Archives

March 2006 – June 2006 June 2006 – December 2006 December 2006 – August 2007 September 2007 – July 2008 August 2008 – September 2008

The use of "Retrieved on" has puzzled me for a long time. It has a suggestion that a snapshot of the referenced page was taken on the date specified and that the link will take you to some wayback machine type archived view of the page. Upon viewing the template, I see that the parameter is "Accessdate". Why not use that: "Accessed on" or "Information gathered on" but NOT "Retrieved on" which is just plain misleading. -- SGBailey (talk) 06:52, 21 December 2008 (UTC)[reply]

That's kind of true. It would be a big change, though. - Peregrine Fisher (talk) (contribs) 08:21, 21 December 2008 (UTC)[reply]

I assume he means just changing the template itself to say "accessed on April 1 2008" rather than "retrieved on April 1 2008", which would actually be a very minimal change. I don't really have a strong view on whether or not it's a good idea. Happy‑melon 15:54, 21 December 2008 (UTC)[reply]

Since it hasn't been mentioned—this would need to be changed in all cite templates that use the accessdate parameter. You'll need to post notices on a few template talk pages. As more templates start using {{Citation/core}} I suspect this will become easier. Pagrashtak 07:14, 22 December 2008 (UTC)[reply]

I believe that we are using the APA style and that uses Retrieved. For a variety of styles, you can have a look at Wikisource cite. The place for the citation preference discussion is Wikipedia:Citing sources. -- billinghurst (talk) 08:26, 22 December 2008 (UTC)[reply]

That's interesting. The APA "Retrieved on" makes sense in an actual paper, which just lists a URL that you can't click. It's stranger here, where we hide the URL under a clickable title. - Peregrine Fisher (talk) (contribs) 22:15, 24 December 2008 (UTC)[reply]

Hi. Suddenly, at least in my browser (Safari), {{cite web}} hyperlinks no longer seem to display. For example in my browser, I do not see the hyperlink:

• "Casodex product insert". www1.astrazeneca-us.com. 2006-03-01. Retrieved 2008-12-27. {{cite web}}: Cite has empty unknown parameter: |coauthors= (help)

produced by:

• {{cite web | url = http://www1.astrazeneca-us.com/pi/casodex.pdf/ | format = PDF | title = Casodex product insert | author = | authorlink = | coauthors = | date = 2006-03-01 | format = | work = | publisher = www1.astrazeneca-us.com | pages = | language = | archiveurl = | archivedate = | quote = | accessdate = 2008-12-27}}

This used to work, but several days ago it stopped working. Has anyone else noticed this behavior? Cheers. Boghog2 (talk) 21:47, 27 December 2008 (UTC)[reply]

Also in FF3. Will investigate... Any ideas when this started? Happy‑melon 21:50, 27 December 2008 (UTC)[reply]

Fixed, but this highlights a whole new set of problems. The template code wants to accomodate every man and his dog when it comes to alternative parameter names. It's just not feasible to do so when there are so many of them. Pick one syntax, hunt down and eliminate the rest, and then remove them from the template code. Otherwise we have to use a mass of #if: statements to choose between them, becuase we can't use parameter defaults as you've just realised. Happy‑melon 10:49, 28 December 2008 (UTC)[reply]

 Done Much better. Thanks for your quick response. I apologize if I have opened a can of worms. Boghog2 (talk) 22:04, 27 December 2008 (UTC)[reply]

Could you please add hu:Sablon:Cite web to the article. Thanks Karmela (talk) 00:16, 29 December 2008 (UTC)[reply]

 Done, but remember you can add interwiki links to almost any template which uses {{documentation}}. The /doc subpages are almost never protected. — Huntster (t • @ • c) 00:42, 29 December 2008 (UTC)[reply]

The parameters accessdaymonth, accessmonthday, and accessyear seem to no longer work.

This code:

• {{cite web | title = Test page | url = http://www.testpage.co.org/ | date = 12 March 2008 | accessdaymonth = 28 December | accessyear = 2008 }}

should produce this:

• "Test page". 12 March 2008. Retrieved on 28 December 2008.

but is currently producing this:

• "Test page". 12 March 2008.
• "Test page". 12 March 2008. {{cite web}}: Unknown parameter |accessdaymonth= ignored (help); Unknown parameter |accessyear= ignored (|access-date= suggested) (help)

Was there any discussion before the removal of these parameters, or was this an oversight? When can this be fixed? — Bellhalla (talk) 03:32, 29 December 2008 (UTC)[reply]

{{editprotected}}

Please revert the template to the 9 November 2008 version. The changes made on 24 December 2008 have broken the alternative access date parameters above, leaving many instances of this template without a valid retrieval date. — Bellhalla (talk) 14:07, 29 December 2008 (UTC)[reply]

 Not done:, but I have  Fixed the template instead. Happy‑melon 14:06, 30 December 2008 (UTC)[reply]

Thanks for getting the retrieval dates back, but why the change in format? Now the the template generates a series of items separated by commas, rather than full stops as was previously the case and is the case with {{cite book}} and {{cite news}} — Bellhalla (talk) 16:05, 30 December 2008 (UTC)[reply]

