Wikipedia talk:Full-date unlinking bot

RFC

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.

I think I should have added some kind of comment when closing the RFC. I am happy that there is large support for this proposal. Supporters want these date links finally removed and appreciate that the proposal is conservative and limited, leaving the possibly controversial cases. Opposers raise concerns about false positives, whether this is a suitable task for an automatic bot and that it might create drama. I think the support is strong enough to proceed. There are many constructive comments that will help in designing the bot. I would have wanted even more participation, but over 100 comments should be enough to represent the general community consensus.

The next step is to start implementing the bot and work out any unclear details as they are discovered. User:harej has started the task at User:Full-date unlinking bot. The bot will be presented for BAG approval later.--Apoc2400 (talk) 16:14, 9 July 2009 (UTC)

---

Please indicate your support or opposition to this proposal, along with your comments. Please use numbers to refer to specific points in the proposal that you are referring to, particularly if you overall support the proposal but oppose particular points, or vice-versa. The RFC will be open for two weeks, until Monday 6 July, 23:59 (UTC). --Apoc2400 (talk) 10:14, 22 June 2009 (UTC)

Support

1. Support as proposer. Date linking for autoformatting has been deprecated since a discussion in August last year. Two RFCs in December, one in January and one in April has confirmed this and further clarified when dates should be linked. There has still been dispute over where and how dates can be delinked leading to an arbitration case. I have not been involved in the conflict, so I hope this discussion can focus on the proposal, not people. --Apoc2400 (talk) 10:35, 22 June 2009 (UTC)
2. Support. This looks like a cautious, thoughtful proposal that proceeds conservatively with respect for all points of view. This is how the matter should have been handled in the first place. – Quadell ^(talk) 11:00, 22 June 2009 (UTC)
3. Support, is this a vote? I sure hope not. Anyway, delinking dates in this way seems like a good plan. I never could understand the value of linking dates as a matter of course, and I think that editors got the impression that it was required, and then imitated what they saw. Abductive (talk) 11:17, 22 June 2009 (UTC)
   • It is not supposed to be a vote. Not more than any other community discussion anyway. --Apoc2400 (talk) 12:10, 22 June 2009 (UTC)
4. Support. As Quadell notes, this is how it should have been done. — Hex (❝?!❞) 11:52, 22 June 2009 (UTC)

   Support. Per Quadell and Hex. Bravo. AKAF (talk) 11:54, 22 June 2009 (UTC) Moved to oppose.

