://www.example.pdf outside all templates. --Redrose64 (talk) 12:21, 30 September 2010 (UTC)
Maybe it's just me thinking outloud but my point is it is a very informative way of making a user aware that they are about to access a page that might take some time to download. I think maybe we should extend this to other xls, doc formats. I'll try and see if it can be done at source but, if not, should we not have some sort of workaround? Rambo's Revenge (talk) 13:25, 30 September 2010 (UTC)
Active content such as XLS and DOC files is commonly regarded as hazardous, and so is blocked by many firewalls. Some users also object to the promotion of proprietary closed formats. Offering the same content rendered as a basic PDF would be a helpful alternative for some of us. Perhaps a discussion at wp:VPT would be warranted.LeadSongDog come howl! 14:51, 30 September 2010 (UTC)

Taken to Wikipedia:Village pump (technical)#External link icons proposal. Rambo's Revenge (talk) 15:20, 30 September 2010 (UTC)

Replied there. ---— Gadget850 (Ed) talk 16:14, 30 September 2010 (UTC)

Cite web and ref names

I need help with using cite web template in named references. If I use cite web template and named references more than once, then result is only one quotation listed in reflist section. Is there a way to use cite web template and to have both named references and listed quotations in reference subtitle?--Antidiskriminator (talk) 07:48, 22 October 2010 (UTC)

If the two citations have different |quote= values then they're different citations, so use different reference names. Rjwilmsi 08:05, 22 October 2010 (UTC)

Should I use different reference names if it is the same web site and same web page that I quote?--Antidiskriminator (talk) 08:29, 22 October 2010 (UTC)

Ah, maybe I slightly misunderstood your first question. If you are citing a website to source fact 1, and then later fact 2, then use the same reference name (it's the same citation even if you are refering to different information). If you are citing a website but using different parameters like |page= or |quote=, the use different reference names (the citations are defined as different). Rjwilmsi 08:53, 22 October 2010 (UTC)

Thanks. --Antidiskriminator (talk) 08:55, 22 October 2010 (UTC)

Assuming that your problem page is User:Antidiskriminator/Drafts of articles/Nationalization of history, I see that there are two refs named "Laboratory", identical apart from the quote; and five named "Hopkins", again, identical apart from the quote. The problem does not lie with the {{cite web}} template, but with how you've used the <ref></ref> construct. If the ref is unnamed, it's used as it stands. But if it's named, the reference processing checks through a list to see if the ref name has previously been used. If it hasn't, the ref is used as it stands, just as if it was unnamed; and then the name is noted in the list. If a ref name is used a second time, the whole of the content of the reuse of that ref is discarded and the ref is linked back to the first use of the name. So, if the following is encountered:
First statement.<ref name="MyName">ref content version 1</ref>
Second statement.<ref name="MyName">ref content version 2</ref>
it is exactly the same as putting this:
First statement.<ref name="MyName">ref content version 1</ref>
Second statement.<ref name="MyName" />
I notice that in each case the quote given is quite lengthy: is it necessary to have all this? I'm thinking mainly of copyright issues. If you can eliminate the quotes entirely, then you can condense the refs like this:
First statement.<ref name="Hopkins">{{Cite web
| url = http://www.scribd.com/doc/2557562/History-of-Globalization
| title = The History of Globalization - and the Globalization of History
| accessdate = October 20, 2010
| author = A. G. Hopkins
| authorlink = A. G. Hopkins
| language = english
Second statement.<ref name="Hopkins" />
Third statement.<ref name="Hopkins" />
If you do need to show the different quotes, you must first remove all the names from the <ref> tags. It would be a good idea to add page numbers too; the Hopkins doc has those, in the bar at the bottom, and they would go in the |page= parameter. You can then leave your article as it now stands; or you can reduce the space a bit like this:
First statement.<ref>{{Cite web
| url = http://www.scribd.com/doc/2557562/History-of-Globalization
| title = The History of Globalization - and the Globalization of History
| accessdate = October 20, 2010
| author = A. G. Hopkins
| authorlink = A. G. Hopkins
| language = english
| page = 4
| quote = Short quote relevant to the preceding statement }}</ref>
Second statement.<ref>Hopkins, p. 6 "Another short quote".</ref>
Third statement.<ref>Hopkins, p. 11 "A third short quote".</ref>
Note that for the second and subsequent cases, {{cite web}} is not used. --Redrose64 (talk) 13:51, 22 October 2010 (UTC)

Thank you very much for detailed explanation. I prefer to have short quotes written in the article if possible, and therefore, I would choose last case that you proposed. --Antidiskriminator (talk) 14:21, 22 October 2010 (UTC)

Large (e.g. 5MB+) web pages as citations

While trying to find a source for the birth year of Professor Boyd Hilton, I found a web page:


that is over 5 MB in length, and minimally formatted for viewing. I was using a old computer at the time, and a relatively slow Internet connection, so when I clicked on it, it took quite a while for my browser to download and render the page. I am adding that link as a reference for the professor's age (I was unable to find a smaller web page of similar authority) and think it would be handy to have a way to tell people that at the time I accessed the page it was 5.9MB in length. {{Cite web}} has no specific parameter for that purpose, and Category:Inline templates has no obvious option either. Any comments or suggestions? Thanks. (talk) 13:28, 1 November 2010 (UTC)

One can always add it as <ref>{{cite web|blahblahblah}} (5.9MB file)</ref>. However, in this case you might prefer to use this source which is much smaller and just as authoritative, if not more.LeadSongDog come howl! 20:43, 1 November 2010 (UTC)
Good find. Thanks. (talk) 11:45, 2 November 2010 (UTC)

Template help

If the web address I use includes [ or ], how can I avoid making a mess of the template? - Theornamentalist (talk) 22:11, 14 December 2010 (UTC)

Common question, detailed answers are in the archives; but in brief, use %5B for [ and %5D for ]. See Percent-encoding for encodings of other characters. --Redrose64 (talk) 22:52, 14 December 2010 (UTC)
Thank you. - Theornamentalist (talk) 22:53, 14 December 2010 (UTC)

propose "editor" field

I have come across a couple of references where "editor" would be appropriate. It seems that such cases would be fairly common on the web. I think that adding an editor field to this template would be useful. Bcharles (talk) 21:43, 19 December 2010 (UTC)

Examples? The template supports the same editor parameters as {{cite book}}. This is undocumented, as it would probably be more appropriate to use {{cite news}} or {{cite journal}}. ---— Gadget850 (Ed) talk 22:23, 19 December 2010 (UTC)
There are a total of twelve different fields available for up to four editors; all but three have synonyms.
  • |editor-last= (synonyms |editor-surname=, |editor1-last=, |editor1-surname=, |editor1=, |editor=, |editors=) is for the surname of the first or only editor
  • |editor-first= (synonyms |editor-given=, |editor1-first=, |editor1-given=) is for the first name(s) of the first or only editor
  • |editor-link= (synonym |editor1-link=) is for the name of the Wikipedia article carrying the biography of the first or only editor
  • |editor2-last= (synonyms |editor2-surname=, |editor2=) is for the surname of the second editor
  • |editor2-first= (synonym |editor2-given=) is for the first name(s) of the second editor
  • |editor2-link= is for the name of the Wikipedia article carrying the biography of the second editor
For the third and fourth editors, use fields as per the second editor but change "2" to "3" or "4" as appropriate, ie |editor3-last= etc. --Redrose64 (talk) 22:45, 19 December 2010 (UTC)
Yes, but why do you need editor for a website? If it is a website for a journal, magazine or the like, then other templates should be used. Odds are very high that a follow on editor will update the template. ---— Gadget850 (Ed) talk 02:47, 20 December 2010 (UTC)
Many websites do not show authors of individual pages. Some have a note of the form "This site maintained by John Doe", either at the foot of a page, on the home page, or a "contacts" or "about us" page. It's unreasonable to expect that Doe wrote every single page, so he should not be credited as author. The editor fields are ideal for this purpose.
If the website is neither book nor journal, {{cite book}}/{{cite journal}} should not be used. --Redrose64 (talk) 16:19, 20 December 2010 (UTC)


Any clue why ref #13 in Hugo Boss (Jewish Virtual Library) and ref #2 in Hassan Musa (Sudplanete) do not give the expected result? Thanks, Racconish Tk 08:08, 3 January 2011 (UTC)

For both articles, the problem was a line break in the middle of the citation. See my diffs: [32] [33] to see where the problems were. Huntster (t @ c) 08:40, 3 January 2011 (UTC)
Thanks for the quick - and useful - answer. Racconish Tk 08:48, 3 January 2011 (UTC)

Ref tags

Can someone put the ref tags in the template descriptions? Or is there some reason that they arent allowed... It would make it a lot easier to just copy the whole thing to where its needed.--Metallurgist (talk) 04:55, 14 December 2010 (UTC)

The {{cite web}} template, and the related templates such as {{cite book}}, do not necessarily occur within <ref></ref> tags; they are also used in bulleted lists, particularly when shortened footnotes are used. Shortened footnotes need not be limited to refs in books, but can also be used for web references: see, for example, the ref to "Wilson 1945" here. Incidentally, although that article has 17 entries under "Notes", only one of these was produced using <ref></ref> tags - all the others are produced by {{sfn}}, which uses {{#tag:ref }} internally, instead of <ref></ref>. --Redrose64 (talk) 16:33, 14 December 2010 (UTC)
<ref></ref> and {{#tag:ref }} access the same underlying code, the only difference is that they are processed at different points in the parser. Happymelon 11:02, 18 January 2011 (UTC)

Titles containing square brackets (such as assumed titles ie [Introduction])

Cite web fails to handle titles containing square brackets correctly __"[Introduction=>__"]. displays where => stands for the linking icon and __ for the extent of underlining. This leaves the close quote incorrectly placed (prior to the square bracket). The underlining and link icon is also incorrectly placed. ie: "[Example]".  Fifelfoo (talk) 01:18, 18 January 2011 (UTC)

For how to handle this, see this answer and this edit. Debresser (talk) 07:36, 18 January 2011 (UTC)
Its a work around, not a solution, and it breaks a fundamental wikimarkup convention that single instances of special characters don't need escaping. Thanks for pointing to it, but the template maintainers need to add this to the bug list. Fifelfoo (talk) 07:42, 18 January 2011 (UTC)
Agreed. Debresser (talk) 07:59, 18 January 2011 (UTC)
There is no "funamental wikimarkup convention" of that nature; this has nothing to do with {{cite web}}, it's a feature of the MediaWiki parser. The link [408113262=1&ids[407943528]=1] (a RevDel link for a couple of main page revisions) has exactly the same issue, and the solution is the same: the first right square bracket encountered is treated as the complement to the initial bracket and closes the link, so to avoid that happening, you need to escape any brackets you don't want to be thus considered. Happymelon 10:49, 18 January 2011 (UTC)

If you take Fifelfoo's original example, and subst: it right out, then eliminate non-display stuff (like the Coins metadata), you find that the last template to be fed title and URL separately is {{Citation/make link}}, which is given the following:

{{Citation/make link
| 1=http://www.google.com/
| 2="[Example]"

which turns that into

[http://www.google.com/ "[Example]"]

which displays as


so as Happy-melon states, it's the MediaWiki parser - no simple template modification will change this behaviour.