That change was per the discussion above, not something I've had any involvement with. Happy‑melon 16:20, 30 December 2008 (UTC)[reply]

Actually, no. The change from periods to commas was because you removed the part setting Sep completely instead of changing "|Sep = {{{separator|{{{seperator|.}}}}}}" to "|Sep = .". Anomie⚔ 18:24, 30 December 2008 (UTC)[reply]

As I said above, this template currently incorporates a large number of alternative parameter names for the same value. This is, not to put too fine a point on it, untenable. We need to settle on one allowed name for each parameter and stop supporting the others. Note that I'm not talking about things such as |author= vs |first= and |last=, but things like |first= vs |given=; two parameters that do exactly the same thing but are just synonyms. There is no reason to support things like these. The list of duplicated parameters are below: a lot of them have been very recently introduced and IMO should be just as quickly removed before they gain any popularity:

• |Surname1 = {{{last|{{{surname|{{{last1|{{{surname1|{{{author1|{{{author|{{{authors|{{{author|}}}}}}}}}}}}}}}}}}}}}}}}
• |Surname2 = {{{last2|{{{surname2|{{{author2|{{{coauthor|{{{coauthors|}}}}}}}}}}}}}}}
• |Surname3 = {{{last3|{{{surname3|{{{author3|}}}}}}}}}
• |Surname4 = {{{last4|{{{surname4|{{{author4|}}}}}}}}}
• |Surname5 = {{{last5|{{{surname5|{{{author5|}}}}}}}}}
• |Surname6 = {{{last6|{{{surname6|{{{author6|}}}}}}}}}
• |Surname7 = {{{last7|{{{surname7|{{{author7|}}}}}}}}}
• |Surname8 = {{{last8|{{{surname8|{{{author8|}}}}}}}}}
• |Surname9 = {{{last9|{{{surname9|{{{author9|}}}}}}}}}
• |Given1 = {{{first1|{{{given1|{{{first|{{{given|}}}}}}}}}}}}
• |Given2 = {{{first2|{{{given2|}}}}}}
• |Given3 = {{{first3|{{{given3|}}}}}}
• |Given4 = {{{first4|{{{given4|}}}}}}
• |Given5 = {{{first5|{{{given5|}}}}}}
• |Given6 = {{{first6|{{{given6|}}}}}}
• |Given7 = {{{first7|{{{given7|}}}}}}
• |Given8 = {{{first8|{{{given8|}}}}}}
• |Given9 = {{{first9|{{{given9|}}}}}}
• |Authorlink1 = {{{author-link|{{{author1-link|{{{authorlink|{{{authorlink1|}}}}}}}}}}}}
• |Authorlink2 = {{{author2-link|{{{authorlink2|}}}}}}
• |Authorlink3 = {{{author3-link|{{{authorlink3|}}}}}}
• |Authorlink4 = {{{author4-link|{{{authorlink4|}}}}}}
• |Authorlink5 = {{{author5-link|{{{authorlink5|}}}}}}
• |Authorlink6 = {{{author6-link|{{{authorlink6|}}}}}}
• |Authorlink7 = {{{author7-link|{{{authorlink7|}}}}}}
• |Authorlink8 = {{{author8-link|{{{authorlink8|}}}}}}
• |Authorlink9 = {{{author9-link|{{{authorlink9|}}}}}}
• {{{date|}}}{{{publication-date|einval}}}
• {{{journal|{{{periodical|{{{newspaper|{{{magazine|}}}}}}}}}}}}
• {{{pages|{{{page|{{{at|}}}}}}}}}
• {{{title|{{{contribution|}}}}}}
• |Place = {{{place|{{{location|}}}}}}
• |PublicationPlace = {{{publication-place|{{{place|{{{location|}}}}}}}}}
• |language = {{{language|{{{in|}}}}}}
• |AccessDate={{#if:{{{access-date|}}}|{{{accessdate|}}}|{{{accessdate|}}} }}
• |Sep = {{{separator|{{{seperator|.}}}}}}

As you can see there are a lot of them, most of which have been unilaterally imported from Citation/core without considering whether they are useful to cite web. Most of them, I believe, are not. Which ones should be retained and which ones removed? Happy‑melon 11:30, 29 December 2008 (UTC)[reply]

Well {{{journal|{{{periodical|{{{newspaper|{{{magazine|}}}}}}}}}}}} can come out LS&B for a start. Are the nonstandard AccessDate, Place, and PublicationPlace used anywhere, or are they cruft? Mr Stephen (talk)

AccessDate is, I think, what comes out as "retrieved on XYZ", so that's useful. Unless they want us to find out where the site's servers are based, I can't think of any possible use for the other two. Happy‑melon 12:36, 29 December 2008 (UTC)[reply]

I see they are internal variables, rather than supported parameters, so skip my comment about AccessDate etc. Mr Stephen (talk) 14:40, 29 December 2008 (UTC)[reply]