5. Support. Well-thought-out and long overdue solution to a problem that everyone would surely want to be moving on from.--Kotniski (talk) 11:58, 22 June 2009 (UTC)
6. Support. Rettetast (talk) 12:42, 22 June 2009 (UTC)
7. Support. Lincher (talk) 12:55, 22 June 2009 (UTC)
8. Support per Apoc, Greg and Phil. Ironholds (talk) 12:56, 22 June 2009 (UTC)
9. Support A good proposal that is in line with consensus. Dabomb87 (talk) 12:59, 22 June 2009 (UTC)
10. Support - A conservative approach to the problem of implementing consensus Ealdgyth - Talk 13:47, 22 June 2009 (UTC)
11. Support—Such a bot is well overdue. Consensus to remove date links has been demonstrated repeatedly & overwhelmingly. It's a huge job, very tedious & time-consuming to do by hand. JIMp _talk·cont 14:25, 22 June 2009 (UTC)
12. Support - Galoubet (talk) 14:56, 22 June 2009 (UTC)-
13. Support - As a preliminary runthrough, this is very good. NW (Talk) 15:28, 22 June 2009 (UTC)
14. Conditional Support - subject to an edit banner akin to the 'request for funding' or 'arbcom elections' letting editors know this will be happening and to request articles for exclusion if they prefer.--Joopercoopers (talk) 15:48, 22 June 2009 (UTC)
15. Support for the reasons already mentioned, and because the lies that result when non-Gregorian dates are reformatted into the ISO 8601 format will cease. I am delighted this will end the debate, because once the runs are finished, the metadata arguments will have no validity, because whatever limited and incomplete metadata used to exist will be gone, and would require massive manual intervention to restore it. --Jc3s5h (talk) 16:24, 22 June 2009 (UTC)
16. Support. I have full confidence that the bot-running community - one with which I am familiar - will be able to handle the technical implementation of this. Take Quadell (above) as just one example of the sort of person that can be made available. - Jarry1250 ^{(t, c, rfa)} 16:43, 22 June 2009 (UTC)
17. Support. This seems like a good, cautious approach to date-delinking, especially with the inclusion of points 1, 2, 3, and 4. –Drilnoth (T • C • L) 16:41, 22 June 2009 (UTC)
18. Support. I never could understand why Wikipedia linked dates to articles of historical trivia that had nothing to do with the subject of the linking article. Once I understood what these links were about, I learned to avoid them. Hardly “valuable” links. Fortunately, the community became aware of its chronic problem with the full gamut of its link‑itis and now no longer permits non-germane date links. But there are still literally hundreds of thousands of linked dates to be delinked. A bot is the only practical way to address all these to improve Wikipedia. The proposal is a perfectly sound next step. Greg L (talk) 17:18, 22 June 2009 (UTC)
19. Support per Greg L above. Ben MacDui 17:44, 22 June 2009 (UTC)
20. Support. There is no doubt that the huge number of overlinked date-triples are there because of, and only because of, the failed method of date autoformatting. Correcting that mistake en bloc will allow editors to concentrate on making a positive decision about where dates should be linked. --RexxS (talk) 18:05, 22 June 2009 (UTC)
21. Support. Seems a sensible approach. These linked dates have always bugged me. Hchc2009 (talk) 18:57, 22 June 2009 (UTC)
22. Support. Let's get this over with. --Zvika (talk) 19:10, 22 June 2009 (UTC)
23. Support. Overdue, please bring it on. -- Banjeboi 21:43, 22 June 2009 (UTC)
24. Support. Recognizing that the original decision to unlink dates was somewhat controversial, that decision is behind us now. We need consistency, and this is the best way to accomplish that. Dr. Cash (talk) 21:56, 22 June 2009 (UTC)
25. Suppport. If we do not want to perpetuate this issue indefinitely we have to clean up. The proposal should be completely non-controversial and is quite conservative. Actually, I am not so sure if point 12 (publicizing the bot code) would be a good idea. Cacycle (talk) 23:52, 22 June 2009 (UTC)
26. Support for overly-conservative proposal. Getting the most tedious, most obvious, and least controversial edits out of the way in a clean, regulated process such as this will be most effective in furthering all the next steps. ~ Amory (user • talk • contribs) 02:19, 23 June 2009 (UTC)
27. Support—per Quadell, RexxS et al. Tony (talk) 03:54, 23 June 2009 (UTC)
28. Support—I created a handful of Featured Articles that had to abide by the datelinking rule and would love to see those dates removed. I'm unwilling to do it by hand since it would require hundreds of deletions, and I think a bot is an excellent idea. JKBrooks85 (talk) 04:34, 23 June 2009 (UTC)
29. Support. I agree with Quadell, Greg L and RexxS. Chris the speller (talk) 05:09, 23 June 2009 (UTC)
30. Support The proposal seems well-stated, especially since relevant date links can be kept without issue by reverting the bot. ThemFromSpace 05:49, 23 June 2009 (UTC)
31. Support A very well thought out proposal for moving forward, especially with the checks and balances and provisions for preserving date links that editors identify as important. -- Tcncv (talk) 06:21, 23 June 2009 (UTC)
32. Support per above. DrKiernan (talk) 07:36, 23 June 2009 (UTC)
33. Support Considering we are all volunteers to this project this seems a reasonable way of making ones life easier when writing and editing articles.--EnigmaMcmxc (talk) 08:55, 23 June 2009 (UTC)
34. Seems well thought-out.  Sandstein  11:12, 23 June 2009 (UTC)
35. Support per Quadell. Darrenhusted (talk) 11:59, 23 June 2009 (UTC)
36. Support Why not? No need to make things harder than they are. Have a bot do it. hmwithτ 14:52, 23 June 2009 (UTC)
37. Support per above. It would save editors' time, assist in giving the encyclopedia greater consistency. Moreover, bluelinking these dates does not help the reader, but only encourages him or her to click away from the article that he or she really wishes to read. -- Ssilvers (talk) 14:53, 23 June 2009 (UTC)
38. Support so as to have done with the whole process once and for all (hopefully). Guest9999 (talk) 16:22, 23 June 2009 (UTC)
39. shoy (reactions) 16:36, 23 June 2009 (UTC)
40. Seems reasonable. R. Baley (talk) 17:50, 23 June 2009 (UTC)
41. Support. Ruslik_Zero 18:33, 23 June 2009 (UTC)

    It is much easier to removed 1,000 links and then readd one useful link manually, than to wait years before editors remove 999 useless links by hand. Ruslik_Zero 07:50, 25 June 2009 (UTC)