Questions like this have come up several times before, not just here but on the talk pages of {{citation}}, {{cite journal}}, etc. so I think we need a stock answer in the form of an FAQ list. We have another problem: the best advice to offer. The problem of square brackets has several solutionsworkarounds, and not all are universal for both URLs and titles - some work in one but not the other. For instance, Debresser provides two links, but these give two different workarounds, the first of which (percent-encoding) is a well-known guaranteed fix for URLs, but doesn't work at all in titles:

{{cite web |title=%5bExample%5d |url=http://www.google.com/}}

His second suggestion (numeric character encoding) works in both URLs and titles:

{{cite web |title=&#91;Example&#93; |url=http://www.google.com/}}

Finding out/remembering these character codes can be tiresome, and they are not intuitive to those who may encounter them later on. There is another way (which works in titles but not in URLs) - type the square brackets, but hide the right-hand one inside <nowiki></nowiki> tags:

{{cite web |title=[Example<nowiki>]</nowiki> |url=http://www.google.com/}}

--Redrose64 (talk) 13:06, 18 January 2011 (UTC)

One thing we can do: update Help:URL. Move this issue to the top of the page and add a list of encodings, ---— Gadget850 (Ed) talk 16:12, 25 January 2011 (UTC)


The doc page has recently been altered to put spaces after the pipes and before the equal signs in the copy&paste examples. That is, instead of this:

All parameters, horizontal layout
{{cite web |url= |title= |first= |last= |author= |authorlink= |coauthors= |date= |month= |year= |work= |publisher= |location= |page= |pages= |at= |language= |trans_title= |format= |doi= |archiveurl= |archivedate= |accessdate= |quote= |ref= |separator= |postscript=}}

we now have this:

All parameters, horizontal layout
{{cite web | url = | title = | first = | last = | author = | authorlink = |coauthors = | date = | month = | year = | work = | publisher = | location = | page = | pages = | at = | language = | trans_title = | format = | doi = | archiveurl = | archivedate = | accessdate = | quote = | ref = | separator = | postscript = }}

Is this something that we want? I, for one, do not like the change at all. --Redrose64 (talk) 15:47, 25 January 2011 (UTC)

Don't like it. I use an external script to cleanup citations and removing extra spaces like this is one of the features. I know of other who do similar cleanup. Yes to a space before the pipe (it especially helps with long urls), but no to the rest. ---— Gadget850 (Ed) talk 16:10, 25 January 2011 (UTC)
I agree. Why is there a need for spaces? Albacore (talk) 17:34, 29 January 2011 (UTC)
I checked the other template documentation— none use whitespace in this manner. Reverted. ---— Gadget850 (Ed) talk 20:08, 29 January 2011 (UTC)
Face-smile.svg Thank you --Redrose64 (talk) 20:22, 29 January 2011 (UTC)


There's an error in the Required parameters section. Where it says, The character "[" must be replaced with "&#91", there should be a colon after the "1". Thank you. LordVetinari (talk) 12:16, 6 February 2011 (UTC)

Yes check.svg Done Actually, a semicolon, not a colon. It's on the documentation page, which is not protected, so I've fixed it. --Redrose64 (talk) 15:01, 6 February 2011 (UTC)

Citing websites based on Flash

Example: a musician has some informative quotes about an album within two subsections of his website; however, there are no direct links to those subsections, since there is no HTML. How should this be represented within a citation? Just the main website address followed by a quote, perhaps..? Mac Dreamstate (talk) 00:24, 7 February 2011 (UTC)

Please provide an example. ---— Gadget850 (Ed) talk 03:20, 7 February 2011 (UTC)
On the official website for guitarist Vinnie Moore, there is a section named 'Music', which leads to subsections for individual albums. In this case, we'll use the Mind's Eye album, which contains a Flash/Java-based link saying "Click here read more about this release". Here, I feel some of the information can be cited within the Mind's Eye Wikipedia article for general reference purposes. However, I'm unsure as to how to cite that particular section of the website, since there is no HTML link to it. Is there even a way to link to something Flash-based? Mac Dreamstate (talk) 03:42, 7 February 2011 (UTC)

Format issue when no "year=" parameter is set.

This template produces a formatting issue when no "year=" parameter has been set. If the first name of the author is an Initial ending with a period, then a double period is produced. See the following example:

This is not an issue with most other cite templates since for most reference a year or date is required. However, websites often do not report the date they were created and therefore often do not have a year parameter.TR 15:48, 15 February 2011 (UTC)

Either do not use the period after the J or add |separator= with no parameter. ---— Gadget850 (Ed) talk 16:00, 15 February 2011 (UTC)

Multiple quotes

Since there is no explanation in the Template article about handling cases where an editor wishes to quote more than one passage using the quote= option, I'm guessing the distribution of the quotes needs to be handled manually? I.e., the editor should insert quotation marks internally but not to surround the whole batch of quotes. Is this so?—Biosketch (talk) 11:01, 18 February 2011 (UTC)

There shouldn't really be a need to quote very much... after all, the use of {{cite web}} suggests that the source is online, so anybody wishing to verify the ref should be able to follow the link and verify the fact on the external site. The documentation for {{cite book}} gives a rather longer description for |quote=, including the crucial note "Should not be excessive in length: More than a few sentences is rarely needed ...". If, as you describe, you wish to quote more than one passage, this suggests that the quoted material is "excessive in length". --Redrose64 (talk) 14:12, 18 February 2011 (UTC)
Agree. Most quotes should be in-text and supported by the citation. Quotes within a citation should support the text in some manner and should be succinct and relevant. -— Gadget850 (Ed) talk 14:27, 18 February 2011 (UTC)

Can we drop all the periods (full stops)?

Currently when Cite web displays it has periods (full stops for my English speakers on the other side of the pond) after every parameter which frequently makes the reference string look clunky. I would like to propose that we remove some of the periods but doing that may require shifting the order of things or adding different punctuation. Does anyone have an opinion on that? If we do the following I think it would flow better and be easier to read and understand.

  1. Remove the period after |author=
  2. Switch the order of display for |work= and |url=/|title= with a : as a separator. Using the example from this templates documentation it would look like Encyclopedia of Things: "My Favorite Things, Part II"
  3. Replace the . preceding |publisher= with published by. --Kumioko (talk) 19:28, 24 October 2010 (UTC)
{{cite web}} uses {{citation/core}}, so this proposal would affect several other templates such as {{cite book}}. --Redrose64 (talk) 19:55, 24 October 2010 (UTC)
Oh well that makes sense I guess. Not sure that it would be a negative affect on the others either but thats good to know. --Kumioko (talk) 20:03, 24 October 2010 (UTC)
  • This proposal would have a series of negative effects on other wikipedia users. Most citations styles, and the citation style that Wikipedia's in house style is very loosely based on, use periods and commas as basic field delimiters; much as they are used in common language. A significant majority of editors and users have either been trained to recognise these field delineation systems, or ought to be as part of their ongoing studies or autodidactic work. Changing the in house style to strip field delimiters would move the in house style further away from its origin style, and from the standard styles in common use. The central fields requiring delimiting are Author block. Year block. Title block. Editor & other contributor block. Containing works block (including volumes). Publisher block. Page reference block. Identifiers block. Within these blocks, delimiters are often useful to differentiate between "Bobs, Mary" and "Bobs, Susan" for example. Fifelfoo (talk) 02:01, 28 February 2011 (UTC)

proposal: cable parameter

Cables from the United States diplomatic cables leak are being widely used by mainstream media, academics, politicians, international governmental organisations, and alternative media as a source of knowledge about the World, i.e. encyclopedic knowledge. The more notable usages of the first 1% of these cables are now being quite extensively referenced here in the English-language Wikipedia, both indirectly to get NOR opinions on what is notable, and directly since they are secondary/tertiary sources about sociopolitical events (or primary sources of e.g. ambassadors' opinions). For comments about the citation of these sources, please see Wikipedia:Requests_for_comment/Use_of_classified_documents. i'm not trying to redo that debate here, and i'm not claiming to make an NPOV summary of the comments on that page!

IMHO it's now obvious that we should think of adding a parameter for easier meta-analysis. The cables all have very standard ID's, e.g. 08ASHGABAT1399 for http://www.wikileaks.ch/cable/2008/10/08ASHGABAT1399.html, of which the first and second parts are semantically obvious (year and city name), and the full ID is written in the "original" (WikiLeaks) version, or can be guessed from e.g. Guardian-published versions by removing 0's in front of the number like 001399.

Proposal: Add a parameter, e.g. "cable ", so the above example would require adding |cable = 08ASHGABAT1399 to the cite web citation template (and also the general Template:citation?). Various possibilities for the parameter name include (please add short arguments for/against within the list, longer discussion below):

  • cable
    • for: short Boud (talk) 02:27, 1 January 2011 (UTC)
    • against: ambiguous? US diplomatic cables have been published prior to Cablegate, and probably will be post-Cablegate, and other countries' embassies also send diplomatic cables, etc. Boud (talk) 02:27, 1 January 2011 (UTC)
  • cablegate
    • for: unambiguous, official name by WikiLeaks Boud (talk) 02:27, 1 January 2011 (UTC)
    • against: POV, because doesn't include USG (US government) POV? original authorship is not by WikiLeaks, and the authors may not see their cables as being similar to Watergate Boud (talk) 02:27, 1 January 2011 (UTC)
  • cable_id
    • against: all the template parameters are ID of some sort Boud (talk) 02:27, 1 January 2011 (UTC)
  • cablegate_id
  • cable_ref_id
  • cablegate_ref_id

i have almost no experience in editing templates, so someone else will have to implement this if there's consensus. Boud (talk) 02:27, 1 January 2011 (UTC)

These are documents, regardless of the host, so {{cite document}} which was merged to {{cite journal}} would probably be more appropriate than cite web. Regardless, the templates in question have an |id= parameter for just this purpose. ---— Gadget850 (Ed) talk 02:38, 1 January 2011 (UTC)
Nice point about cite document, and you're right that the ID parameter is included for cite document or the generic citation. (i had tried it for web cite, but it didn't print any info visible to the reader.) i've added Template:Cablegate so that people can use {{cite document | ... |id={{cablegate|01CITYNAME1234}} | ... }}. Since it's a template, people can discuss what should be the "obvious" URI over at that template or a cablegate-related page and then update the template. Thanks for the fast response! Boud (talk) 23:04, 1 January 2011 (UTC)
  • Document type is determined by creation or publication mode. These documents comprise elements of a series of correspondence. They're not websites. Seek another template, or manual citation. Fifelfoo (talk) 02:18, 28 February 2011 (UTC)


I removed the nocode section {{#time: j F Y}} that implies "wikipedia is doing original research". Read wikipedia <code>{{or}}</code>. 14 September 2014Special+Utilizator+$ (talk) 01:42, 28 February 2011 (UTC)

Your edit was reverted, because there seems to be no connection between your edit summary "wikipedia shall not be doing original research", and your action; further, it reduces the usefilness of the copy&paste code chunks. --Redrose64 (talk) 12:58, 28 February 2011 (UTC)


Resolved: Redrose64 (talk) 13:18, 28 February 2011 (UTC)

Why am I showing a bare URL on ref 16 in State reptile?TCO (talk) 04:59, 28 February 2011 (UTC)

The URL has a space in it. Rjwilmsi 07:50, 28 February 2011 (UTC)
Correct. Fixed. --Redrose64 (talk) 13:18, 28 February 2011 (UTC)
Thanks!TCO (talk) 14:34, 28 February 2011 (UTC)

Edit request

The |title field does not put the title in italics like the related templates {{cite}}, etc. could someone please fix this for consistency sake?


