Wikipedia:Bots/Requests for approval/Citation bot 5


 * The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section. The result of the discussion was Symbol keep vote.svg Approved.

Citation bot 5
Operator: Martin  (Smith609 – Talk) 

Automatic or Manually Assisted: Automatic

Programming Language(s): PHP

Function Overview: Change 'Cite ArXiV' to 'Cite Journal' where appropriate

Edit period(s): Continuous

Already has a bot flag (Y/N): Y

Function Details:


 * Suggestion copied from User_talk:Citation_bot

Many cite arXiv could be converted to cite journal For example, this page http://arxiv.org/abs/cond-mat/0605258v1 gives the journal into with this was published Nature 442, 54-58 (2006). Why not change it to a cite journal (with |id= )?Headbomb {{{sup|ταλκ}}κοντριβς – WP Physics} 20:25, 28 March 2009 (UTC)


 * That sounds like a great idea. I'm not entirely familiar with the ArXiV system; could you give me a set of rules which I could turn into bot code?  I would need to know the situations where converting to cite journal is appropriate, because presumably there are many scenarios where it isn't (or cite ArXiV wouldn't exist).  Martin  (Smith609 – Talk)  01:32, 29 March 2009 (UTC)

As far as I'm aware, when you can find the journal the article has been published into, it's appropriate to change cite arXiv to cite journal, otherwise they should remain cite arXiv (I'm not very familiar with citation, but I suspect there wouldn't be any additional problem than those mentionned for cite journal). As long as you adapt the parameters from cite arXiv to into an  in cite journal, there should be no problem. and are optional parameters; there is no need to modify the logic used when  is not used, as  will produce the correct output; however if class is not present, the brackets should not be present (i.e, it should look like  rather than ). All other fields from cite arXiv (i.e., author, title, etc...) overlap with those of cite journal.


 * For example produces the following:   This can be duplicated using cite journal with  which produces the following (not incorporating other changes by DOI bot, such as add doi, journal, volume, page, etc...):
 * Headbomb {{{sup|ταλκ}}κοντριβς – WP Physics} 02:26, 29 March 2009 (UTC)


 * I would still recommend a trial period (50 or so edits) on random articles from Special:WhatLinksHere/Template:Cite_arXiv just to make sure that I didn't miss something, or that I'm not aware of some subtleties.Headbomb {{{sup|ταλκ}}κοντριβς – WP Physics} 02:35, 29 March 2009 (UTC)

Discussion
BAGAssistanceNeeded
 * Yes, this sounds good... but I think a trial run would help us get a feel for what could go wrong. – Quadell (talk) 13:49, 1 April 2009 (UTC)

It seems that the suggested changes are only equivalent if cite journal's journal parameter is specified; otherwise, cite journal puts the title in italics where cite arXiv uses quotes. Is it possible for the bot to determine the needed journal parameter? Anomie⚔ 18:33, 5 April 2009 (UTC)


 * My understanding of the request was that action would only be taken if a journal parameter either existed or could be found. Martin  (Smith609 – Talk)  21:13, 5 April 2009 (UTC)


 * Sounds good. Anomie⚔ 21:16, 5 April 2009 (UTC)


 * So what's the status? – Quadell (talk) 23:09, 15 April 2009 (UTC)
 * I'm afraid I've not had chance to implement this yet. Martin  (Smith609 – Talk)  19:15, 16 April 2009 (UTC)
 * The coding's done; User:Citation bot 2 will be bravely demonstrating its new capabilities.
 * NB. Most of its edits don't actually involve this change: here's one that does; here's anohter. Martin  (Smith609 – Talk)  00:59, 19 April 2009 (UTC)

Looks good to me, so far. Let us know when the bot has made somewhere around 50 edits. – Quadell (talk) 01:34, 19 April 2009 (UTC)
 * It's edited about 45 distinct pages now. Martin  (Smith609 – Talk)  01:49, 19 April 2009 (UTC)


 * Your bot edit summaries don't make it immediately obvious which of these are arXiv edits, but I've looked through a number of them and they look good. Any idea why this edit left it with "cite arXiv"? – Quadell (talk) 02:25, 19 April 2009 (UTC)
 * That last one is not yet published in any journal (although it has been submitted to The Astrophysical Journal), so the bot can't (and shouldn't) change it to a cite journal. Headbomb {{{sup|ταλκ}}κοντριβς – WP Physics} 04:36, 19 April 2009 (UTC)

I'm reviewing every edit, and I've found a few weird ones. This might not be exclusive to this task (aka, it might simply be how citation bot usually work)


 * Pointless whitespace edit
 * Pointless whitespace edit
 * Bot overrides human input for authors
 * Bot overrides human input for authors
 * Bot overrides human input for authors
 * Bot overrides human input for authors
 * Bot overrides human input for authors
 * Bot overrides human input for authors
 * Bot overrides human input for authors
 * Bot overrides human input for authors
 * Bot overrides human input for authors (in this case it is an improvement, but it still shouldn't have done it.
 * Bot overrides human input for authors
 * Etc..