42. Support. Unfortunately there is so much of this linkcruft that it can't all be removed by hand. And it must be removed to ensure that meaningful links don't get overlooked in the blue sea. This bot proposal is very conservative, and that's exactly what we need in this situation. Hans Adler 18:51, 23 June 2009 (UTC)
43. Conditional Support. My first concern is with item #4. While I understand that a bot should not get into a revert war with editors, this item may well prevent the bot from fixing new editing errors in articles. I'm not sure how this can be resolved, but it should at least be looked at. My second concern is that when the bot begins operation, there should be some limits on the number of articles edited per day. This will allow time to fine tune the bot. My fear is that there will be conditions that we have not anticipated. Given the past history and many of the oppose concerns, a slow start would reduce the chance for this bot to create new controversy. Vegaswikian (talk) 19:29, 23 June 2009 (UTC)
44. Support. The clear consensus is that nearly all dates should not be linked, and the current situation is that very many are, so the most efficient starting point for getting to the consensus situation by manual editing is by defaulting to unlinked dates where they have not been identified as requiring links. Phil Bridger (talk) 20:25, 23 June 2009 (UTC)
45. Support—as per GregL above.  HWV258  22:33, 23 June 2009 (UTC)
46. Support, although month-day links, even without a linked year, should also be unlinked. — Joe Kress (talk) 01:40, 24 June 2009 (UTC)
47. Support as a much-needed step toward clearing out the non-useful links, and toward making dates less confusing for editors who learn by imitating what they see. Sssoul (talk) 05:59, 24 June 2009 (UTC)
48. Support. The fact that more drastic solutions have been controversial should not weigh against this moderate proposal, which incorporates well-conceived checks and limitations. Nor should the fact that it is only a partial solution count against it. That's all we can realistically hope to implement, at this stage.–_⊥^{¡ɐɔıʇǝo}Noetica!^T– 23:16, 24 June 2009 (UTC)
49. Support. At this point I think most of the significant articles have already delinked. This is just cleaning up the crumbs. Kaldari (talk) 23:19, 24 June 2009 (UTC)

    There are many articles that still have linked dates (especially biographies), although yes, many articles of important subject matter have been delinked due to the traffic they receive, as well as our featured articles and lists, which were part of the initial consensus gathering on date unlinking. Dabomb87 (talk) 23:27, 24 June 2009 (UTC)

50. Support Conservative proposal for a much-needed first step. cheers, Struway2 (talk) 12:39, 25 June 2009 (UTC)
51. Support - the proposal is well thought out, will not cause significant damage, is a required and highly useful first step in fixing the date mess, and should definitely be done. Ale_Jrb^talk 13:13, 25 June 2009 (UTC)
52. Support. Let's just put this issue to bed once and for all by getting rid of these autoformatted dates. --Malleus Fatuorum 14:05, 25 June 2009 (UTC)
53. Support A reasonable and conservative implementation of existing consensus. I like the fact that the bot will never intentionally visit a page twice, but I wouldn't care if it accidentally revisited a page that had been moved in the interim. GRBerry 16:47, 25 June 2009 (UTC)
54. Yes. –Juliancolton | ^Talk 19:22, 25 June 2009 (UTC)
55. Support This is a carefully crafted and tightly circumscribed proposal to remove only those date links that are clearly there solely to support the autoformatting the community has declared it doesn't want. It's a big step towards an end to this lengthy and painful dispute. Colonies Chris (talk) 21:32, 25 June 2009 (UTC)
56. Support - Let a bot do the work. - Peregrine Fisher (talk) (contribs) 00:11, 26 June 2009 (UTC)
57. If approved, the bot will delink dates that are not supposed to be linked in any case, at least in this way. A bot can get most of the useless links out of the way, leaving editors free to finally make a meaningful separation between relevant and irrelevant dates. The proposal seems well-designed and sufficient to prevent any potential problems, so... Why not? Waltham, The Duke of 08:52, 26 June 2009 (UTC)
58. Support - I think it is clear that there is a consensus to unlink the dates. This bot seems a good way to do so. -- Donald Albury 11:49, 26 June 2009 (UTC)
59. Support, although I would support a step further - delinking date fragments as well. Karanacs (talk) 16:03, 26 June 2009 (UTC)

    I'd advise waiting for this to be completed before looking at that issue. Although consensus in now clear on fragments, editorial judgement is going to be required, so although a bot might be used to produce a list of pages with fragments, I suspect that actual delinking may have to be performed by humans. Let's cross that bridge when we come to it. --RexxS (talk) 16:33, 26 June 2009 (UTC)