Someone. "An example of the error".  ({{cite web}})


Someone, An example of the error  ({{cite}})

Thanks, --Jeremy (blah blahI did it!) 02:27, 2 March 2011 (UTC)

An individual web page is considered to be part of a larger work, thus it is treated as a chapter and placed in quotes, whereas the website is italicized.
Someone. "An example of the error". Name of the website. 
If you have a specific example that should be different, please provide it.---— Gadget850 (Ed) talk 04:51, 2 March 2011 (UTC)
This is pretty much a duplicate of the thread raised at Template talk:Cite book#Edit request, and an example of why they differ is given there. --Redrose64 (talk) 12:00, 2 March 2011 (UTC)
I looked up the reasons why this is in the major citation guidelines, (Harvard, Chicago) and found the reasons why. Article titles and chapter titles are in quotations while book and the publications they appear in, including web sites, are italicized. There are other formatting inconsisencies across the various citation templates I discovered though, and I will be addressing them soon. --Jeremy (blah blahI did it!) 21:21, 2 March 2011 (UTC)
Take a look at User:Gadget850/Citation templates. Some templates use the same style, others were created without regard for the style of the other templates. ---— Gadget850 (Ed) talk 23:26, 2 March 2011 (UTC)

archiveurl and url

If I am not mistaken, a while ago the template produced an error message if both, "archiveurl" and "url" were provided. So I removed "url" from those citations and everything was fine, until... Today I was very surprised to see that the template produced an error message ("Error: If you specify |archiveurl=, you must first specify |url=") if only "archiveurl" was supplied. Is this a feature or a bug? (Not too fond of fixing references every couple of months...) Another question, if the old (non-working) url is deemed important nowadays, why can't it be automatically extracted from the archiveurl? In the case of archive.org, the archiveurl contains the old url. In the case of webcite, the old url is depicted in the top right of the archived page (see for instance [34]). bamse (talk) 14:12, 4 March 2011 (UTC)

The error message concerned was added 19 October 2009 with this edit and the circumstances were amended 16 November 2009 with this edit. The relevant discussion is now at Template talk:Citation/core/Archive 7#Error messages. --Redrose64 (talk) 16:12, 4 March 2011 (UTC)
Thanks for the reply. To be honest, the pages you link to are a bit too technical for me. What I don't understand is, why is "url" required if "archiveurl" is provided? Also, I am pretty sure that I did not see the error message ("Error: If you specify |archiveurl=, you must first specify |url=") until recently (certainly not in 2010). bamse (talk) 16:44, 4 March 2011 (UTC)
As I thought, I initiated that.
What you are seeing now is how it should work. I can't answer for what you saw in the past.
We can't control what archive site might be used and not all archive sites may readily include the original URL. If an archive site ever goes dead, we still have the original URL to work with.
---— Gadget850 (Ed) talk 16:48, 4 March 2011 (UTC)
OK. Thanks. Makes more and more sense to me now. bamse (talk) 16:52, 4 March 2011 (UTC)

Now that I thought I understood everything I got confused again. Sorry for being ignorant, but reference 36 in this article has only an "archiveurl" without a "url" and still ther is no error message. Why??? bamse (talk) 20:48, 5 March 2011 (UTC)

Looks like some of the error checking was removed. This template gives an error if {{title}} is not specified. {{Citation/core}} check for |archiveurl= without {{archivedate}}.
---— Gadget850 (Ed) talk 21:27, 5 March 2011 (UTC)
Hm!? bamse (talk) 21:55, 5 March 2011 (UTC)
It's the way that the relevant parameters in {{cite web}} are mapped to those in {{citation/core}}, and which combinations produce error messages. {{citation/core}} complains when |ArchiveURL= is missing, or when |ArchiveDate= is missing, or when both of |OriginalURL= and |IncludedWorkURL= are missing. {{cite web}} fills these four in from its own parameters as follows:
  |IncludedWorkURL = {{#if:{{{archiveurl|}}}|{{{archiveurl|}}}|{{{url|}}}}}
  |ArchiveURL = {{{archiveurl|}}}
  |OriginalURL = {{{url|}}}
  |ArchiveDate= {{{archivedate|}}}
The significant thing here is the first one: if an |archiveurl= is provided, that parameter is passed through as |IncludedWorkURL=. Unfortunately I don't know why it's done that way. --Redrose64 (talk) 14:44, 6 March 2011 (UTC)

Deprecated parameters

Hey all. I propose the removal of the following code from this template:

{{#if:{{{accessdaymonth|}}}{{{accessmonthday|}}}{{{accessday|}}}{{{accessmonth|}}}{{{accessyear|}}}{{{day|}}}{{{access-date|}}}{{{dateformat|}}}|[[Category:Pages containing cite templates with deprecated parameters|{{NAMESPACE}} {{PAGENAME}}]]}}

These parameters were deprecated over a year ago. Now we are just adding overheads to our thousands of {{cite web}} transclusions (ditto {{citation}} but I'm just going to stick to this template for now. Regards, - Jarry1250 [Who? Discuss.] 22:08, 5 February 2011 (UTC)

As you saw on Arana College some of these incorrect parameters still get added, so I don't think we can drop the maintenance category just yet. Rjwilmsi 09:19, 6 February 2011 (UTC)
I accept that. Two thoughts come to mind, however. Firstly, loads of misspelt or just plain wrong parameter names must get added as well, and we don't attempt to detect for those. Secondly, the template is transcluded over 800,000 times. That's 800,000 pages which experience worse loading times (caching aside), for maybe a handful of useful catches. - Jarry1250 [Who? Discuss.] 11:03, 6 February 2011 (UTC)
We check only for the most common thing, which is adding deprecated parameters, not typos. The reason being that there have been attempts to actively remove them. As to your second point, there is a policy page somewhere which says not to worry about things like loading times etc., when writing Wikipedia. Debresser (talk) 19:27, 6 February 2011 (UTC)
If they had been deprecated a few weeks ago, I would agree with you. But they were deprecated well over a year ago, and the number of article routinely in the maintenance category must surely demonstrate that very few, if any people, are using the parameter names because they can remember them working over 14 months ago. I believe you are referring to WP:PERF, from which I quote: "Particularly in the area of template design, optimising server performance is important, and it's frequently done by users with a great amount of impact". This, I believe, is one of those moments. - Jarry1250 [Who? Discuss.] 19:45, 6 February 2011 (UTC)
(Indeed, it is worth noting that the issue at Arana College was caused by {{citation}}, with its 60-odd thousand transclusions. You see my point here: I am not denying there is a cost to removing the check, but that the benefits outweigh those costs.) - Jarry1250 [Who? Discuss.] 11:57, 6 February 2011 (UTC)
Over at Template talk:Infobox locomotive we recently had a drive to search out and remove unrecognised parameters, also to or fix misspelled or duplicated parameters. It needs the assistance of somebody with the ability to query the database, and so produce something like this. --Redrose64 (talk) 14:56, 6 February 2011 (UTC) amended Redrose64 (talk) 16:40, 6 February 2011 (UTC)
Indeed. I think that's probably the preferable option here too. - Jarry1250 [Who? Discuss.] 15:49, 6 February 2011 (UTC)

{{citation/core}} no longer supports these parameters, so their inclusion in this template doesn't actually do anything but add the maintenance category. I don't see that this check is needed— removal can be added to AWB if it isn't already there.

What about DateFormat={{#if:{{{dateformat|}}}|{{{dateformat}}}|none}} ? ---— Gadget850 (Ed) talk 16:22, 6 February 2011 (UTC)

The category is used to assist in attempts to actively clean up the use of deprecated parameters. Note that part of these have been added as they became deprecated. Debresser (talk) 19:27, 6 February 2011 (UTC)
Perhaps then it would be better to have a new category for any invalid parameters, so that the value outweighs the cost. I expect we'd instantly get several thousand pages needing cleanup. Rjwilmsi 22:44, 6 February 2011 (UTC)
Not sure how you'd do that onwiki. Why not just grep a database dump? Anyone got one handy? - Jarry1250 [Who? Discuss.] 22:57, 6 February 2011 (UTC)
I'll do a database scan. Rjwilmsi 23:18, 6 February 2011 (UTC)
253 distinct invalid parameters having 5 or more uses in January 2011 dump for cite web (1153 different ones with 1 or more use). Rjwilmsi 00:20, 7 February 2011 (UTC)
(collapse table) Excellent. Now, what to do about this: a bot, perhaps? Or a competent AWB user (I assume you kept a list of the articles affected RJ?) Also, the bigger numbers obviously need some care and attention. But what I'm also getting is a non-need for the maintenance categories. Does anyone disagree? - Jarry1250 [Who? Discuss.] 16:58, 7 February 2011 (UTC)
For the ones that are obvious typos I'll do the cleanup (example). We'll have to see about the other ones. Rjwilmsi 19:35, 7 February 2011 (UTC)
Citations using parameters such as editor, journal or isbn — if non-empty — make me wonder if cite web is appropriate. Did we ever have a deadurl parameter? The rest are mostly typos, and a lot could be copy/paste from another template. Regardless, I think that if this bit is cleaned up, we shouldn't need to worry about it anymore. ---— Gadget850 (Ed) talk 20:45, 7 February 2011 (UTC)
Shall we proceed with removal of the tracking category as ineffective? - Jarry1250 [Who? Discuss.] 21:38, 15 February 2011 (UTC)
Why not keep them? Editors are likely to use them again once in a while, and this way they are tracked and can easily be fixed. Especially since this same detection is active on another 20 citation templates. Debresser (talk) 13:51, 16 February 2011 (UTC)
In addition, I think that adding detection of these parameters to AWB is a must before this detection could be removed. Debresser (talk) 13:53, 16 February 2011 (UTC)

What can we do with deadurl? I think it's added by a bot when the original url is dead and replaced by a link to some webarchive. Do you think we should add it to the manual? -- Magioladitis (talk) 10:01, 12 March 2011 (UTC)

subtitle parameter

Does anything that would serve the purpose of a "subtitle" tag exist? For example, consider this page: http://your.kingcounty.gov/elections/2003nov/respage18.htm . The title of the page is "King County Election Results", but if I want the citation to refer to the Redmond results (but not the other city on the page) it would make sense to indicate "subtitle=Redmond" if such a tag (or something similar) exists.

Obviously, I could use "title=King County Election Results: Redmond", or press the "quote" tag into service as place for a subtitle, but neither solution seems perfect.