I've removed essentially all the parameters that were not present before the conversion to the meta-template; which pretty much resolves the problems. Happy‑melon 14:12, 30 December 2008 (UTC)[reply]

You also removed coauthors, doilabel, and editor that were present before the conversion. Also, FWIW, you left in author[2-9], authorlink[2-9], dateformat, day, first[2-9], and last[2-9]. Anomie⚔ 18:32, 30 December 2008 (UTC)[reply]

You've not outlined what damage these extra parameters cause. The benefit is that giving editors a range of parameters makes the template easier to use, and most of these 'surplus' parameters are common to other citation templates. Removing parameters without thorough testing does seem a bit premature, especially as many are already present in articles. Martin (Smith609 – Talk) 20:10, 3 January 2009 (UTC)[reply]

I don't know what changes were made recently, but the coauthors tag seems to have been broken in the process?! -- Avi (talk) 16:39, 30 December 2008 (UTC)[reply]

Happy-melon removed it in this series of edits, probably because he didn't realize it wasn't one of the many "extra" parameters added in the recent edits to this template. Anomie⚔ 18:27, 30 December 2008 (UTC)[reply]

I've restored the template to the 12/27 version, that should fix the issues, and then any pruning should be done very carefully in the future. -- Avi (talk) 18:48, 30 December 2008 (UTC)[reply]

There used to be brackets around the date in cite web Google cached from 21 Dec. They have now gone. Why? I see no consensus for this change. Can it be changed back please, Rambo's Revenge (talk) 20:15, 30 December 2008 (UTC)[reply]

Consensus was obtained here, with the argument being that cite web's output should be consistent with Wikipedia's other citation templates. Martin (Smith609 – Talk) 20:19, 3 January 2009 (UTC)[reply]

{{editprotected}} Please revert this template to the stable 9 November version. Then you guys can play around with the template sandbox and testcases page before implementing the changes. This is a highly transcluded template that shouldn't be ping-pong-ing back and forth, as it is disruptive to many articles. This reversion has once again broken the alternate access date parameters accessdaymonth, accessmonthday, and accessyear. — Bellhalla (talk) 20:22, 30 December 2008 (UTC)[reply]

Looking at all the recent mess, I support this proposal. Rambo's Revenge (talk) 20:40, 30 December 2008 (UTC)[reply]

 Done -- Avi (talk) 20:43, 30 December 2008 (UTC)[reply]

Are the accessdate parameters the only things that are broken in the 27th Dec version? Please list any other errors here so I can fix them in one go. Martin ' 18:41, 3 January 2009 (UTC)[reply]

That's all I'm aware of but there may be other issues. May I please repeat the suggestion above of using the sandbox and testcases pages? Thanks in advance. — Bellhalla (talk) 19:23, 3 January 2009 (UTC)[reply]

I agree that the sandbox and testcases should be rigorously checked before any edit is enacted. Unfortunately the accessdate parameters slipped through my original testing of the stable 24th Dec version. I've now corrected this and would be grateful if any further editors would be willing to find further problems with the new template. Here are the testcases and here is the sandbox template. Thanks, Martin ' 19:43, 3 January 2009 (UTC)[reply]

The sandbox still contains all the problems highlighted in the #Consolidating parameters and #URL hyperlink not displaying sections above. ‑ 19:46, 3 January 2009 (UTC)

The URL problem is now resolved in the sandbox, and I've reinstated a few of your tweaks along the way. I'm reluctant to prune too many of the standard parameters, as these often end up being useful in rare cases. Can you convince me that they cause any harm worth worrying about? Martin (Smith609 – Talk) 20:15, 3 January 2009 (UTC)[reply]

As explained above, using parameter defaults is not acceptable in this situation. So all constructs like {{{last2|{{{surname2|{{{author2|{{{coauthor|{{{coauthors|}}}}}}}}}}}}}}} need to be replaced by a construct like {{#if:{{{last2|}}}|{{{last2}}}|{{#if:{{{surname2|}}}|{{{surname2}}}|{{#if:{{{author2|}}}|{{{author2}}}|{{#if:{{{coauthor|}}}|{{{coauthor}}}|{{{coauthors|}}} }}}}}}}}, an incredibly inefficient and messy parser construct that will introduce a huge increase in ParserFunction counts for absolutely no benefit whatsoever. Why do we need five alternative syntaxes for saying the same thing? Find the original parameter names that were used to define these values (ie what the templates 'in the wild' are actually using), and remove the unused and unnecessary alternatives. Happy‑melon 10:39, 8 January 2009 (UTC)[reply]

Why is accessdate still broken after so many weeks? If there's no consensus for changing the formatting, simply revert back the formatting that worked before instead of breaking it. Madcoverboy (talk) 15:49, 7 January 2009 (UTC)[reply]

{{editprotected}}

I've removed redundant parameters (= those not documented in the documentation) from the current sandbox to address HappyMelon's concerns. The constructs Happymelon suggested are used where necessary; in some cases they aren't (because they are effectively employed in the citation/core template instead). I've put a couple of Template:Cite web/testcases testcases up with blank parameters and they work fine. As no other bugs have been spotted, I suggest that we go live with the current sandbox by copying it to Template:Cite web. Thanks, Martin (Smith609 – Talk) 23:31, 10 January 2009 (UTC)[reply]