60. Support Definitely. It would be much more efficient for a bot to do the job of putting consensus into action, rather than having editors do the work manually. Timmeh 00:46, 27 June 2009 (UTC)
61. Support - Overly conservative in my view, but worth doing nonetheless. It would be totally impractical to do this job by hand. Pfainuk talk 08:43, 27 June 2009 (UTC)
62. Support purging linkcruft. Let bot do it. bishzilla _ROARR!! 11:23, 27 June 2009 (UTC).
63. Support. Rather painless way to clear up pointless links in articles. – Joe N 22:13, 27 June 2009 (UTC)
64. Support. Seems well planned so that false hits should be minimal. • Anakin ^(talk) 05:05, 28 June 2009 (UTC)
65. Support, for reasons stated by Apoc2400 above and in consideration of the proposal. --JBC3 (talk) 07:31, 28 June 2009 (UTC)
66. Support --NE2 10:33, 28 June 2009 (UTC)
67. Support. -- Magioladitis (talk) 12:19, 28 June 2009 (UTC)
68. Support Sarah777 (talk) 01:49, 29 June 2009 (UTC)
69. Support. Well planned, well thought out, low false positives, and finally progress in that area. Headbomb {^ταλκ_{κοντριβς} – WP Physics} 19:35, 29 June 2009 (UTC)
70. Support yes please. Happy‑melon 22:17, 29 June 2009 (UTC)
71. Support per Headbomb. -- Quiddity (talk) 03:04, 30 June 2009 (UTC)
72. Support Long overdue: The community has come firmly against date linking for autoformatting, and has agreed that date linking outside that should only be done in very specific circumstances: Ones that a bot can easily avoid. Furthermore, in the exceedingly rare cases where linking the date is useful, and got caught by this bot, it's almost never useful enough that it being accidentally removed would be a serious problem. Furthermore, autoformatting breaks pages, by making it uncertain what text will be shown to someone. It was a hgorrible idea, and should be removed as quickly as possible - this will allow the code to be deleted, one would hope. Shoemaker's Holiday (talk) 03:37, 30 June 2009 (UTC)
73. Support Fixes overpresence of useless links in articles. Orderinchaos 19:20, 30 June 2009 (UTC)
74. Support. 99.9% of date links are improper anyway. The overuse of date linking is annoying and adds to the sea of blue links that already flood many articles. Overlinking is an annoyance and date links just add to the problem. Niteshift36 (talk) 02:57, 1 July 2009 (UTC)
75. Support. Appears to be a cautious approach which will have minimal false positives. Plastikspork (talk) 05:32, 2 July 2009 (UTC)
76. Support. All has been said. It only remains to be done. Finell (Talk) 22:26, 3 July 2009 (UTC)
77. Past due The arguments for this have been made repeatedly and are extremely persuasive. Using a bot in the manner described will overcome many of the problems raised by critics. I'm also a writer of FAs who's been frustrated in the past at the need to include daft links to meaningless dates and I cannot wait to see them gone. --Dweller (talk) 08:53, 6 July 2009 (UTC)
78. Support In exactly the form proposed, and with the understanding that the bot will bypass triple dates that include a piped link. Any changes, and, well… No, seriously, this is a calm, reasoned, and measured approach to moving forward on removing links that can reasonably be assumed as being for no other purpose than autoformatting. I wish there was more community contribution to this specific discussion, but c'est la vie. Full marks to everyone involved in putting this together. Mlaffs (talk) 23:47, 6 July 2009 (UTC)

Oppose

1. Oppose Something as controversial as date delinking should be done by individual editors (if at all). Cedars (talk) 12:00, 22 June 2009 (UTC)
2. Oppose. The recent RfAR shows that there is no consensus to delink dates, let alone to allow a bot to do so. Plus there has already been discussion about how linking dates is important in some circumstances. A bot couldn't decide when it is and isn't appropriate. Who then was a gentleman? (talk) 17:11, 22 June 2009 (UTC)
3. Oppose. Leave this to editors, as Cedars says. 84user (talk) 17:50, 22 June 2009 (UTC)
4. Oppose, the previous flamewars show that this area is one in which automatization would be great source of unnecessary drama. Titoxd^{(?!? - cool stuff)} 20:32, 22 June 2009 (UTC)
5. Oppose as unnecessary. The consensus was that autoformatting was unhelpful and that dates should only be linked when relevant. As a bot can only use rough heuristics to judge relevance, unlinking dates should be left to editors to do by hand, just as linking for autoformatting was. Eluchil404 (talk) 04:20, 23 June 2009 (UTC)
6. Oppose. I am really confused, aren't we just waiting for a technical solution that allows autoformatting + optional linking? Per Titoxd, a bot will cause too much drama.--Commander Keane (talk) 10:10, 23 June 2009 (UTC)

   Some people are waiting for better autoformatting, others think it is a solution in search of a problem. This RfC will determine which attitude prevails. --Jc3s5h (talk) 17:05, 23 June 2009 (UTC)

7. Oppose (changed vote). The proposer's lack of understanding of the end cases suggests this bot may damage articles, even in cases where the unlinking is clearly (and can be seen to be clearly by the bot) a good idea. I may change my !vote when the code is published, but there seems to be a likelyhood of significant damage caused by the bot. — Arthur Rubin (talk) 17:18, 23 June 2009 (UTC)

   Do you mean cases like [[April 4]] [[2004 in Sports|2004]]? --Apoc2400 (talk) 10:29, 25 June 2009 (UTC)

   I too would like to know more about what you mean here. – Quadell ^(talk) 12:56, 25 June 2009 (UTC)