Would a subtitle tag (or some other means of indicating a section within a web page that includes a lot of material unrelated to the part that matters) make sense? — Steve98052 (talk) 12:04, 11 March 2011 (UTC)

Seems a short page to go to that trouble, but you can use |at=. ---— Gadget850 (Ed) talk 12:19, 11 March 2011 (UTC)
I was actually using that page to make the example easier to read; I was more interested in the general case of a page where it's actually difficult to find the appropriate subtitle. Anyway, thanks a bunch for the quick answer. That resolves my problem. — Steve98052 (talk) 21:55, 11 March 2011 (UTC)

Accessdate to copy&paste


the German Wiki has a similar template (de:Vorlage:Internetquelle#Kopiervorlagen). If you take a look at it, you can see a section that can be copied. It looks like this: {{Internetquelle | url= | titel= | zugriff=2011-03-15}}. This is the most basic version of the template. The URL, a title and the access-date. The access-date is filled by default to the actual date. I think it would be practical to make a similar template to copy and paste for the english wiki. --MartinThoma (talk) 19:35, 15 March 2011 (UTC)

We already have this; but the access date is pre-filled in the two formats preferred by English Wikipedia. --Redrose64 (talk) 20:39, 15 March 2011 (UTC)
Oh, sorry, I didn't see this. I only saw the complete template (a bit lower). Thanks Redrose64! This discussion can be closed. --MartinThoma (talk) 07:35, 16 March 2011 (UTC)


Is there a difference (or preferrence) between "cite web" and "Cite web"? The posted examples for annotation are lower case, but the names of the templates are then capitalizeed. In several WP articles they are being "corrected" to capital. Can the topic be made more clear in the article please?MartinezMD (talk) 04:23, 19 March 2011 (UTC)

There is no difference in the rendering of the citation between "cite web" and "Cite web". Whether there is a preference: well, previous discussion established that most editors don't see any point in changing existing capitalisation either way. If the "correction", as you call it, is recent (last few weeks), and you want to complain about it, you could probably take it up with User:CBM or User:Xeno since they are admins who contributed to previous discussion in this area. Rjwilmsi 11:45, 19 March 2011 (UTC)
For me, I don't care. If it makes no difference, that's fine. I would like that point then made on the template article. I think it would avoid unnecessary work (and maybe arguments).MartinezMD (talk) 13:18, 19 March 2011 (UTC)
It is a trivial matter and applies to all templates. I see no need to mention it. ---— Gadget850 (Ed) talk 14:56, 19 March 2011 (UTC)
I don't care about capitalization either. Just wanted to point out that some of the "corrections" are due to automatic tools such as checklinks or WP:AWB and possibly others. bamse (talk) 14:58, 19 March 2011 (UTC)

Supression of "ed." where authors named?

It seems that the appendage "ed." for editor is not shown when authors are named.

For example:
{{cite web|url=http://www.example.com |editor=Editor Name|title=Example.com}} yields
Editor Name, ed. "Example.com". http://www.example.com.

With an author:
{{cite web|url=http://www.example.com |first=Firstname |last=Lastname |editor=Editor Name|title=Example.com}}
Lastname, Firstname. "Example.com". In Editor Name. http://www.example.com.

This gives rise to an inconsistency between references. Is this intentional and in accordance with style guidelines? Thanks, Skomorokh 18:59, 15 April 2011 (UTC)

There are no style guidelines associated with citation templates. They do whatever they do. Jc3s5h (talk) 01:32, 16 April 2011 (UTC)
But this series of templates is intended to work together and are intended to have the same style. ---— Gadget850 (Ed) talk 11:50, 16 April 2011 (UTC)
And yes, this is how this works. See {{cite book}}, the examples for Citing a chapter in a book with different authors for different chapters and an editor and Citing a chapter in a book with two joint authors and an editor. ---— Gadget850 (Ed) talk 13:38, 16 April 2011 (UTC)


WebCiteBOT added a lot of dead=no parameters when the URL was online but an archiveurl was provided. The purpose of the "dead" parameter was to swap the links, and mantain the real link first, and the archive url the last.

Now, some people are working in a new WebCiteBOT, so, probably, this parameter is going to be more used. Regards. emijrp (talk) 16:19, 21 April 2011 (UTC)

Problem with authorlink

Please advice me on this reference.[1]

  1. ^ [Andrew] Check |authorlink= value (help) (9 October 2007). "Why the square-root rule for vote allocation is a bad idea". Statistical Modeling, Causal Inference, and Social Science. Columbia University website. Retrieved 30 April 2011. 

As you can see, the authorlink is formatted in an unintended way. I tried but could not get rid of the brackets and bar. Tomeasy T C 10:47, 30 April 2011 (UTC)

|authorlink= is the name of the article on the subject, not an external link.
{{cite web | last=Gelman | first =Andrew | authorlink = Andrew Gelman | title =Why the square-root rule for vote allocation is a bad idea | work =Statistical Modeling, Causal Inference, and Social Science | publisher =[[Columbia University]] website | date =9 October 2007 | url =http://www.stat.columbia.edu/~cook/movabletype/archives/2007/10/why_the_squarer.html | accessdate =30 April 2011 }}
Gelman, Andrew (9 October 2007). "Why the square-root rule for vote allocation is a bad idea". Statistical Modeling, Causal Inference, and Social Science. Columbia University website. Retrieved 30 April 2011. 
---— Gadget850 (Ed) talk 10:57, 30 April 2011 (UTC)
Thanks. Tomeasy T C 11:14, 30 April 2011 (UTC)

Citing Credo Reference

I have a Credo Reference like this:

"poplin." The Thames & Hudson Dictionary of Fashion and Fashion Designers. London: Thames & Hudson, 2007. Credo Reference. Web. 05 June 2011.

When reformatting this to use the {{Cite web}} template, what parameter should I use for "Credo Reference"? Thanks! - PKM (talk) 18:19, 5 June 2011 (UTC)

What's a Credo Reference? ---— Gadget850 (Ed) talk 19:22, 5 June 2011 (UTC)
See Credo Reference, unless I'm very much mistaken. In such cases I recommend the use of {{subscription required}} in addition. --Redrose64 (talk) 20:06, 5 June 2011 (UTC)
Ah— I wondered what happened to Xrefer. It is a host and we don't include the host in the reference. ---— Gadget850 (Ed) talk 22:10, 5 June 2011 (UTC)
Pardon my stupidity, but why not? Surely it falls under "cite where you got the information"? - Jarry1250 [Weasel? Discuss.] 22:20, 5 June 2011 (UTC)
A host that has nothing to do with the creation of the content is no different from my local library or the top of my bookshelf. Google Books, Scribd or Project Gutenberg have noting to do with the content and it appears that Credo does not either. ---— Gadget850 (Ed) talk 02:27, 6 June 2011 (UTC)
Yes, good point. I suppose if it were a less reliable host we would note it, though. - Jarry1250 [Weasel? Discuss.] 09:33, 6 June 2011 (UTC)
I'm not sure what would make a host reliable or unreliable. If it hosts content that it has no permission then that is a problem. ---— Gadget850 (Ed) talk 10:37, 6 June 2011 (UTC)
Thanks. - PKM (talk) 03:16, 7 June 2011 (UTC)

When to use authorlink= parameter?

Please see Template talk:Citation/core#authorlink parameter on a question of when the |authorlink= should be used and when the |author= and |editor= fields can be linked directly. —  HELLKNOWZ  ▎TALK 12:17, 16 June 2011 (UTC)


Note: The below was posted to Wikipedia talk:Template:Cite web, seemingly by accident, so I have moved it here. Avicennasis @ 00:00, 30 Sivan 5771 / 2 July 2011 (UTC)

Couldn't there be some automatic auto-archiving of links when this template is used. Since link rot is a real problem. I often use WebCite for links of which I think they may not stay around forever. SpeakFree (talk) 00:08, 30 June 2011 (UTC)

arxiv, asin, etc etc

What are the fields for arxiv, asin, "jfm", "lccn", "mr", "ol", "osti", "rfc", "ssrn" and "zbl" for and what do they all stand for? The documentation has no information, and the last time I used the template they weren't there. Thanks, Matthewedwards :  Chat  04:24, 23 June 2011 (UTC)

No one knows? :( Matthewedwards :  Chat  22:55, 30 June 2011 (UTC)
They are mostly additional identifiers for journal papers that are of use in {{cite journal}} and {{citation}} so were added to Template:Citation/core. It seems unlikely that they would be used in cite web. When they were added the documentation of the templates wasn't updated. (talk) 08:46, 1 July 2011 (UTC)

bo.lt as reference

As part of this edit I just discovered that bo.lt is not an URL shortening service. As they put it, "When you put a page on BO.LT, it is a copy. Don't worry, the original page on the original site is still safe. Your page, however, is nimble, changeable, and under your control." Given its stated purpose, I suggest someone look at http://bo.lt/ more closely and if it is what it seems to says it is, put it on the WP:BLACKLIST or ad it to some other automated check system such as WP:FILTER or ClueBot (task list · contribs). (talk) 05:21, 11 July 2011 (UTC) P.S. I posted here in spite of the {{notice}} on the talk pate because WP:CTT has fewer watchers.


Ideally, each newly introduced reference that uses a template like 'Cite web' gets an archiveurl and archivedate parameter, e.g. by WebCiteBOT. Unfortunately, the latter appears defunct and even if ever repaired it will have a tremendous backlog to handle. So we are stuck with a number of references that do not allow to find an archived page (or to know whether that even exists).
Finding a dead link (be it signalled by the {{dead link|date=.. ....... ....}} template or not) entices to find out where the web page may have been moved to. When found, it is reasonable to replace the dead url in the url= parameter with the live one. The problem however, is that one cannot compare the old url's page with the newly found: it may have changed and not (as well) represent the text it had originally referenced; it could make the originally introducing editor look stupid. I therefore suggest to keep the original url and the original accessdate in separate fields: The updating editor might simply rename '|url=' to '|introductionurl=' and '|accessdate=' to '|introductionaccessdate=' (content for simpler parameter names like oldurl and oldaccesdate might become replaced when a web page got moved a second time), and add his/her found '|url=' and '|accessdate='. It would maintain a trace that might allow another editor who is capable of finding the archive of the original url upon his/her suspecting it to have said something else, and then restore while filling in the archiveurl and archivedate fields.
Programs that handle templates are not supposed to malfuntion in case a parameter name is not recognized. I need to repair a few links as described here above, and will apply '|introductionurl=' and '|introductionaccessdate='; at least the necessary info is then retained and the reference will function.​▲ SomeHuman 2011-02-04 06:26 (UTC)
In fact, the field '|introductionaccessdate=' might be inserted with its date copied from '|accessdate' upon creation of a reference (automatically by a bot): it warrants simply finding back the article version that became referenced, even if the url and (or only) the (last) accessdate become updated by an editor. This can facilitate maintenance of an article for detectig unsourced material having replaced or been inserted into properly sourced text, without ever having been noticed. It would just as well facilitate finding the originally introduced url from the relevant old version. Hence, whether an often very lengthy '|introductionurl=' would best go into the template call (upon the need arising as above described), may be a matter of opinion with respect to technical capabilities.​▲ SomeHuman 2011-02-04 07:09 (UTC)

authorlink does not work

I think authorlink does not work. For example {{cite web | first = John | last = Smith | authorlink = http://www.johnsmith.name/ | title = A book | publisher = Some publisher | date = August 2011 | url = http://www.someurl.com | accessdate = August 2011}}


[John] Check |authorlink= value (help) (August 2011). "A book". Some publisher. 