Looking at test cases, the last parameter doesn't appear functional. Disabling the edit request for now while this is sorted out. Pagrashtak 19:16, 14 January 2009 (UTC)[reply]

Numerous parameters (Surname1, Surname2, Authorlink1, Year, Place, DOI, At, Sep) still use parameter default constructions. More work needed. Happy‑melon 19:43, 14 January 2009 (UTC)[reply]

Good point. I see that the proposed version would add support for last1 in the Surname1 parameter, for example—why do we need to introduce this if the current version doesn't use it? More questions: Why does the sandbox reference {{Citation/core/sandbox}} instead of {{Citation/core}}? Would it be possible to give us a list of what parameters would be added/removed with this change? For example, last1, last2, last3, etc. are being added, while editor is being removed. Pagrashtak 20:13, 14 January 2009 (UTC)[reply]

Happymelon, the proposed version uses parameter default constructions in a perfectly adequate and functional way. Please give an example of how this could cause a problem. Pagra, I'm probably being blind, but I couldn't see where the |last= parameter didn't work. Could you give an example? The sandbox links to the sandbox citation core because it's a sandbox version of the template. Parameters are detailed below. Martin (Smith609 – Talk) 23:51, 19 January 2009 (UTC)[reply]

• Editor: There is currently no editor parameter documented, which makes sense because I have never seen a web page which cites an editor. Therefore by HappyMelon's logic I didn't add a parameter which was unused.
• Last1/last2/etc. : The 'last' parameter only works for web pages with a single author. It seems logical to extend it.

{{editprotected}}

The queries above seem to have been addressed, all Cite web/testcases are working, and I cannot produce an example where the new template will cause incorrect output. I'd be grateful if an admin would replace Template:Cite web with the current sandbox. When copying the code, please remove '/sandbox' from the first line of the template. Thanks, Martin (Smith609 – Talk) 14:59, 23 January 2009 (UTC)[reply]

Is there a reason you didn't remove it now? Pagrashtak 15:33, 23 January 2009 (UTC)[reply]

Edit request disabled, the test cases are not working as you say. Smith, I appreciate the work you do with these templates, but you need to slow down—you tend to rush and let things fall through the cracks. Pagrashtak 15:38, 23 January 2009 (UTC)[reply]

Pagra, it would be very helpful indeed if you were to let me know which test cases aren't working, and what the problem is. I've looked again and still can't see the problems you're alluding to. Could you please be specific? Martin (Smith609 – Talk) 23:39, 23 January 2009 (UTC)[reply]

Actually, I didn't mention the specific problem because I wanted you to take a look at all the test cases instead of focusing on the parameter I had in mind. It appears this wasn't enough to draw attention to it, so I'll be more specific below. Pagrashtak 22:37, 25 January 2009 (UTC)[reply]

Indeed. In fact, there has been no attempt to address my concern:

  |Surname1 = {{{last|{{{last1|{{{author|}}}}}}}}}
  |Surname2 = {{{last2|{{{coauthors|}}}}}}

That's a parameter default.

  |Authorlink1 = {{{authorlink|{{{authorlink1|}}}}}}

So is that.

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

This is a particularly nasty one: all that carefully-constructed parser coding will be totally ignored if someone calls the template like {{cite web|url=http://www.example.org|date=January 23 2008|year=|month=|day=|...}}. How many times do I have to say it? You Can Not Do That. Pick one parameter to pass in each situation (either by only accepting one parameter in the first place, or by fiddling through them with parserfunctions) and pass only that parameter. Why do you continue to insist on using numerous different synonyms for the same thing? Happy‑melon 16:18, 23 January 2009 (UTC)[reply]

Hi, the example you cited seems to work to me:

• {{cite web/sandbox|url=http://www.example.org|date=January 23 2008|year=|month=|day=||title=title...}}
• "title..." January 23 2008. {{cite web}}: Check date values in: |date= (help); Cite has empty unknown parameters: |1=, |month=, and |day= (help)
• {{cite web/sandbox|url=http://www.example.org|date=|year=2008|month=January|day=23|title=title...}}
• "title..." 2008. {{cite web}}: Unknown parameter |day= ignored (help); Unknown parameter |month= ignored (help)

It's much more useful if you can give examples of where the parameter defaults don't work, rather than just stating that they don't. You may want to look at Template:Cite journal, Template:Cite book and Template:Citation, which all employ parameter defaults without a problem. Martin (Smith609 – Talk) 23:39, 23 January 2009 (UTC)[reply]

You're right, thanks to the way the /core template handles the date input. This will still fail though:

{{cite web|url=foo|title=title...|year=|month=|day=|publication-date=January 24 2009}}

"title..." January 24 2009. {{cite web}}: Check date values in: |publication-date= (help)

Now consider the plausible situation of someone wanting to split a list of coauthors from the |coauthors= parameter to a list of |last#= parameters:

{{cite web|last=Smith|first=John|url=http://www.foo.com|title=title...|coauthors=Jane Smith, Fred Bloggs, George Bush}}
→ {{cite web|last=|url=http://www.foo.com|title=title... |last1=Smith|first1=John|last2=Smith|first2=Jane|first3=Fred|last3=Bloggs|first4=George|last4=Bush}}

Smith, John. "title..." {{cite web}}: Unknown parameter |coauthors= ignored (|author= suggested) (help)

Smith, John; Smith, Jane; Bloggs, Fred; Bush, George. "title..."