8. Oppose. This has ben done so many times please don't start another one sop soon after the last has finished. I will accept any consensus or style but I am not gonna argue it and would rather it not started. It seems late to take it mosnum when there are already opinions here. SimonTrew (talk) 17:53, 23 June 2009 (UTC)
9. Oppose. The use of a bot for something which is unnecessary to perform edits to nearly all articles is clearly a waste of time and resources. If dates are to be unlinked then this should be done when other productive edits are made to the article and with clear regard for those dates that are to be linked in an article. Keith D (talk) 21:23, 23 June 2009 (UTC)
10. Oppose. I am not in favour of linked dates, and have spoken out against them in previous RFC and discussions. I even use the banned Lightmouse's very useful script for delinking dates in articles (importScript('User:Lightmouse/monobook.js/script.js'). However, the use of a bot to unlink dates was a cause of concern, and may be again. The ArbCom decision on mass delinking is: "All mass date delinking is restricted for six months. For six months, no mass date delinking should be done until the Arbitration Committee is notified of a Community approved process for the mass delinking." which suggests that careful thought should be put into it. This proposal seems too soon, and I am concerned that User:Greg L is taking part in the discussion, given that he is "topic banned indefinitely from style and editing guidelines, and any related discussions." - this discussion is flying close to the spirit of that ban. I feel this proposal is too soon and a little unwise, and may lead to drama. I feel that a proposal to mass date delink should carefully consider drama implications, and avoid the use of bots which will unlink dates that some may consider to be usefully linked. There is no hurry here - it is haste that led to this problem in the first place. Let people unlink dates in articles they edit, using Lightmouse's script, or any other useful gadget that somebody can devise. But automated edits in such a contentious and emotionally charged area is going to end up in trouble. SilkTork *^YES! 00:20, 24 June 2009 (UTC)
11. Oppose A recent RFC earlier this year made it clear that how useful a date link is (as opposed to autoformatting) depends on the context. A bot can't judge context the way a real editor can. - Mgm|^(talk) 07:48, 24 June 2009 (UTC)
12. Oppose - This would result in the articles on years (like 2004) becoming orphaned. Unlinking should be done by a person who can determine where it is and is not appropriate to link a date. --Joshua Issac (talk) 23:01, 24 June 2009 (UTC)

    Far from it. It has many thousands of links now, and a tiny fraction of them are relevant. The chance of a relevant link to 2004 coming just after a link to March 16 is minuscule. --Apoc2400 (talk) 23:47, 24 June 2009 (UTC)

    Note, as Apoc described, that the bot would only unlink autoformatted dates (i.e. [[Month day]], [[year]]), not date fragments. Dabomb87 (talk) 00:01, 25 June 2009 (UTC)

13. Oppose I never want to see a date written as "June 25, 2009" so until we have a working autoformatting system that does not need linking, stick to what we have. — RHaworth (Talk | contribs) 01:38, 25 June 2009 (UTC)
    • "I never want to see..."—so I take it you never go near articles (such as—to take a completely random example—this page) that have hard-coded dates in the mmm dd, yyyy format? (I'm aware that my example page has nothing to do with your username, but I was just put in mind of Rita)  HWV258  01:46, 25 June 2009 (UTC)
14. Oppose per Arthur Rubin. Mr.Z-man 02:14, 25 June 2009 (UTC)
15. Oppose per Arthur Rubin and Titoxd. Rebecca (talk) 07:10, 25 June 2009 (UTC)
16. Oppose (moved from support) I support this in theory, but there seem to be considerable technical questions surrounding the implementation of this bot which are, as yet, not sufficiently addressed. I agree with some of Arthur Rubin's concerns. AKAF (talk) 10:35, 25 June 2009 (UTC)
17. Oppose. MOS:UNLINKDATES strikes me as inherently very subjective (should not be linked unless their content is germane and topical to the subject) and thus entirely inappropriate for automation. I don't want a bot presumptively judging against the appropriate linking of dates on a huge variety of articles such as John the Baptist, Declaration of Independence, or Bastille. Even if a list of exceptions could be compiled, that compilation would be itself an undertaking both monumental and litigious; and would take place in bot administration space rather than article or article talk space, and thus be invisible to the majority of concerned editors. - Smerdis of Tlön (talk) 15:40, 25 June 2009 (UTC)

    The exclusion list would look something like this, and I'm it would be widely publicized (notifications at all major noticeboards and such, and probably a watchlist notice). Users would be given a month to add articles to the exclusion list. Also, there are very few instances, if any, in which both the month-day link and the year link would be relevant; the bot would only target these type of formatted dates, not date fragments. Dabomb87 (talk) 15:50, 25 June 2009 (UTC)

    I hope you don't feel badgered, but I would like to answer too. Very few of the currently linked dates are relevant. The number of cases where triple date like 1 January 2009 should be linked must be miniscule. I have not seen a single case. --Apoc2400 (talk) 15:59, 25 June 2009 (UTC)

    You still will need to exclude entire categories of articles; everything in Category:Saints or List of saints, just for example. Every such article ought to link to a calendar date, even if the years aren't important. A bot won't be able to tell why a date ought to be linked at Battle of Puebla but not at Battle of Tacámbaro, or at United States Declaration of Independence but not Constitution of the United States. These are all matters requiring some discretion and understanding. - Smerdis of Tlön (talk) 17:43, 25 June 2009 (UTC)

    Correct me if I'm wrong, but is your objection based on the argument that the article Battle of Puebla should have a link to the article May 5? If so, I'd recommend you actually take a look at "May 5" and see if you can find any value a reader might derive from going there from "Battle of Puebla". There's already a perfectly good link in the next sentence from "Battle of Puebla" to Cinco de Mayo - why dilute it with an irrelevant link to a list of trivia that happened to occur on the same calendar date throughout history?

