Wikipedia:Requests for comment/Dead url parameter for citations

From Wikipedia, the free encyclopedia
The following is a closed discussion of the proposal. Please do not modify it. Subsequent comments should be made in a new section on the talk page. No further edits should be made to this section.

Proposal has unopposed support after 19 days listing. Discussion has now ceased. There is clear consensus to implement adding a |deadurl= parameter to the citation templates. SilkTork *Tea time 10:44, 9 June 2011 (UTC)[reply]

This RfC is to add a |deadurl= parameter to the citation templates. This is so the readers and editors can differentiate between a dead original source link |deadurl=yes or a pre-emptively archived one |deadurl=no. Currently there are two reasons an |archiveurl= parameter is added – 1) dead link replacement (example) and 2) pre-emptive archiving (example). However, there is no way to tell the two edits apart if they are not somehow marked. The status quo until recently has been to treat |archiveurl= as an indication that the url is dead. But as linkrot is becoming a more pressing matter (see Wikipedia:WikiProject External links/Webcitebot2), more editors are archiving material pre-emptively. This is a good cause, but it leaves all citations showing the archived url as the main link indiscriminately. Thus, I believe the need for differentiating the two cases when archive urls are present – whether the original links is a dead url or a live url. I do not currently propose to change the display of citations without a |deadurl= parameter. In fact, only |deadurl=no would affect the display.

The problem with archiving live links is that the main link from the citation is to the archived version (see below). This means clicking the link will take the reader to the archiving service (WebCite, Wayback, etc.) and not the actual original url. The archiving services are notoriously slow and sometimes unreliable. Obviously, the editor may have intended this behaviour, so adding |deadurl=yes or leaving the citation intact would keep the citation linking to the archived version. But most of pre-emptive archiving does not require linking to the archived version.

There are several bots and tools that have been approved to detect dead links reasonably well. They can be easily modified to use the |deadurl=yes when archiving a dead link. Similarly, there are several tools for pre-emptive archiving. They can also be easily modified to use |deadurl=no. (Note that bots will never change "deadurl=yes" to "no". Firstly, live link detection is unreliable as opposed to dead link detection and, secondly, they would not know is the editor intended to show the archived version.) As time goes on, users will add more archived links and more links will rot. Both cases are easily covered by archiving bots – marking the link as |deadurl=yes. For pre-emptive archiving, the editor can themselves add |deadurl=no or this will be done by the tool they use. Of course, many citations won't have the parameter; but as I do not propose to change the status quo display for a missing parameter, the citation display will remain unaffected and cause no issues.

Current behaviour
{{Cite web |url=http://example.org |title=Example web-page |author=Smith, John |archiveurl=http://replay.web.archive.org/20050312094334/http://example.org/ |archivedate=2005-03-12}}
Smith, John. "Example web-page". Archived from the original on 2005-03-12.
Proposed behaviour
{{Cite web |url=http://example.org |title=Example web-page |author=Smith, John |archiveurl=http://replay.web.archive.org/20050312094334/http://example.org/ |archivedate=2005-03-12}}
Smith, John. "Example web-page". Archived from the original on 2005-03-12.
{{Cite web |url=http://example.org |title=Example web-page |author=Smith, John |archiveurl=http://replay.web.archive.org/20050312094334/http://example.org/ |archivedate=2005-03-12 |deadurl=no}}
Smith, John. "Example web-page". Archived from the original on 2005-03-12.
{{Cite web |url=http://example.org |title=Example web-page |author=Smith, John |archiveurl=http://replay.web.archive.org/20050312094334/http://example.org/ |archivedate=2005-03-12 |deadurl=yes}}
Smith, John. "Example web-page". Archived from the original on 2005-03-12.
{{Cite web |url=http://example.org |title=Example web-page |author=Smith, John |archiveurl=http://replay.web.archive.org/20050312094334/http://example.org/ |archivedate=2005-03-12 |deadurl=something else}}
Smith, John. "Example web-page". Archived from the original on 2005-03-12.
{{Dead link}} simplification and moving closer to the url

Per comment below, added 17:38, 30 May 2011 (UTC). A citation without |archiveurl= parameter but with |deadurl=yes will display a {{Dead link}} template next to the url. This will make placing {{Dead link}} after the citation unnecessary. The change can be easily implemented by bot. This would also make syntax consistent with above and reduce workload in case archive parameters are added in the future.