Bang, there goes Mr Smith. In fact, what happened to the rest of the authors? Something's not right there. Happy‑melon 00:23, 24 January 2009 (UTC)[reply]

Okay, thanks for the examples. I'll get onto it. Martin (Smith609 – Talk) 01:12, 24 January 2009 (UTC)[reply]

I've removed all of the parameter defaults mentioned by HappyMelon, with the exception of misspellings (separator/seperator; doi/DOI), because it's highly unlikely that these will both be present with different values in the same citation. (Someone with a sharper eye than me might want to make sure I've not missed any.) Are there any other things that need to be addressed before we go live? Martin (Smith609 – Talk) 01:24, 25 January 2009 (UTC)[reply]

Still not ready—you appear to making DOI a required parameter when it should stay optional and you are removing support for the first parameter. Pagrashtak 22:37, 25 January 2009 (UTC)[reply]

 Fixed but not tested. Anything else? Martin (Smith609 – Talk) 03:46, 26 January 2009 (UTC)[reply]

That looks much better from my perspective. Thankyou for your patience, Martin. Happy‑melon 08:42, 26 January 2009 (UTC)[reply]

Good. Thank you for checking my code so vigilantly! The test cases are all looking happy, and I don't think there are any outstanding concerns. Would anyone object to me requesting an edit? Martin (Smith609 – Talk) 14:35, 28 January 2009 (UTC)[reply]

No need.  Done. Let's see how this version goes down :D Happy‑melon 22:42, 28 January 2009 (UTC)[reply]

You broke support for accessmonthday/accessdaymonth/accessyear. Pagrashtak 15:04, 29 January 2009 (UTC)[reply]

What is the background behind these parameters? They seem rather redundant to me. Happy‑melon 16:20, 29 January 2009 (UTC)[reply]

When it was more-or-less mandatory to specify accessdate in YYYY-MM-DD format and the accessdate was automatically linked, the accessmonthday (or accessdaymonth) and accessyear provided a way to provide a spelled-out date of access. During the time these parameters were available, they were added to some unknown number of articles; to remove them now will break citations that were made in good faith. --Gerry Ashton (talk) 17:53, 29 January 2009 (UTC)[reply]

Seems like this category Happy-Melon just added will tell us in the near future which articles use the deprecated accessmonthday/accessdaymonth/accessyear fields. After that it's just a matter of going through the articles in that category and moving the values of those fields into the accessdate field. Depending on how many there ends up being, I wouldn't think it would take that long to do, even if we were to do it manually. It only took me half a minute or so to replace the field on 2008-2009 Canadian parliamentary dispute.[1] --Bobblehead ^(rants) 20:58, 29 January 2009 (UTC)[reply]

This could be done with a bot. Is it reasonable to conclude that a page which uses |accessmonthday= and |accessyear= should display dates in the "mdy" format, and equivalently for |accessdaymonth=? If so, the bot could combine the accessdate parameters into |accessdate=, combine the |day=, |month= |year= parameters into |date=, and set the appropriate |dateformat= to display them both correctly. If there's over a threshold number of cites using this format on the page, perhaps it should set the dateformat on all citation templates. Happy‑melon 08:47, 30 January 2009 (UTC)[reply]

That seems reasonable. Can a bot be programmed to count the different date formats and go with the majority? --Gerry Ashton (talk) 20:54, 30 January 2009 (UTC)[reply]

I've noticed that the "date" section does not automatically turn 2009-01-03 into January 3, 2009. The "accessdate" does do this, but not the basic "date" (for publication date). I'd fix this if I knew how, but I think it needs to be done to be consistent with the other dating styles that are present.  BIGNOLE  (Contact me) 13:34, 3 January 2009 (UTC)[reply]

Neither date nor accessdate currently changes the format of the date entered: "Example". 2008-01-01. Retrieved 2008-02-02. For a short time recently, attempted cleanup by Happy-Melon had broken things to force DMY ordering, but that has been reverted now.