    I gather that the reason you dislike the links is because you subjectively dislike the almanac articles they link to ("irrelevant link to a list of trivia"), but no consensus has ever been mustered to mass delete the date articles; delinking is a sort of blockade. Yes, any article that's connected to an annual observance ought to also link to the calendar date. - Smerdis of Tlön (talk) 17:09, 26 June 2009 (UTC)

    Not at all. Firstly, I do like the almanac articles (although I think they should be renamed to "List of things that happened on/in X day/year" until they reach the quality of 1345). Just because they are trivia doesn't mean I don't like them. What I don't like is a link to them when it is irrelevant. It does no service to the article where the link is made. Secondly, nor am I proposing deletion of such articles. They are start-points for better things, but while they remain irrelevant, don't link to them. Finally WP:DATEPOLL gave the option for commemorative links: 18 supported that; while 256 preferred "References to commemorative days (Saint Patrick's Day) are treated as for any other link." So while I respect your desire to make an exception to the relevance criterion, I don't see much consensus for it. --RexxS (talk) 19:36, 26 June 2009 (UTC)

18. Oppose Not convinced that getting a bot to do this is a good idea. Personally I am apathetic either way on date links (never contributed to any of the RFC's before) but feel that many editors who are even peripherally aware of the controversy that has already taken place over this area, would be put off from adding articles to the exclusion list for fear of getting jumped on by those who support removing pretty much every date link. Discussion over whether a date link sould be in place should happen on the article talk page where those who contribute/watch the page can take part, not at a centralised list where it is those who are interested in date links and not the article, who would discuss the issue. Also if the page is not on the exclusion list, editors who are not really informed about this discussion will feel constrained from challenging the bot's edits, even if the link is "germane and topical", as obviously the bot would not have been approved unless the edits it is making are always correct. Davewild (talk) 18:31, 25 June 2009 (UTC)
19. Oppose per Davewild's very reasonable and thoughtful comments. ShondaLear (talk) 20:10, 25 June 2009 (UTC)
20. Oppose About a year ago I asked whether birth/death dates in biography articles should be formatted or not, & never received a sensible answer. I did have a slab of text (which was alleged to be policy) cut-n-pasted at me, was told auto-formatting was deprecated, as well as received a number of memorable but unhelpful responses, but never noticed anyone actually reading what I wrote, considering my points in favor of it, & answering my question. Seeing how this would be the largest single exception to the "linking to dates in discouraged" rule, I believe it is reasonable to wait on this until we discuss this exception to the rule. -- llywrch (talk) 21:48, 25 June 2009 (UTC)
21. Oppose The problem with unlinking full dates (MDY) is that one of the two links (either MD or Y) may be a useful link, so delinking them all the way is problematic. Furthermore, I think the idea that articles like 25 May should be changed to List of events that happened on 25 May or something like that is a good idea, but coming up with some criteria for inclusions in such articles would need to be constructed. Linking to such articles would be fine as long as they were included in the list; furthermore, delinking after such articles are created would be easy to do as those that didn't link to the article directly would quickly be identified. Lastly, user pages should not be changed as some people simply like them; while the practice is deprecated, it doesn't mean that it is banned or use in such a manner is frowned upon. — BQZip01 — ^talk 16:45, 27 June 2009 (UTC)
22. Oppose The main (only?) reason for not advocating date linking is due to a deficiency in the wiki software not correctly localizing the dates. Since this deficiency in the software could be removed, I do not agree that using a bot in this way is desirable. Bots IMO should be reserved for non contentious edits.- (User) Wolfkeeper (Talk) 17:00, 29 June 2009 (UTC)
23. Oppose I regard this as the latest in a series of attempts to remove date autoformatting without having established a consensus. A lot of us still want the ability to autoformat dates. The software still has the ability, so we're using it. I say let those who oppose the current mechanism wait until a replacement exists. In the meantime, quit making articles uglier, whether manually or with bots! Ntsimp (talk) 18:21, 29 June 2009 (UTC)

    Could you please point to any of the myriad discussions in the last year on this point that did not come to a consensus against autoformatting? Phil Bridger (talk) 18:57, 29 June 2009 (UTC)

    The only discussion IMO that drew a wider group than just the MOS regulars was the big one, where a huge minority argued in favor of autoformatting. The attempts since then to enforce the tyranny of the majority on this point have met with considerable resistance, yet the view inexplicably persists that consensus has been reached. Ntsimp (talk) 20:12, 29 June 2009 (UTC)

    Then you seem to have missed WP:Manual of Style (dates and numbers)/Date Linking RFC, where 247 supported the proposition Dates should not be linked purely for the purpose of autoformatting (even though in the past this was considered desirable), with only 48 opposing. That's the consensus that exists. Hope you find it a little less inexplicable now. --RexxS (talk) 20:41, 29 June 2009 (UTC)