I repeat for posterity: [|Smith, John] (August 2011). "A book". Some publisher.

The extra brackets and bar disappear when removing authorlink

Smith, John (August 2011). "A book". Some publisher. 

Turesable (talk) 16:25, 10 August 2011 (UTC)

The authorlink field is for creating internal links to a Wikipedia article about the author. It is not designed to support external links to the author's website. --RL0919 (talk) 16:28, 10 August 2011 (UTC)

Request to add deadurl parameter

Per Wikipedia:Requests for comment/Dead url parameter for citations, could the following change please be made: Sandbox Source ([35]).

This request is one of several follow-ups on implemented change to Citation/core. I am requesting {{cite web}} alone first to make sure everything will be in order.

The request adds an optional |deadurl= parameter that is passed down to {{Citation/core}} as |DeadURL=. Additionally, if |deadurl=no then the main |IncludedWorkURL= link sent to Citation/core will be the original link and not the archived one. This will (in Citation/core) change the archive text to lead to the |archiveurl= (see RfC or sandbox below for preview).

Here are all the test cases I could think of: User:H3llkn0wz/Sandbox3 (currently limited to {{Cite web}} in actual sandbox implementation, but showing all that use it to make sure nothing breaks; note that post-expand size is too big there). —  HELLKNOWZ  ▎TALK 17:28, 12 June 2011 (UTC)

Thanks. —  HELLKNOWZ  ▎TALK 08:33, 13 June 2011 (UTC)
Would there be any way to account for the use of |deadlink being accidentally used instead of |deadurl? Or is that something that would be better handled by an AWB citation bot? NW (Talk) 21:39, 10 August 2011 (UTC)
Already covered, by bot task. Rjwilmsi 09:35, 11 August 2011 (UTC)
That settles that. Thanks for the quick reply. NW (Talk) 18:42, 11 August 2011 (UTC)

Link order ([Original→backup] instead of [Backup→original])

>"List of psychotropic substances under international control" (PDF). 30 April 2005. Archived from the original on 11 September 2005. Retrieved 6 July 2005. 
Wouldn't it be better to put the original link first and only then the backup copy address? The majority of users will not be caring which link would it be to be clicked, though the original publishers will be losing their webtraffic and thus being more inclined to forbid archiving of their projects. [Original→backup] order would've also decreased the load on IBA projects, which, I guess, are non-profit and will be more inclined to care about their sustainability and solid 24/7/etc availability . DaemonDice (talk) 19:56, 30 August 2011 (UTC)

That is a function of {{citation/core}}— please discuss there. ---— Gadget850 (Ed) talk 20:16, 30 August 2011 (UTC)
You mean |deaurl=no"List of psychotropic substances under international control" (PDF). 30 April 2005. Archived from the original on 11 September 2005. Retrieved 6 July 2005.  —  HELLKNOWZ  ▎TALK 20:21, 30 August 2011 (UTC)
Yes. Sorry, I overlooked that. But I meant activating that option by default. (And adding something like {{para|deaurl|yes}} if needed) 30-08-2011, 20:50 — DaemonDice (talk)
It cannot be activated by default now, because it first needs to be checked if the original is still live. The point of an archived copy is that the original is presumed to be dead. But lately, a lot of pre-emptive archiving has been done in fear of linkrot, so the optional parameter was introduced. —  HELLKNOWZ  ▎TALK 07:18, 31 August 2011 (UTC)
Why is it impossible in cases of pre-emptive backuping then?
Anyway, my intention was simply to state here the issue which in my opinion does exist. If it's impossible to solved at the moment, then I guess one may only regret about that and move on. Thank you for helping to sort the nuances out. 31-08-2011, 14:01 — DaemonDice (talk)
It's supposed to be used with pre-emptive archiving, which is what the deadurl parameter was added for. All the other (>90%) of cases are dead links though, and that's thousands of articles. I'm just saying changing their display to reverse the links is not helpful. —  HELLKNOWZ  ▎TALK 14:51, 31 August 2011 (UTC)

language parameter... vs language icon

It appears that the format that the cite web template uses for the language parameter is different than the format used by the language icons. I recall having seen the language icon used in references before. Is this an inconsistancy? example that shows both in action Jason Quinn (talk) 03:45, 3 September 2011 (UTC)

Work and publisher fields

Hi, see thread at citation templates. I originally posted there because I thought my query may apply to some of the other citation templates, although in hindsight that may not be the case. Please respond there in order to follow the thread. Thanks. Eldumpo (talk) 16:52, 6 September 2011 (UTC)

Missing URL

... or not. I occasionally come across cases where {{cite web}} has a URL, but it isn't in the actual |url= parameter, i.e. it's in an unnamed (positional) parameter caused by the simple omission of the crucial characters url= - see, for example, this edit. This is treated in the same way as if the URL were omitted entirely: that is, no clickable link is created and no error is thrown. This is in contrast to the case where |title= is blank or omitted, which does throw an error. I consider that a missing URL is a greater error than a missing title, so how about testing for a missing or blank |url= parameter and throwing a similar error (although not identical: how about {{Citation error|no |url= specified|Cite web}}). --Redrose64 (talk) 17:22, 30 September 2011 (UTC)

OK by me. ---— Gadget850 (Ed) talk 19:12, 30 September 2011 (UTC)

Question about deprecated parameters

Although the article Lead has a {{cite web}} template that contains |dateformat=dmy, the article is not included in Category:Pages containing cite templates with deprecated parameters. I don't understand the cite web source code: I see {{{dateformat|}}} in an if statement that appears to put the article in this category, but I also see |DateFormat={{#if:{{{dateformat|}}}|{{{dateformat}}}|none}}, which makes it appear that the parameter is still functional. Could someone please explain this to me? Thanks! GoingBatty (talk) 16:33, 2 October 2011 (UTC)

The problem on Lead is that the same ref name has been used for two different refs.
<ref name="leadorg">{{cite web|url = http://www.ldaint.org/technotes2.htm|title = Primary Lead Refining Technical Notes|publisher = LDA International|accessdate = 7 April 2007 |archiveurl = http://web.archive.org/web/20070322191856/http://www.ldaint.org/technotes2.htm <!-- Bot retrieved archive --> |archivedate = 22 March 2007}}</ref>
<ref name="leadorg">{{cite web|url = http://www.infomine.com/commodities/lead.asp|title = Global InfoMine{{ndash}} Lead Mining|publisher = GlobalInfoMine|dateformat = dmy|accessdate = 17 April 2008}}</ref>
When this happens, the content of the second is ignored. The |dateformat=dmy only occurs in the second one, which is why it's not coming up in Category:Pages containing cite templates with deprecated parameters. --Redrose64 (talk) 16:48, 2 October 2011 (UTC)
In answer to your second q: although {{cite web}} does indeed pass the dateformat through to {{citation/core}}, via the latter's |DateFormat= parameter, {{citation/core}} ignores that parameter - it was removed in February 2009 with these four edits. --Redrose64 (talk) 16:54, 2 October 2011 (UTC)
Thanks for the quick response! I'll post on Talk:Lead and ask for help with this cleanup. GoingBatty (talk) 16:56, 2 October 2011 (UTC)

What is the accessdate used for?

Resolved: We use it here because all citation style authorities demand this information.

I use WP:Twinkle which can fill in the accessdate field. The field seems useless to me and the resulting "Retrieved some date" clutters the reference list. Why does the template display the date? --Marc Kupper|talk 21:33, 30 July 2011 (UTC)

Because web page are prone to amendment, and few are dated with any form of decent revision date: often the only date on a page is the copyright year, which is at best vague (covers a 365-day span) and at worst always seems to be the current year, even if the page in question was last amended a year ago. For the purposes of verifiability, it is very important that we be able to detect if a current version of a web page is not the same as the one which existed at the time that the reference was added. --Redrose64 (talk) 21:45, 30 July 2011 (UTC)
I still don't see how retaining and displaying the accessdate helps with this. Let's say a page is dated. In that case we can document it with the date field and someone later verifying the citation can see if the date matches. If a web page is not dated then I don't see how accessdate allows me to detect that the "current version of a web page is not the same as the one which existed at the time that the reference was added." Let's say I look at the page and see that it seems to support the Wikipedia article. Did the accessdate date help me? No. Has the web page been revised since it was used as a citation? I have no idea. Let's say I look at the page and see that it does not seem to to support the Wikipedia article. Does knowing the accessdate help me? No as I don't know if the web page changed, if the editor who added the cite was mistaken, or if the cite ended up where it was in article as a result of editing.
One potential value of accessdate is it allows me to know when someone added that citation to the article. Using the article history I could see what it looked like at that time and see if the citation supports that version. However, if someone knows Mediawiki well enough to do that then an editor likely knows they can view the page in edit mode (or view-source for protected articles). The other option is to have the accessdate available as hover text (float the mouse here to see what I mean) meaning there would be less reflist clutter while still making the information available. --Marc Kupper|talk 23:09, 30 July 2011 (UTC)
The access date is important because the content may change. It may be removed, it may be altered, it may be expanded. Besides, sources may not have a date at all, like a team roster. MediaWiki page history is an unreliable way to determine access date, simply because adding a citation does not equal retrieving the page. One may have written a draft months ago, copied content, restored long-term blanking vandalism, etc. One of the other main reasons to fill in access date is so that the page can be properly archived and restored later on, when it inevitably goes 404. Be it via archive.org, URL change, or a different citation. —  HELLKNOWZ  ▎TALK 07:53, 31 July 2011 (UTC)
Hi, just thought I'd flag this up here, the access date appears to be broken, usually when one comes to get the template the access date is always today's date and today it says 28 July 2011. What's up, Doc? CaptainScreebo Parley! 20:52, 1 August 2011 (UTC)
I'm seeing 1 August 2011 (today is 2 August). You're probably viewing a cached copy of the page. Try purging. --Redrose64 (talk) 15:02, 2 August 2011 (UTC)
That would be it then ,cheers, now says 2 August, will bear this in mind for the future. CaptainScreebo Parley! 15:39, 2 August 2011 (UTC)
I likewise see no point in cluttering up reference lists with lists of the last access date. It's great that it's stored in the code for posterity, it does have some benefits for someone creating or searching for an archive (and therefore modifying the code), but it should not be displayed, should be behind some javascript magic, or should be so tiny that it doesn't blend in with the rest of the list. It's a HUGE amount of visible text clutter for no benefit. I've tried to use javascript on my end to hide it, but because it's not enclosed in any tags, I can't. It's pointless and needs to go. Foxyshadis(talk) 02:31, 9 September 2011 (UTC)