At the moment, we have no acceptable way to do so, as forcing either DMY or MDY format for all uses would conflict with the date format used in the prose of many articles, link-based date formatting has been rejected by the community, any non-link-based replacement is still trying to gain consensus (see WT:MOSNUM), and forcing every use of the template to supply a "dateformat" parameter would be decidedly sub-optimal. Anomie⚔ 16:05, 3 January 2009 (UTC)[reply]

Wasn't me guv! I was just trying to fix someone else's broken code, although I admit I did make things worse... Happy‑melon 16:43, 3 January 2009 (UTC)[reply]

Putting it all in "long-hand" format is excessive waste of coding space for the article when we can automatically make it change to that style in the template. Otherwise, we now have hundreds (probably thousands) of articles that have never been adjusted since the date linking issue first came up and was removed. Thus, all those article's references that use the template read, "YYYY-MM-DD"...which is kind of hard to read on the fly, and looks grotesque on the screen.  BIGNOLE  (Contact me) 17:45, 8 January 2009 (UTC)[reply]

This went unnoticed (at by myself) until I was challenged on why I was using YYYY-mm-dd dates in cite templates. From hunting, I notice that this undiscussed edit[2] from 2008-11 removed all of the previous examples and explicit advice to use YYYY-mm-dd. The editor in question appears to have made no previous contributions to this template or its documentation before, nor further edits immediately after this point. As far as I can tell, it appears to have been done unilaterally.

Could this please be reverted so that we don't end up with a complete mess of non-machine-parsable dates in the references section. {{cite/doc}} is still very explicit about the use of YYYY-mm-dd dates in connection with cite templates. —Sladen (talk) 18:09, 14 January 2009 (UTC)[reply]

Agreed, this should be reverted since the template itself still looks for and wants the ISO format. And from all discussions I've read so far, even when the code does go back to fixing the format, it will still prefer ISO. -- Collectonian (talk · contribs) 18:50, 14 January 2009 (UTC)[reply]

{{editprotected}}. Okay lets get it fixed then (nb. to admin: this is for /doc). —Sladen (talk) 19:13, 14 January 2009 (UTC)[reply]

Disabling editprotected, as Template:Cite web/doc is not a protected page. Pagrashtak 20:27, 14 January 2009 (UTC)[reply]

 DoneI tried[3] to merge the obvious changes; however in the mean-time, the examples had also been meddled with, changing them from highly-diffable multi-line format, to massive single-line blocks (which aren't much good as examples anyway!). Please review and add back another further that I missed in the revert cleanup. —Sladen (talk) 03:38, 15 January 2009 (UTC)[reply]

I think you've gone a little too far back. There are now references to "Using non-linked retrieved date" examples, which are from the old version of the template when linking was the default. Pagrashtak 17:39, 16 January 2009 (UTC)[reply]

Related side question... was the date linking and formatting turfed?

The docs ref that the date should self-generate a link and I could have sworn that the "yyyy-mm-dd" numbers would convert to "Month dd, yyyy" depending on where in the article the template was placed.

- J Greb (talk) 16:41, 17 January 2009 (UTC)[reply]

When a web site disappears, and a reference is changed to refer to an archived copy elsewhere, there's no point in having the generated code contain a link to the now vanished original web site. For example, at present, input like this:

{{cite web
|title=Example
|url=http://example.org/example.html
|archiveurl=http://web.archive.example/yyyymmddhhmmss/http://example.org/example.html
|archivedate=YYYY-MM-DD
}}

generates output that looks rather like this:

"Example". Archived from the original on YYYY-MM-DD.

That link to "the original" is pointless if the original URL is inaccessible (due to a web site being reorganised, content being deleted, or a company going out of business). I'd like an option to leave the original URL unlinked, for use when the original URL is known to be inaccessible. I suggest an extra input parameter like urlinaccessible=YYYY-MM-DD, to cause the output to look more like this:

"Example". Archived from http://example.org/example.html on YYYY-MM-DD. The original is inaccessible as of YYYY-MM-DD.

—AlanBarrett (talk) 19:29, 15 January 2009 (UTC)[reply]

What if the original URL comes live again? Or if the archive goes dead? How would this affect the Link Checker tool? --—— Gadget850 (Ed) ^talk - 20:27, 15 January 2009 (UTC)[reply]

If the original URL is rendered in plain text, not as a link, then any link checker tool will ignore it. If the archive goes dead, then somebody will have to find another archive and edit the reference. If the original URL becomes live again, then probably nobody will notice, but I'd suggest that this proposed new feature be used only where the original URL is expected to remain dead. In the case that triggered my suggestion, it seems clear that the original URL won't come back (the domain name has been abandoned by the original publisher and is now an advertising site), and the Internet Archive seems likely to be stable. —AlanBarrett (talk) 20:57, 15 January 2009 (UTC)[reply]

This would cause the citation to be much longer and would likely lead to wrapping problems. Pagrashtak 22:39, 15 January 2009 (UTC)[reply]

I'd suggest this is a case of not needing to mess with something that already works fine. — Huntster (t • @ • c) 23:39, 15 January 2009 (UTC)[reply]

I don't think there should ever be a need to render a url in plaintext except where the article actually discusses such a thing in an encyclopedic context. If a link doesn't work, we should remove it. I'm not sure how the template currently functions with respect to the |archiveurl= parameter being specified but not |url=; if it's an acceptable output then it should be the approved practice; if not then the template should be adjusted to allow it. If it is abundantly clear that the link is not ever coming back, then the url should be removed altogether. In more dubious cases, it can just be comented out using HTML comment tags, and thus is still present in the wikitext for users to test and restore if necessary. Happy‑melon 00:05, 16 January 2009 (UTC)[reply]