    That discussion produced a huge majority against the current system of autoformatting, which is what this proposal addresses. If any other system of autoformatting comes along and gets consensus then the work that has gone into bot programming for this proposal will mean that it can be implemented easily. This RFC is about this autoformatting method, not about some pie in the sky. Phil Bridger (talk) 21:06, 29 June 2009 (UTC)

    Ntsimp and RexxS, for what it's worth, you've both linked to the exact same RFC — ironic, no? The most recent RfC is actually this one, which indicated that while there's a substantial minority (40% on a straight headcount basis) that are in favour of the general concept of autoformatting, there's very little support for dates to be linked unless they're relevant. Let's please keep in mind that those are two separate concepts, the conflation of which is a large reason for why we've ended up in this mess. Mlaffs (talk) 21:10, 29 June 2009 (UTC)

    (ec) This current RfC ignores that the current method of date autoformatting can be tweaked to prevent links from being generated and can be changed to display a default format for unregistered users. It also prefers simply delinking the dates instead of using the Werdna's {{#formatdate}} parser function. It sure sounds like it's anti-autoformatting. —Ost (talk) 21:21, 29 June 2009 (UTC)

    Having links not actually display as links would be very strange, and I am sure there is very little support for that idea. The community has showed it is against the current system of autoformatting many times during the past year. I am not going to comment on any future system until I see it. I am not convinced that wrapping all dates in {{#formatdate}} is worth it for the very little benefit. --Apoc2400 (talk) 23:16, 29 June 2009 (UTC)

    I can understand your point about using square brackets without linking and that {{#formatdate}} can be unwieldy, but the syntax can really be changed as seen fit within reason if the community wished. If a bot is going to unlink dates, it can change them to another autoformatting syntax if desired. There are valid concerns for or against autoformatting and I thought the community was expected to revisit this issue without the parties barred by ArbCom. You may be correct that there is lack of support for the feature, though I am less certain. For better or worse, this RfC will sweep the issue under the rug if it doesn't bring the issue into the spotlight when all dates are changed. —Ost (talk) 14:37, 30 June 2009 (UTC)

    Pardon me if I'm being ignorant, but I don't think removing date links precludes users and developers from continuing to come up with a better system of autoformatting that can garner community support. What's to prevent another bot run in the future if we want to implement a new system of autoformatting? Dabomb87 (talk) 14:40, 30 June 2009 (UTC)

    (<--)I assumed it would be easier to do in one pass rather than create a mess of dates to be standardized for the article. I would think this more true now since the present syntax likely makes it easier to parse for dates. The community will always struggle to have support if they can't get the dissenters to clearly express their concerns or to contribute to a spec that lists the features that are believed to be ignored by date autoformatting implementations.

    You are correct that a new system may be developed and implemented after a bot unlinks the dates, but there is WP:NODEADLINE for when dates need to be cleaned up and it is likely beneficial to discuss options with the new participants dialogue. I agree with comments from users like User:Anomie who thinks this is too soon after the ArbCom and WP:DATEPOLL. I also think that an RfC that specifically targets dates linked for the purpose of autoformatting should discuss the alternatives to having a bot merely remove the current syntax and that it's worth the extra time to explore options. However, I also acknowledge that I am likely in the minority with many editors wanting to be done with the issue and I do appreciate the spirit of resolve in the RfC. —Ost (talk) 17:00, 30 June 2009 (UTC)

    To be accurate, the mess of dates has already been created by the original editors, not by this bot. Also, it is not the case that the present syntax makes it easier to parse for dates: a two-step process now; a one-step if this bot runs. See #General comments --RexxS (talk) 20:42, 30 June 2009 (UTC)

    I guess it depends what you're trying to fix. If you're trying to fix all dates in an article, you may be right that a second bot is beneficial. If you only plan on targetting linked dates, I don't understand your logic that it's a two-step process if autoformatting but only a one-step process if the dates are unlinked. Likely parsing will be done with regular expressions to find what looks like dates and whether the dates are bracketed or not, the bot would have to find the dates to edit them. The unlinking bot will already be a two step process: looking for the dates (step one) and unlinking them (step two). I assumed step two would assure that the dates would be changed into an appropriate—though perhaps inconsistent—format and not leave oddly formatted dates like 200906-30 (which some reason now renders incorrectly even with autoformatting). Thus, formatting the date can be included in step two at the minimal expense of using a new syntax instead of simply remving brackets and reconfiguring. This prevents having to find the dates again with a new script that has to look both at linked and unlinked dates. That being said, I'm not championing autoformatting and I don't know if anyone will if this RfC succeeds; I venture it will be forgotten about after this RfC removes its use from articles and that's probably what is irking me most. If this is supposed to be the end-all of how to deal with linked dates, it should first be decided how dates will be handled moving forward. I don't really care if it's one or two bots are used, but I would like to know the community's opinion on how dates should be formatted moving forward. —Ost (talk) 22:16, 30 June 2009 (UTC)