Current usage: {{Cite web |url=http://example.org |title=Example web-page |author=Smith, John |accessdate=30-05-2008}} {{Dead link}}
Smith, John. "Example web-page". Retrieved 30-05-2008. [dead link]
Proposed syntax: {{Cite web |url=http://example.org |title=Example web-page |author=Smith, John |accessdate=30-05-2008 |deadurl=yes}}
Smith, John. "Example web-page" [dead link]. Retrieved 30-05-2008.
Discussion

As nom. —  HELLKNOWZ  ▎TALK 07:34, 21 May 2011 (UTC)[reply]

Looks good, very good in fact. Can you clarify one thing: that |deadurl= with no value will be equivalent to |deadurl=no? Also I've checked existing usage from March 2011 dump and there are 9900 yes, 790 no, and about 15 other values that I'll sort out manually. Rjwilmsi 12:07, 21 May 2011 (UTC)[reply]
Any non-documented value would be treated the same way as not setting the parameter at all, just to be sure. I don't think we need additional cases for now, so we could categorize the page as problematic or show an error for incorrect usage? Also the dump had more uses than I thought it would. —  HELLKNOWZ  ▎TALK 15:08, 21 May 2011 (UTC)[reply]
I think it is OK if people think this is a significant problem, though I haven't run across it as a problem myself. My only worry would be that this will be one more mysterious parameter that uninitiated users might not realise they don't normally have to bother with, and we have too many of those already, such as "publisher" in "cite news", which editors constantly think they have to fill in with something, even though in the vast majority of cases it is unnecessary. -- Alarics (talk) 12:16, 21 May 2011 (UTC)[reply]
It is only needed if archive parameters are used; in fact, it only makes a difference if the archived urls are pre-emptive ones and the actual links are live. Since |archiveurl= is used by more experienced editors anyway, I think this optional parameter will not cause confusion. —  HELLKNOWZ   TALK 15:27, 21 May 2011 (UTC)[reply]
OK well then just go ahead and do it. -- Alarics (talk) 21:20, 21 May 2011 (UTC)[reply]
  • Support - This would be helpful and sounds like a good idea. There are some similar discussions from March 2009, September 2009 and February 2010. - Hydroxonium (TCV) 22:47, 21 May 2011 (UTC)[reply]
  • Support. Sensible and helpful. -- œ 20:16, 22 May 2011 (UTC)[reply]
  • Support - This seems useful in case large-scale preemptive archiving will be performed. Toshio Yamaguchi (talk) 11:26, 23 May 2011 (UTC)[reply]
  • Support. Clearly handling archiving issues is one of the few unambiguous advantages of cite templates, and this builds on that. Really wish, though, that WMF would prioritise finding a way to separate out ref data from the main body of an article, to make wikitext easier to parse... Rd232 talk 09:36, 29 May 2011 (UTC)[reply]
  • Support Whenever I pre-emptively archive, I always hate it when the first link in the ref goes to the archived version. My only concern is with the citation style; I dislike "Smith, John. "Example web-page". Archived copy on 2005-03-12". This seems ... awkward. What about keeping the same current format ("Smith, John. "Example web-page". Archived from the original on 2005-03-12") but simply changing the links so it becomes "Smith, John. "Example web-page". Archived from the original on 2005-03-12"? /ƒETCHCOMMS/ 02:22, 30 May 2011 (UTC)[reply]
    • Good point. Rephrased to be closer to the original wording. 17:19, 30 May 2011 (UTC)
  • Support. I like this idea. Building on what User:Fetchcomms said; the Status quo actually makes me unwilling to preemptively archive. Which will cause more frustration later on. I don't care much about the exact wording as long as it makes absolute grammatical sense and preferably "deadurl= 'yes' or 'no'" would have slightly different outputs. That way readers can learn out what to expect. Awg1010 (talk) 02:51, 30 May 2011 (UTC)[reply]