The template doesn't allow |archiveurl= to be used without |url=. Changing the template to allow that seems like a reasonable solution to this problem. Then the original |url= could be commented out in the wikitext, both to retain it as a kind of documentation, and to instruct the template not to render it. —AlanBarrett (talk) 07:08, 16 January 2009 (UTC)[reply]

If we allow that, editors won't just comment it out, they'll remove it completely. Pagrashtak 17:09, 16 January 2009 (UTC)[reply]

Given that the link doesn't work, that's not really a serious problem. If a user added a broken link to an article, it would be reverted, no? We're only keeping the link against the faint possibility that it might one day fix itself. Removing it entirely is a perfectly acceptable action. Happy‑melon 17:55, 17 January 2009 (UTC)[reply]

No, that is a problem. We need to know the original source used. Not all broken links would be cause for immediate reversal. Pagrashtak 16:54, 20 January 2009 (UTC)[reply]

Why not just use the archiveURL in the |url= field, and not specify a URL? Martin (Smith609 – Talk) 23:03, 19 January 2009 (UTC)[reply]

Because you lose semantic clarity that way: it is no longer possible for a bot (or a human for that matter) to know with certainty that the url is going to link to the original page. |url= is for direct link, and |archiveurl= is for archive links. If it's not possible to use one without hte other, we need to fix the template, not change our behaviour to accomodate it. Happy‑melon 23:17, 19 January 2009 (UTC)[reply]

(outdent)I think the current functionality works just fine. Having the original URL (which has been pointed out above could possibly go live again) is important to show the source of the citation and serves as a historical record of where the information came from. Unless the original domain is hijacked by, say, someone spreading viruses or a porn site, I see no harm in retaining the links. — Bellhalla (talk) 16:41, 23 January 2009 (UTC)[reply]

I agree with Bellhalla. Mr Stephen (talk) 19:08, 23 January 2009 (UTC)[reply]

I'm confused as to what to use. For example, would this be correct usage of both the work and publisher fields: {{cite web |url=http://www.autoblog.com/2008/03/04/geneva-2008-bugatti-veyron-fbg-by-hermes-scepter-and-empire-no/ |title=Geneva 2008: Bugatti Veyron Fbg by Hermes, scepter and empire not included|last=Ramsey|first=Jonathon|date=2008-03-04 |work=autoblog.com |publisher=Autoblog|accessdate=2009-01-1 7}} What about something like this: {{cite web |url=http://www.bugatti.com/en/veyron-16.4/technology/handling.html |title=Driving the Ideal Line |accessdate=2007-10-02 |accessmonthday= |accessdaymonth= |accessyear= |author= |last= |first= |authorlink= |coauthors= |date= |year= |month= |format= |work=bugatti.com |publisher=Bugatti SAS |pages= |language= |doi= |archiveurl= |archivedate= |quote= }} — Mr. Grim Reaper at 01:53, 24 January 2009 (UTC)[reply]

I would think giving the name of the website, when it is obvious, isn't really necessary. It might be more helpful if you were citing a play titled "Henry VIII" on a site named "The Complete Works of Shakespere" but the web address gave no hint of this (say it was http://www.a9263.invalid). --Gerry Ashton (talk) 02:28, 24 January 2009 (UTC)[reply]

(EC) No. In the first example, work isn't needed, nor in the second. Work should be used primarily for newspaper and other periodical/publication websites where the name should be italicized, such as here, for a reference using an online magazine (which shouldn't use cite journal): {{cite web |url=http://www.animefringe.com/magazine/2003/06/reviews/11/ |title=Tokyo Mew Mew Vol.2 |first=Patrick |last=King |work=Animefringe |year=2003 |month=June |accessdate=2008-04-18 }}. Publisher is used for general website/company names, such as {{cite web |url=http://www.hersheyicecream.com/ |title=Hershey's Ice Cream |publisher=Hershey Creamery Company |accessdate=2009-01-07 }}. The url generally shouldn't be in there at all except, of course, in the URL field. Usually you use one or the other, though sometimes you may use both if the company that publishes a periodical is different from the work itself. For example, here are two that use both:

{{cite web |url=http://www.referenceforbusiness.com/businesses/G-L/Hershey-Foods-Corporation.html |title=Hershey Foods Corporation |publisher=[[Advameg]] |work=Reference for Business |accessdate=2009-01-07 }} which gives you:

"Hershey Foods Corporation". Reference for Business. Advameg. Retrieved 2009-01-07.