Overriding human input is definitely out of line for this bot. I've pointed this out several times now and it still hasn't been fixed. Changing cite arxiv to cite journal seem to work fine however.Headbomb {{{sup|ταλκ}}κοντριβς – WP Physics} 04:47, 19 April 2009 (UTC)
 * I disagree. This isn't a situation where humans make creative decisions; it's crediting someone for their work. In every case you list, the bot's version is better than the human version, since the bot pulls the name or names the way they wanted to be credited, the way the journal officially credits them. I think these changes are appropriate. – Quadell (talk) 05:20, 19 April 2009 (UTC)
 * Writing J. Bob et al. instead of Bob, J. et al or Long list of 103 authors is a stylistic decision. Changing this often (see nearly always) introduces inconsistencies in the citation style used for a particular article. The bot is neither able to make those calls, nor approved to.Headbomb {{{sup|ταλκ}}κοντριβς – WP Physics} 05:43, 19 April 2009 (UTC)
 * These edits are unrelated to this task; they are edits which the bot makes anyway in its role of adding missing information to citations. There are pretty good reasons for not specifying et al in the template, which I won't go into here.  Note that if more than 10 authors are specified, only the first 9 are shown by the templates.
 * The whitespace quirk is something which I'll look into (but, again, unrelated to this request). It's slightly complicated and I have to fix a couple of things with a different bot first. Martin  (Smith609 – Talk)  14:10, 19 April 2009 (UTC)
 * There are also pretty good reasons to specify et al in the template, and to use J. Smith over Smith, J. or Smith, John. The bot isn't equipped to make these decisions; so it shouldn't make 'em. It more-oftens-than-not introduces inconsistencies in how citations are presented. Again, not related to this task, but it should be fixed. If it keeps doing this, I'll stop the bot. Headbomb {{{sup|ταλκ}}κοντριβς – WP Physics} 15:09, 19 April 2009 (UTC)
 * As I've said, this isn't a discussion for here. At Template:Citation, only the last and first parameters are documented.  I haven't spotted the related discussion in the talk page and am not in the mood for trawling through the archives, but the implicit consensus is that the use of two parameters is better, for reasons including metadata availability and compatability with the harv family of citations.  One solution would be for the template to accept a parameter which controls whether authors are displayed as first-last or last-first, and whether the first 9 or the first 1 authors are displayed before the et al.  This would retain metadata while allowing precise format tweaking.  If this issue rankles you, I would suggest that you discuss it more centrally, where other editors are able to chip in, to make sure that the solution we come to reflects a broad consensus. [I should add that the reason that I'm not immediately stopping this function is because it's quite complicated to un-code it, and I have a couple of other urgent coding jobs to do presently.]  Martin  (Smith609 – Talk)  15:47, 19 April 2009 (UTC)
 * Encouraging first/last is not the same thing as making it mandatory. Consistency and style trumps metadata. I don't care if the bot adds first/last/etc... if they wouldn't end up being displayed and screwing up the style of an article. I'll start a discussion if need be, but this is a task that should have been approved first. 16:20, 19 April 2009 (UTC)
 * I think that changing the templates so they can produce both style and metadata is the only solution that will satisfy both of us.  That will require discussion at the template pages.  Martin  (Smith609 – Talk)  16:45, 19 April 2009 (UTC)
 * I've also noted that some people place "|journal=arxiv" (in both cite arxiv and cite journal, see for example). The bot should recognize this and switch cite journal to cite arxiv in those cases (or simply comment the journal parameter away when found in cite arxiv, as there is no use for that parameter. There might be some technical problems associated with that, but this shouldn't be hindering this task. Headbomb {{{sup|ταλκ}}κοντριβς – WP Physics} 04:52, 19 April 2009 (UTC)
 * I'll get round to this - is it all right if I get round to it at a later date? Until this request is complete, it makes maintenance of the bot more time consuming (because I have to attend a new and old version), so I am loath to prolong the time before I can go live with the arxiv->journal change. Cheers,  Martin  (Smith609 – Talk)  14:10, 19 April 2009 (UTC)
 * Like I said, this isn't related to this bot task (well it is, but nothing major IMO), so it shouldn't impede its approval IMO.Headbomb {{{sup|ταλκ}}κοντριβς – WP Physics} 15:09, 19 April 2009 (UTC)

, though Headbomb's concerns may need to be addressed separately. – Quadell (talk) 14:11, 20 April 2009 (UTC)


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