The accessdate parameter is first of all confusing. From the usage patterns for the last few years, most editors put both |date= as the date of publication and |accessdate= as the actual date they added the citation to the article; however the documentation currently says that |accessdate= "should be given if the publication date is unknown; see Citation styles - Webpages" - so |date= and |accessdate= are presumed to be mutually exclusive! How did it happen that so many editors got this wrong - was it a recent policy change? Anyway, this is not emphasized enough in the documentaion, so I moved the description of accessdate to immediately follow date for now. --Dmitry (talkcontibs ) 09:28, 17 September 2011 (UTC)

You have to be careful with the wording though. If the date is given by |year=, the citation would still need an access date. Also, |accessdate= should not be used instead of the |date=, so the "or" part does not work both ways. —  HELLKNOWZ  ▎TALK 10:55, 17 September 2011 (UTC)
The documentation suggests specifying both |year= and |month=, not the year alone. This was not clear at the first glance, so I have moved the parameter names to be closer together, and also changed the description of |accessdate= to read "should only be given etc." --Dmitry (talkcontibs ) 11:30, 17 September 2011 (UTC)
That's assuming month is even available, which is not a fact. Anyhow, I clarified the documentation so you see what I meant -- if |year=2001 is set, then |accessdate=November 11, 2011 should still be set. —  HELLKNOWZ  ▎TALK 12:27, 17 September 2011 (UTC)
I restored the previous language and placement. Accessdate is not an alternative option for the other dates. It is most important (or required per Citation styles - Webpages) if the publication dates are not available. However, it is always good practice to include the access date for a web page. With the access date, it is often possible to recover information from dead urls using the Internet Archive. olderwiser 13:03, 17 September 2011 (UTC)
I am really looking for the language I thought I've seen that accessdate is not required where there is a publication date. (Tips would be appreciated.) But having glanced over several similiar discussions I wonder if I might offer some summarization. "Publication" has traditionally implied a definite form or content, pretty much fixed as of the date of publication. Publishers might subequently alter the content, but this is generally deemed a revison, or even a new edition, and is distinguished by a different publication date. Web pages, and possibly other sources of similiar malleability, can be continuously modified ("prone to amendment", as Redrose64 said above), and therefore must be specified as of a specific time (date alone is usually considered sufficient); this is equivalent to a time of publication. Where material from a web page (not otherwise published in a definite version) is used in an article the access date is when the material was accessed, not when it was included in the article.
Perhaps that helps. One of the issues I've seen is where editors reference a web form of material with a definite publication date (such as a book, or a journal article); it seems over pedantic to cite when a book was "accessed". Although I have seen pdfs of journal articles that were subsequently modified (corrected??). That seems to warrant a revision date (do we have that?), perhaps even a file size or (!!) CRC of the file content. But these verge off-topic, and I mention them only to show that access date is not the only method of identifying content. _ J. Johnson (JJ) (talk) 21:54, 10 October 2011 (UTC)
  • Template:Cite web states: "accessdate: Full date when item was accessed. Should not be wikilinked. This should be given if the publication date is unknown." This is included as one of the optional parameters.
  • Wikipedia:Citing sources#Web pages states: "Citations for World Wide Web pages typically include ... the date of publication (if known), the date you retrieved the page, for example Retrieved 2008-07-15. (this is required if the publication date is unknown)."
  • Template:Cite book states: "accessdate: Full date when url was accessed. Should be used when url parameter is used."
  • Template:Cite news states: "accessdate: Date when the news item was accessed, if it was found online." This is included as one of the optional parameters.
  • Template:Cite journal states: "accessdate: Full date when URL/DOI was last checked." This is included as one of the common parameters.
Hope this helps! GoingBatty (talk) 02:31, 11 October 2011 (UTC)
When correcting wp:LINKROT (dead links), it is very useful to know what date the url was accessed. It often helps in identification of the correct snapshot archived by the Wayback machine (or other archiving engine). LeadSongDog come howl! 15:08, 11 October 2011 (UTC)

(Outdent). A much shorter answer is that all citation style authorities of any note demand that proper source citations in any kind of formal writing (scientific papers, legal briefs, student essays, post-graduate dissertations, etc., etc.) provide date of access of online materials. Regardless whether all Wikipedians find this information useful, enough people in the real world do find it useful and expect it, that we'd be collectively idiotic to disallow our citation templates from including such metadata. — SMcCandlish Talk⇒ ʕ(Õلō Contribs. 18:31, 14 December 2011 (UTC)

Linking to publisher and work entity articles

Hello, there is a bot approval request at WP:BRFA/H3llBot 9 for adding wikilinks to work/publisher fields where the entity can be unambiguously identified from a pre-selected list. Comments welcome, thanks. —  HELLKNOWZ  ▎TALK 16:18, 28 August 2011 (UTC)

As long as it actually gets this right. I cannot tell you how many times (surely over 1000 by now) I have personally had to correct |publisher= to |work= because for some reason this template in particular seems to short-circuit people's brains and make them treat the name of a website as the name of the publisher instead of the work. This makes as much sense as confusing the album name of Led Zeppelin's The Song Remains the Same with Rhino Records, the label that issued it, or confusing the game Doom with id Software, the software company that produced it. It completely mystifies me that this happens at all, and the only explanation that makes sense to me is that broken tools are doing it automatically. I know for a fact that WP:Cite4Wiki was originally doing this, which is why I took it over and fixed it when the original authors were non-responsive. — SMcCandlish Talk⇒ ʕ(Õلō Contribs. 18:37, 14 December 2011 (UTC)

Reqest to allow URL to be used with parameter authorlink

As many cited authors don't have a lemma on Wikipedia, but have internet pages with valuable information about themselves and their work, allowing WP readers to better judge their credibility, I would suggest to allow also URLs for the parameter authorlink. Many users of this template assume this to work anyway (see #authorlink does not work and #Problem with authorlink), because the name link is confusing and can be used for both: a WP link and an external link as well. In case it's too complicated to implement both usages in one parameter, I'd suggest to change the authorlink parameter name to a less confusing authorlemma and have a new parameter authorurl in addition. For downward compatibility authorlink could still be used in the code, but documented as obsolete and replaced by authorlemma.--Berny68 (talk) 12:03, 25 September 2011 (UTC)

For one, I think the name authorlemma would be much more confusing than authorlink. I have never heard the word “lemma” used to refer to Wikipedia article, and I think most other people too. User<Svick>.Talk(); 21:06, 6 December 2011 (UTC)

Edit request from , 10 November 2011

  1. The page lists 3 possible archive parameters, but speaks of "both" of them. "Deadurl" is the new 3rd one. A grammar fix is needed; I suggest "(if one of the first two is used, the other must be used as well)".
  2. I propose a 4th possible parameter, archived=[WC|IA|no|maybe] which defaults to maybe.
  • archived=IA is substituted with [http://wayback.archive.org/web/*/[URL] Internet Archive copy]. That is, it displays a link to the IA page showing links to all archived copies of the cited page.
  • archived=WC is substituted with HTML so a click brings up a WebCitation.org page showing the results of a query for the base URL.
  • archived=no produces nothing
  • archived=maybe produces a combination of =IA and =WC: Possibly working archive search links: [http://wayback.archive.org/web/*/[URL] Internet Archive possible copy search]; WebCitation archive search.

The idea is to make it easier to create working archive links, and less likely for folks to simply remove dead links simply because they're dead, which is AGAINST POLICY.

Clear? Thoughts? I'm really tired of seeing folks damage articles by removing dead links simply because they're dead, when there are archives available.

Elvey (talk) 21:46, 10 November 2011 (UTC)

I set the edit request to no for now because you don't have a clear change to make immediately; see Wikipedia:Edit requests. The archive parameters are passed to and implemented in {{citation/core}}, which is where any new parameters must be added. ---— Gadget850 (Ed) talk 00:16, 11 November 2011 (UTC)

author parameters

This template recognises |author= and |authorn= where n is an integer between 2 and 9: |author1= is invalid. This is in contrast to {{cite book}} where n may be from 1 to 9, and |author1= is a synonym for |author= - is there a reason for the difference? --Redrose64 (talk) 14:06, 20 November 2011 (UTC)

Actually, {{cite web}} only supports |author= and |author2=. The author and editor parameters should be updated from {{cite book}}. ---— Gadget850 (Ed) talk 14:19, 20 November 2011 (UTC)
Hmm, you're right. I rarely use the author parameters, preferring |lastn=|firstn= - but I was tidying a {{cite web}} citation which had used |author= to hold two names. I altered it to |author1=|author2= - and saw that it didn't work, so checked for the validity of those two. I kind of assumed then that if |author2= was valid, |author3= etc. would be too. --Redrose64 (talk) 15:41, 20 November 2011 (UTC)
I truly cannot see any issues, so I made the updates. ---— Gadget850 (Ed) talk 17:05, 20 November 2011 (UTC)
Could you please also update the documentation at Template:Cite web/doc? Thanks! GoingBatty (talk) 20:58, 20 November 2011 (UTC)
I have been working on it since the update. Had a real life interrupt. ---— Gadget850 (Ed) talk 21:35, 20 November 2011 (UTC)
When you have a chance, could you please add |editor= to the appropriate place in the All parameters section?
Also, the Optional parameters section states "The template automatically adds "ed." after the editor's name unless the chapter parameter is used in which case the template adds "in" before the editor's name which appears after the chapter and before the title." However, in Formation and evolution of the Solar System, reference 108 uses "in" even though the |chapter= parameter is not used. Is this an issue with the code or documentation? Thanks! GoingBatty (talk) 01:38, 16 December 2011 (UTC)
Fixed in documentation. This template does not have |chapter=, but |title= is always formatted as a chapter. ---— Gadget850 (Ed) talk 13:34, 16 December 2011 (UTC)
Since the |author= is deprecated, shouldn't the example on Template:Cite web/doc be changed? DarkFrog (talk) — 09:06, 29 December 2011 (UTC)
I don't think that |author= is deprecated. There are times when |last=|first= are unsuitable - for example, corporate authors which differ from the publisher. Another time when |author= may be the better choice is for foreign names when the person adding the ref may not (as a native English speaker) be fully aware of the name conventions in that language. Consider somebody like Kim Jong-il: we know that his parents were Kim Il-sung and Kim Jong-suk, and his son is Kim Jong-un, so the family name is likely to be Kim. If he had written a book which we use as a ref source, should we put |first=Jong-il|last=Kim - or play safe with |author=Kim Jong-il? --Redrose64 (talk) 11:49, 29 December 2011 (UTC)
I strongly favor separating first and last names, and deplore your example, because I have seen quite a bit of editors sloppily (lazily?) stuffing everything into one field, sometimes inverted, sometimes not. Having said that, I agree that |author= should not be deprecated: there are cases where "author name" is a single entity, and it would be incorrect to classify it as "first" or "last". ~ J. Johnson (JJ) (talk) 19:23, 29 December 2011 (UTC)
Technical comments: These parameters are aliases and feed into the Surname1 meta-parameters: last, surname, last1, surname1, author1, author, authors and author. There is no difference in visual or meta-data output.
  • Use lastn and firstn for normal use. This creates separate fields in the meta-data rendering and creates proper anchors for shortened and harv footnotes. Author should be deprecated for this use.
  • The last and first parameters are not ideally suited to authors whose surname is usually written first (e.g. as in Chinese) or for a corporate or academic entity credited as the author. Use last to include the same format as the source. Or use author, since it doesn't really matter.
---— Gadget850 (Ed) talk 21:03, 29 December 2011 (UTC)
Hmmm... There is an extra author parameter in the template. Looks harmless, but should be cleaned up. ---— Gadget850 (Ed) talk 21:07, 29 December 2011 (UTC)
I think what should be deprecated is this certain usage of 'author', not of 'author' itself. ~ J. Johnson (JJ) (talk) 23:03, 29 December 2011 (UTC)
Note that Reflinks uses |author= when it adds citation templates. GoingBatty (talk) 00:45, 30 December 2011 (UTC)