{{cite web |url=http://www.msmary.edu/alumni/whats-new/featured-alumni/Holder_Family.html |title=A Sweet Legacy: For generations, the Holder family has been part of the ice cream business and part of the Mount |first=Lisa |last=Gregory |publisher=Mount St. Mary's University |work=Mount Magazine |month=Fall |year=2008 |accessdate=2008-12-30 }} generates:

Gregory, Lisa (2008). "A Sweet Legacy: For generations, the Holder family has been part of the ice cream business and part of the Mount". Mount Magazine. Mount St. Mary's University. Retrieved 2008-12-30. {{cite web}}: Unknown parameter |month= ignored (help)

Hope that helps some? -- Collectonian (talk · contribs) 02:34, 24 January 2009 (UTC)[reply]

Yes indeed. Thank you very much. — Mr. Grim Reaper at 03:00, 26 January 2009 (UTC)[reply]

Is there a better name for those fields? I put newspapers into publisher with apostrophes for the italics for a long time. - Peregrine Fisher (talk) (contribs) 03:28, 26 January 2009 (UTC)[reply]

Check it out. - Peregrine Fisher (talk) (contribs) 01:26, 25 January 2009 (UTC)[reply]

... I've lost track, is it broken again, broken still, or obsolete? {{cite web |title=My favorite things part II |work=Encyclopedia of things |url=http://www.example.org/ |accessmonthday=July 6 |accessyear=2005 }}

"My favorite things part II". Encyclopedia of things. {{cite web}}: Unknown parameter |accessmonthday= ignored (help); Unknown parameter |accessyear= ignored (|access-date= suggested) (help)

Mr Stephen (talk) 23:05, 28 January 2009 (UTC)[reply]

Obsolete IMO, although I'd wait to hear other opinions. Happy‑melon 08:35, 29 January 2009 (UTC)[reply]

I just fixed Scout Moor Wind Farm a "featured article" with this problem. 71.250.136.115 (talk) 13:26, 29 January 2009 (UTC)[reply]

{{editprotected}} Please revert the template to this version that recognizes the accessdaymonth and accessmonthday . Some of us have article under review at WP:FAC that use these parameters. This same sort of disruption occurred almost exactly a month ago.

Why do these parameters keep disappearing? Is there some particularly compelling reason for these parameters to not be included in the template? — Bellhalla (talk) 20:40, 30 January 2009 (UTC)[reply]

Should be a relatively easy fix, just move the contents of accessdaymonth/accessmonth day and access year into accessdate. You don't need to have the date in ISO format and there isn't any auto-date formatting on cite web anymore. --Bobblehead ^(rants) 20:56, 30 January 2009 (UTC)[reply]

Also, in answer to your question, there isn't exactly a compelling reason to keep the parameters. The main reason why they were created is because back in the days when date autoformating was encouraged, the accessdate field autolinked. Some people didn't like that so they added the accessdaymonth/accessmonthday and access year parameters so the autoformatting could be avoided. Now that the autolinking has been removed from autodate (on cite web), there isn't a reason to keep the additional parameters. --Bobblehead ^(rants) 21:11, 30 January 2009 (UTC)[reply]

It would also be "a relatively easy fix" to retain the alternate parameters. If consensus is that they are to be deprecated (i.e. no longer recommended) they can be removed from the documentation page (where they still are listed, by the way), bit no articles would need to be "fixed". If, however, they are made obsolete, as is currently the case (was there consensus for making them obsolete?), it requires much more effort to "fix" them than would be spent in coding them to continue working. — Bellhalla (talk) 21:20, 30 January 2009 (UTC)[reply]

On the contrary, it would be an extremely complicated and messy fix, since unlike the |date= aliases, there is no alternative but to mash the various accessdate parameters together before it hits the {{citation/core}} template that normally does all this. There are about 1,350 pages that use these parameters, out of probably over 350,000 total (it was 319,000 last September). I will work on a bot script to make these transitions. Happy‑melon 21:38, 30 January 2009 (UTC)[reply]

(edit conflict)Fixed them on the article you're concerned with, so that shouldn't be a problem anymore. Granted, I don't believe that's your point...;) --Bobblehead ^(rants) 21:41, 30 January 2009 (UTC)[reply]

Thank you for the edits, Bobblehead.

Thanks for the explanation, HappyMelon. While you are working on the bot script, can you also design it so that it will perform the same search-and-replace on the following templates that call {{cite web}}:

• {{cite Arnold Hague Convoy Database}}
• {{cite DANFS}}
• {{cite Miramar}}
• {{cite PMARS}}

Thanks in advance. — Bellhalla (talk) 22:20, 30 January 2009 (UTC)[reply]

Coding.... Happy‑melon 22:39, 30 January 2009 (UTC)[reply]

BRFA filed Wikipedia:Bots/Requests for approval/MelonBot 12. Happy‑melon 23:02, 30 January 2009 (UTC)[reply]

Thanks Bellhalla for mentioning {cite DANFS}...I was wondering why the dates were off in USS Nevada (BB-36) the other day... :) And thank you Happy-melon for coding and doing all of this! —Ed 17 ^{(Talk / Contribs)} 01:20, 3 February 2009 (UTC)[reply]

{{cite web}}

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

<span class="error">Error: no {{para|title}} specified when using {{tl|cite web}}</span>

<span class="error">Error: no {{para|url}} specified when using {{tl|cite web}}</span>