    • Do you mean |deadurl=yes would have a different output to not having |deadurl=?
  • Support Sounds like a fantastic idea. One question, is there any way to have |deadurl=yes indicate that the link is a deadlink even if there is no archive link up? That way, this could replace {{dead link}}. NW (Talk) 13:39, 30 May 2011 (UTC)[reply]
    • That is a good point, it would move the [dead link] right after the link and not after the citation. It would also use the same template (I assume that shouldn't cause problems even though template doc says "Avoid complicated formatting. Nesting inside a template is generally not supported well.") so whatlinkshere and categorization remain the same. The change can also be easily implemented with a bot. —  HELLKNOWZ  ▎TALK 17:38, 30 May 2011 (UTC)[reply]
      • I like that. It's easier to see what the deadlink is, instead of how we currently have that template and the end of the whole ref. /ƒETCHCOMMS/ 20:30, 30 May 2011 (UTC)[reply]
      • I support this as well as an alternative to directly using {{dead link}}. —Ost (talk) 18:25, 2 June 2011 (UTC)[reply]
  • Support, Awg1010's point about the current status quo discouraging preemptive archiving is dead on. It may be that not many people will use this, but it'll be an significant improvement for many articles which do, and that's good enough. --joe deckertalk to me 17:43, 30 May 2011 (UTC)[reply]
  • Support. It looks like a very constructive improvement, and I cannot think of any problems with it. --Tryptofish (talk) 18:43, 30 May 2011 (UTC)[reply]
  • Support iff this gets added to all {{cite foo)) templates. {{Cite News}} and {{Cite Book)) can have the same issue --Guerillero | My Talk 01:19, 31 May 2011 (UTC)[reply]
    • This would be applicable to all "cite xxx" templates that use |archiveurl= —  HELLKNOWZ  ▎TALK 09:21, 31 May 2011 (UTC)[reply]
  • Support This will make the template more useful.Curb Chain (talk) 08:55, 31 May 2011 (UTC)[reply]
  • Support Great idea.--v/r - TP 22:15, 31 May 2011 (UTC)[reply]
  • Support. Will the checklinks tool be updated to handle the deadurl parameter for cites without archiveurl? -84user (talk) 13:33, 1 June 2011 (UTC)[reply]
    • That's a question to ask Dispenser (talk · contribs). /ƒETCHCOMMS/ 13:41, 1 June 2011 (UTC)[reply]
    • I would notify the bot/tool owners that I know of. But, of course, I cannot vouch for how soon or if at all they would implement it. —  HELLKNOWZ  ▎TALK 14:19, 1 June 2011 (UTC)[reply]
  • Support - I think that this will promote the inclusion of archive links while helping users to find the active page if it exists. My only minor gripe is that the two external links are barely separated in the new version; I concede that the EL icon makes this less an issue than if they were two wikilinks, but I'm not sure new users know what the icon means. —Ost (talk) 18:25, 2 June 2011 (UTC)[reply]
    • Personally, I find the little external link arrow to separate them really well. In any case, proper cites would include |work= or |publisher= and often other fields that would "push" the archive info away from the link. The only solution would be to change the wording, but I think the current one is as good as it may get while staying consistent and grammatically correct. —  HELLKNOWZ  ▎TALK 18:36, 2 June 2011 (UTC)[reply]
  • Support Great idea --SPhilbrickT 22:30, 2 June 2011 (UTC)[reply]
The above discussion is preserved as an archive of the proposal. Please do not modify it. Subsequent comments should be made in a new section on this talk page. No further edits should be made to this section.