Instead of deprecating author, why not just note the aliases:

Aliases: last, author; last1, author1 through last9, author9

-— Gadget850 (Ed) talk 02:00, 12 January 2012 (UTC)

URL parameters

What does this mean: "Any parameter that provides a URL (excluding archiveurl, isbn, issn, etc.) may be substituted." ---— Gadget850 (Ed) talk 09:40, 2 December 2011 (UTC)

I think that means you can do ...|url={{Allmusic|class = foo}}... and it works. Rjwilmsi 11:35, 2 December 2011 (UTC)
The url field gets merged with the title to produce an external link, so anything other than a proper URL will mangle the link.
Example: [Led Zeppelin at AllMusic "Led Zeppelin"]. .
---— Gadget850 (Ed) talk 11:55, 2 December 2011 (UTC)
Yes, it only sort of works. There are quite a lot of citations like this, particularly with the {{Allmusic}} template (not personally in favour, just providing information). Rjwilmsi 12:09, 2 December 2011 (UTC)
I think it might be confused with the action of {{cite journal}}, where if you have provided |pmc= (or similar), then |url= may be omitted, but the title is still set up as an external link. Consider the following, which has no |url=:
Might be worth asking the editor who amended the doc. --Redrose64 (talk) 16:30, 2 December 2011 (UTC)

Needs edition parameter

Like just about an other work, this should have an "edition" parameter, with code borrowed directly from {{cite book}}. This is especially important as books move online and are frequently cited with {{cite web}} instead of {{cite book}}. In the docs, it should not be included in the "quick copy-paste" versions, since it won't be used much. — SMcCandlish Talk⇒ ʕ(Õلō Contribs. 18:41, 14 December 2011 (UTC)

Yes check.svg Done Also added |others= since it was documented but not actually in the markup. Doc updated. ---— Gadget850 (Ed) talk 20:56, 14 December 2011 (UTC)

D'oh! I wonder how that got left out. While we're at it, is there anything else we can do to make the options and output between these two templates more consistent? For those of us who cite, cite, cite it can be a bit of a hair-pulling exercise to try to remember all of the quirks these templates display. — SMcCandlish Talk⇒ ʕ(Õلō Contribs. 09:43, 15 December 2011 (UTC)
Probably added to the doc in a copy/paste. I have actually been working on making the CS1 templates more consistent bit by bit, as well as updating other templates to comply with CS1. ---— Gadget850 (Ed) talk 13:21, 15 December 2011 (UTC)
That's what I mean. I wasn't thinking about stuff outside CS1, which I believe will ultimately go extinct. — SMcCandlish Talk⇒ ʕ(Õلō Contribs. 12:36, 16 December 2011 (UTC)


I noticed that this template documents an |embargo= parameter:

  • embargo: The date that an article will be freely accessible at PubMed central

However the markup uses |pmc-embargo-date=:


Which doesn't work, since {{citation/core}} does not support |Embargo=. What is this? ---— Gadget850 (Ed) talk 21:06, 14 December 2011 (UTC)

Added to core on 17 Oct 2009.[36] Removed on 6 March 2011. [37] I am going to clean up here and possibly other templates. ---— Gadget850 (Ed) talk 13:43, 15 December 2011 (UTC)
Should this be added to the list of deprecated parameters in the template (at least temporarily), so the citations using this parameter can be cleaned up?

Thanks! GoingBatty (talk) 00:33, 16 December 2011 (UTC)

{{Cite journal}} has markup supporting this. I did a search and found the paramter in one article where it was blank. ---— Gadget850 (Ed) talk 03:03, 16 December 2011 (UTC)
Sounds like useless cruft, including in a more generalized {{cite journal}} or {{cite news}} context; if the source it not yet available to the public, it fails WP:V, so it shouldn't be available in any of these templates. — SMcCandlish Talk⇒ ʕ(Õلō Contribs. 12:34, 16 December 2011 (UTC)
It's the full text that is not yet public, the citation and/or abstract will still be there to cover WP:V. Rjwilmsi 18:36, 4 January 2012 (UTC)
If not used, still cruft. It's also grossly misleading, and if it ever gets used, people will misuse it to cited press-embargoed stuff that cannot be cited per WP:V. If kept, it should be renamed something like |pubmed-release=. — SMcCandlish Talk⇒ ʕ(Õلō Contribs. 02:12, 5 January 2012 (UTC)
As noted, this parameter was removed from core; I removed it here since it did nothing. Only cite journal now supports this. -— Gadget850 (Ed) talk 10:11, 5 January 2012 (UTC)
That makes sense. Rjwilmsi 10:21, 5 January 2012 (UTC)

Where to put a "(150 MB)" warning note?

The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
Moved to Template talk:Citation/core#Where to put a download-size warning note?: Affects all citation templates, not just this one.

If I want to put a warning note like "(150 MB)" for a pdf file, where do I enter that? I'd expect it to be outside of and right after the title label (underlined text). -DePiep (talk) 11:46, 4 January 2012 (UTC)

You could use |format=. This will place it in parenthesis after the title. ---— Gadget850 (Ed) talk 14:20, 4 January 2012 (UTC)
But you'll risk it removed by one or more tools/scripts/bots as an invalid |format= field value. —  HELLKNOWZ  ▎TALK 14:25, 4 January 2012 (UTC)
What ia a valid format? What tools/scripts/bots check this? And on reflection, 150 MB is not really that big. ---— Gadget850 (Ed) talk 14:35, 4 January 2012 (UTC)
150mb is huge in respect toa pdf file and takes ages to load--Andrewcrawford (talk - contrib) 14:38, 4 January 2012 (UTC)
It's not a blanket decision. If I saw "150 Mb" in the format field, I'd probably remove it unless I took my time to analyze who added it and why assuming they left any note. I'm just saying with so much misuse of different fields, it might not be obvious to others and checking each edit is very very time consuming. "150 Mb" is certainly not an actual format value. I've seen AWB runs doing this among other things, not that I could recall now who it was or what they used. So perhaps a comment next to it would be useful. —  HELLKNOWZ  ▎TALK 14:46, 4 January 2012 (UTC)
Just checked, the PDF-icon is added irrespective of the content of |format=. -DePiep (talk) 14:49, 4 January 2012 (UTC)
Actually, for me now it could be OK since I want to create a templated reference. So no AWB visiting there easily, it is under /documentation, etc. The |format=150 Mb solution does give the result I prefer (it is where I got the idea from). -DePiep (talk) 15:07, 4 January 2012 (UTC)
The template does not generate the icon. See Help:External link icons. ---— Gadget850 (Ed) talk 15:35, 4 January 2012 (UTC)
Indeed, it is added. Still, when (mis)using the parameter format= this way, no other harm is possibly done except what is mentioned here? -DePiep (talk) 15:48, 4 January 2012 (UTC)
The icon Icons-mini-file acrobat.gif is generated by code within MediaWiki:Common.css and is largely dependent on the linked URL containing the characters ".pdf" (upper or lower case). --Redrose64 (talk) 18:17, 4 January 2012 (UTC)
AWB genfixes removes |format=HTML as this is unneeded and against {{cite web}} documentation. It also moves {{dead link}} outside the citation per Template:Dead link documentation. Otherwise no other changes. rev 7907 Additional unit test: |format= parameter not changed in citation template if contains (PDF) size information. Rjwilmsi 21:33, 4 January 2012 (UTC)

The above discussion is preserved as an archive. Please do not modify it. Subsequent comments should be made in a new section.


Please don't use <nowiki> or other parser tags in values. This causes the COinS metadata to expose the strip markers. This is hidden, but will show up in reference management systems. For example:

  • {{cite web|title=foo}} properly generates rft.btitle=foo
  • {{cite web|title=<nowiki>foo</nowiki>}} improperly generates title=%7FUNIQ1c9a3b4432e3e11e-nowiki-00000002-QINU%7F

— Preceding unsigned comment added by Gadget850 (talkcontribs) 16:36, January 10, 2012

Ouch! I have been using <nowiki> tags in titles a lot lately. Hopefully I will remember as I do not consider COinS to be important but I know that others do. – Allen4names 03:16, 11 January 2012 (UTC)
I know not everyone really cares about metadata, but it is used and this template supports it. As soon as I saw the doc change to use <nowiki>, I knew this was not going to be a good thing, but I was surprised to see the strip markers exposed.
I am going to look at the issue of encoding in the core, but will need to investigate how expensive it is. ---— Gadget850 (Ed) talk 12:16, 11 January 2012 (UTC)
It may be a good idea to have a bot perform replacements (ie. "["=>"&#x5B;", "]"=>"&#x5D;" and "|'=>"&#x7C;") in the title parameter and removing the <nowiki> tags as I am sure I am not the only one who has been making such edits. – Allen4names 17:40, 11 January 2012 (UTC)

Publisher full name

When listing the publisher, should they be listed under their full name, i.e. Rovi Corporation, or truncated to Rovi? Mac Dreamstate (talk) 00:24, 30 January 2012 (UTC)

It's better to give the full name in case of ambiguity. Can you be sure that there isn't an entirely unrelated (for example) Rovi Products, Inc someplace? The normal times that we shorten publishers names are to drop the "Inc", although with an outfit like Time Inc. we would show the "Inc", because "Time" is just too ambiguous. --Redrose64 (talk) 11:42, 30 January 2012 (UTC)
OK, that makes sense. But what about the way in which I did it above, where the shortened Rovi still links to Rovi Corporation's article? Does the full name always have to be externally conspicious, even though internally the link points to the correct article in question? Mac Dreamstate (talk) 16:19, 30 January 2012 (UTC)
It's OK. Is it essential that the reader have a link to the publisher? If not, then don't link it. Is there enough information to identify the source? ---— Gadget850 (Ed) talk 19:57, 30 January 2012 (UTC)
Well now I'm not sure.. From what you're saying, does that mean hardly anything at all (besides maybe a well-known author) should be Wiki-linked in the template? What about the 'work', if it has an article? I always thought everything should be piped to something if there's an article for it, for the sake of informativity. Whether or not it's essential to the reader never crossed my mind. Have I been going about this the wrong way? Mac Dreamstate (talk) 23:52, 30 January 2012 (UTC)
Never mind, I've read the discussion here and it answers all my questions. Therefore I'll use my own judgement when citing, since there seems to be enough freedom on Wikipedia to do so. Mac Dreamstate (talk) 06:03, 31 January 2012 (UTC)

Subscription required parameter?


may I ask why there is no parameter for sources which requires subscription, something like subscription=yes/no? It is especially useful for new users (or for lazy people like me...). Regards.--♫GoP♫TCN 18:02, 31 January 2012 (UTC)

See Help:Citation Style 1#Subscription or registration required. If you really want this added to the template, take it to {{citation/core}}. ---— Gadget850 (Ed) talk 18:15, 31 January 2012 (UTC)
Actually we don't know if this template meets WP:ACCESS.--♫GoP♫TCN 09:30, 4 February 2012 (UTC)
Which template? How would this affect accessibility? ---— Gadget850 (Ed) talk 18:53, 4 February 2012 (UTC)
The one which looks like your signature :)--♫GoP♫TCN 11:56, 6 February 2012 (UTC)


If the page title contains words in all caps, should they be left that way when citing? For some reason I've always changed them to lower/title case to make it look less like shouting on the refs list, but there doesn't seem to be any guidelines on that. Mac Dreamstate (talk) 15:41, 14 February 2012 (UTC)

You are correct to do so in general; see MOS:ALLCAPS. You may find http://titlecase.com/ handy. ---— Gadget850 (Ed) talk 16:15, 14 February 2012 (UTC)
They should definitely be changed to title case. — SMcCandlish   Talk⇒〈°⌊°〉 Contribs. 07:46, 25 February 2012 (UTC)

Minor fixes

Propose to fix some minor issues:

  • Add nopp to match other Citation Style 1 templates (yes, page numbers are not often used with cite web, but they are used)
  • Remove deprecated parameters from AccessDate (the template already detects these)
  • Remove unsupported DateFormat
  • Add AuthorSep, NameSep, Trunc and amp to match other Citation Style 1 templates
  • Add laysource and supporting parameters
  • Add others

---— Gadget850 (Ed) talk 21:38, 15 February 2012 (UTC)

Now in sandbox. ---— Gadget850 (Ed) talk 09:35, 25 February 2012 (UTC)
  • Support: Maximum consistency between these templates would be a good thing. — SMcCandlish   Talk⇒〈°⌊°〉 Contribs. 07:42, 25 February 2012 (UTC)

Yes check.svg Done ---— Gadget850 (Ed) talk 11:41, 26 February 2012 (UTC)

Is the "others" field working?

I have just used it in the Mona Bone Jakon article. Neither syntax –

others=trans. Michael Valenzuela
others="trans. Michael Valenzuela"

– produces any output. Varlaam (talk) 03:51, 25 February 2012 (UTC)

I don't see it when I view the template code, and it's not included in the list of all parameters, so I removed it from the documentation. GoingBatty (talk) 04:43, 25 February 2012 (UTC)
It works in {{citation}} but not {{cite web}}. It switched the article, but absent an author, it still looks pretty ich. LeadSongDog come howl! 05:16, 25 February 2012 (UTC)
Yeah, in the interim just use |last2=Valenzuela|first2=Michael (trans.) to add someone like that. — SMcCandlish   Talk⇒〈°⌊°〉 Contribs. 07:38, 25 February 2012 (UTC)
I added it to the sandbox version, with the other updates listed above. ---— Gadget850 (Ed) talk 09:36, 25 February 2012 (UTC)
Is it in {{citation core}}?
{{tl|citation/core supports |Other=. Per above, |others= in now in {{cite web/sandbox}}. ---— Gadget850 (Ed) talk 02:59, 26 February 2012 (UTC)

Fixed ---— Gadget850 (Ed) talk 11:38, 26 February 2012 (UTC)

Where should |others= be placed in the list of all parameters in the documentation? GoingBatty (talk) 21:10, 26 February 2012 (UTC)

deadurl should default to no

Is there a good reason why the "deadurl" parameter defaults to "yes"? Tim1357 talk 01:12, 26 February 2012 (UTC)

|deadurl= doesn't default to anything. When set to no, then it swaps the order of the original and archive URL links. See Help:Citation Style 1#Web archives. I agree that the implementation is a bit confusing. ---— Gadget850 (Ed) talk 01:28, 26 February 2012 (UTC)
When there is no |deadurl= paramter, the template puts the archive first. Shouldn't it be the other way around? Tim1357 talk 02:17, 26 February 2012 (UTC)
Backwards compatibility mostly. If it would default to "no", it would invalidate a whole lot of existing citations (100k+ last I checked). You'd need to put a |deadurl=yes to all of them. Editors who put archive params wouldn't necessarily know this even exits. For the sake of implementation it was easier to get consensus to make changes to the citation templates as opposed to citation templates and 100k+ articles. It could default to "no" subject to updating all existing citations (but not preemptively archived ones) and all tools being aware they need to us it. —  HELLKNOWZ  ▎TALK 09:23, 26 February 2012 (UTC)
The default is that deadurl does nothing, so the templates render just as they did before deadurl was added. ---— Gadget850 (Ed) talk 10:56, 26 February 2012 (UTC)
Yeah, I know, I proposed that, I was replying to Tim. I meant "defaults" as a simplified concept of "it happens, because it is not coded not to happen", not the specific implementation. Also I had a "yes/no" typo there and that may have been the issue? Sorry. —  HELLKNOWZ  ▎TALK 12:11, 26 February 2012 (UTC)
Perhaps the template can be changed so that if archiveurl is set but deadurl is not it will add the page it is on to a maintenance category. For example Category:Pages with archive links to be checked. – Allen4names 18:41, 26 February 2012 (UTC)
It would make sense. Bots can check if the link is dead pretty well. Bots cannot really check if the link is live. —  HELLKNOWZ  ▎TALK 18:45, 26 February 2012 (UTC)
What is the intent of deadurl? As best I see, it simply swaps the original and archived links. ---— Gadget850 (Ed) talk 20:04, 26 February 2012 (UTC)
Yes, that's it really. It's meant to be used when the links aren't dead. Linkrot's has been getting pretty bad and many editors have started archiving preemptively, mainly in FAs. It's a much bigger pain to restore (and detect) links when they die. It's very easy with preemptive archiving (just swap to |deadurl=yes). Without that, the archived url appears as the main url. It's slow and sometimes unreliable service, and strains their servers unnecessarily. Original RfC: Wikipedia:Requests for comment/Dead url parameter for citations; some info on linkrot: Wikipedia:WikiProject External links/Webcitebot2. —  HELLKNOWZ  ▎TALK 20:16, 26 February 2012 (UTC)
But |deadurl=yes doesn't do anything different from the original action of the template; unless a bot is supposed to look at it? ---— Gadget850 (Ed) talk 20:42, 26 February 2012 (UTC)
Exactly. Tools and editors know the link is dead and not "may be dead, may be preemptively archived". Checking links is expensive (time and bandwidth) and it's a blessing if a bot can skip links. There was also idea of putting {{dead link}} next to the url when |deadurl=yes is set but the params aren't. —  HELLKNOWZ  ▎TALK 21:12, 26 February 2012 (UTC)
OK. Any bots actully doing that? And it looks like markup in {{citation/core}} is supposed to do some error checking if deadurl is set without other parameters, but I can't invoke the error and have not dug into the markup yet. ---— Gadget850 (Ed) talk 21:36, 26 February 2012 (UTC)
I think there are bots (or something); this self-rv of mine put that page into User:West.andrew.g/Dead links (it's now in User:West.andrew.g/Dead links/Archive 439). --Redrose64 (talk) 22:28, 26 February 2012 (UTC)
User:WebCiteBOT uses |deadurl=no. User:DASHBot may be checking for it. I don't know about Checklinks but I remember mentioning it. Currently all other bots I know are down. —  HELLKNOWZ  ▎TALK 09:58, 27 February 2012 (UTC)


I have several BOT generated archiveurls in an article I'm updating. When I click on the hotlink in the reference to go to the web page, I get some "Wayback" page, but it never takes me to the archived page. My questions is will this pose a problem since the article in question is a featured article. Do I need to update the cite web reference within the article in any way? thanksCcson (talk) 06:18, 27 February 2012 (UTC)

Sometimes Wayback will display a generic page for a period, whilst it reassembles the actual page in which you're interested. This can take some time. How long have you let it process the page for? --Redrose64 (talk) 09:44, 27 February 2012 (UTC)
I left the house and returned ata least 1 hour later and it was still on the Wayback Page. There was an "Impatient?" link at the bottom right corner of the page which would display the archived page if I click on impatient, but other than that, no page was presented.Ccson (talk) 17:31, 27 February 2012 (UTC)
What article and references? ---— Gadget850 (Ed) talk 09:45, 27 February 2012 (UTC)
See List of Alpha Phi Alpha brothers, reference number "14", Alpha Phi Alpha Educators when I waited at least 1 hour for the page to display. There are nineteen more (i.e. 1, 20, 47, 63, 119, etc.) where a BOT has updated references with the archived page. I searched within the page for "the original" to see which pages the BOT had udpated.Ccson (talk) 17:31, 27 February 2012 (UTC)
Those archived links work for me; for example: Reference 14 ---— Gadget850 (Ed) talk 17:43, 27 February 2012 (UTC)
I tried using the link you supplied and still no page except the wayback page. Perhaps my browser doesn't correctly redirect ans the user below hast stated. Is there a troubleshooting page to indicate how the browser settings should be set for wayback to work properly.Ccson (talk) 17:48, 27 February 2012 (UTC)
Try another browser or PC. Reboot your modem and router. ---— Gadget850 (Ed) talk 19:23, 27 February 2012 (UTC)
I will try both. thanksCcson (talk) 23:14, 27 February 2012 (UTC)
I tried rebooting and using the computer at the office, still no page is displayed.Ccson (talk) 23:59, 28 February 2012 (UTC)
My first impression without details is that your browser doesn't correctly redirect on the Wayback page. —  HELLKNOWZ  ▎TALK 09:59, 27 February 2012 (UTC)
thanksCcson (talk) 17:31, 27 February 2012 (UTC)
Just a thought, do you have javascript blocked? That page uses it. LeadSongDog come howl! 04:44, 29 February 2012 (UTC)

If I needed to include a path?

I recently noticed that a user added a ref to a page on my watch list. Upon checking out the link, it appeared that the source was not there. Upon further digging, I found that the source was there, but it did not have a unique URL. Therefore, the only way for a user to be able to find the information is through a path to it. What parameter would be best to use to include a path? - Purplewowies (talk) 16:28, 16 March 2012 (UTC)

You can use |at=. Could you provide the example? Perhaps we can tweak the documentation. ---— Gadget850 (Ed) talk 16:53, 16 March 2012 (UTC)
The ref in question is #11 at PrankStars. The URL in question is http://www.barb.co.uk/report/weekly-top-programmes . The page that links to is more or less a search page (with no visible way to link to the results). Selecting the search parameters "Disney Channel" (for the channel) and the date 2011 November 21-27 shows what the reference was supposed to link to. I'm not sure how exactly I'd provide a path like that within the ref. - Purplewowies (talk) 18:12, 16 March 2012 (UTC)