    /*This thread really ought to be moved */ What to fix is a potentially thorny problem. Bearing in mind that many were involved in an ArbCom case because of contested consensus, it is surely safest to fix only what seems likely to have consensus. I believe that "Dates should not be linked purely for the purpose of autoformatting" has consensus. An arbitrator, John Vandenburgh, has stated on this page "the deprecation of autoformatting has been accepted as consensus since August last year" so I feel less insecure about proposing FullDateDelinkingBot to fix the deprecated current autoformatted dates. (As an aside, I believe that there is also consensus that date fragments should only be linked if they are germane and topical, but that is unlikely to be fixable by bot.) DateConsistencyBot would have the problem of deciding what format to regularise to, and quite possibly may not be implementable. So, I hope you can see why this proposal is being restricted to a very narrow remit. I'm sorry I was not clearer when I used the phrase "two-step process" - I was referring to the process this bot needs to identify a full autoformatted date: (1) find two pieces of linked text (2) check if that text is a date. I hope it now makes sense that, after the bot is run, identifying a full date would require only the second step. Detail 7 of the proposal states "The bot will also fix uncontroversial errors such as missing commas and spaces", so I think it will fix the majority of malformed dates such as June 302009 ([[June 30]][[2009]]) and leave the ones where it is unsure - those are probably jobs best done by hand anyway. As for the community's opinion on date formats moving forward, I expect that the consensuses I referred to above, along with WP:Mosnum#Full date formatting, is all we can expect to have consensus on. I doubt that anyone will be other than ultra-cautious with proposals, following the chilling effect of ArbCom's recent sanctions, so I would expect progress to be slow and move in very small steps. --RexxS (talk) 00:20, 1 July 2009 (UTC)

24. Oppose - We've seen where large-scale date delinking leads. This isn't an urgent problem. Spend your energy improving articles, not fighting over formatting. Guettarda (talk) 02:16, 2 July 2009 (UTC)

    To your first point: true, but that was large-scale date delinking when many believed there was no consensus for it; I think we have sufficient consensus on date delinking now. To your second point: no, the problem is not "urgent", but I think overlinking (which dilutes the important links) and the inconsistencies that most of our readers see are issues that should be remedied. Dabomb87 (talk) 02:38, 2 July 2009 (UTC)

    I'd be amazed if there were no ill feelings left by the last round of date delinking. It generated a lot of bad blood, and involved a lot more people than actually were dragged in front of the arbcomm. New votes don't erase bad blood. So from that perspective, I think this is "finger in your eye" kind of unconstructive. Don't know, but that's the feeling I get. As for overlinking - that's a problem in a tiny percentage of articles, and if there are so many links in an article that date links are a problem - then there are too many links in that article. That's a problem that only exists in a tiny handful of articles. The cost of running a bot like this is that it renders your watchlist useless. If a human editor delinks dates in the process of otherwise editing an article, you expect that s/he will also have removed obvious vandalism. If you edit an article on my watchlist, I don't have any reason to check whether there was vandalism before your edit. But when a bot does it, you know there's no human checking for vandalism. So there's a real cost to having a bot flood your watchlist. Overall, I see non-trivial costs that generate a trivial benefit. Doing this is creates a net negative, gets us further away from our goal of writing an encyclopaedia. An aesthetic benefit, with a real cost in quality and community cohesiveness? Nah. Guettarda (talk) 03:27, 2 July 2009 (UTC)

    To clarify further, per this question: It's true that not everyone who edits articles catches obvious vandalism, but many do, and others make it clear from their edit summaries that they're doing automated edits. Most of the time, when a lower traffic article pops up on my watchlist (and I notice it), I look at who the editor is, and what the edit summary says. And I then decide whether to investigate further - hover over the diff link and it shows you the edit; hover over the history link, and you see recent edits. Most of the time that's enough to tell me whether I should check for vandalism. But when your watchlist is flooded by a bot, you can't check all the articles. And, quite frankly, it's easy to miss the human edits interspersed among the robot ones. So, for a while, the functionality of your watchlist is degraded, and vandalism is more likely to creep in. That's true of any bot edit, of course, or even of many AWB runs. Not good, but at least it's usually an improvement to the enecyclopaedia.

    The proposed task is almost entirely cosmetic. If cosmetic changes are to be made at a potential cost to encyclopaedic value, I'd say it's a BAD THING. And this isn't just a matter of letting some vandalism creep into a few obscure articles - many of these "obscure" articles are top-level country articles. So I see plenty of downside to this idea, and no real upside. If enough people feel strongly about delinking dates, it could be done manually. And that would add a much-needed pair of eyes to all sorts of obscure articles. That would be a GOOD THING.

    This option produces a real cost as a side-effect of a trivial benefit. Of course, the effort is trivial as well. Doing it manually produces a real benefit (articles that still have dates linked probably need some more eyes) as a side effect of a trivial benefit. Granted, it takes some effort. But so does writing the world's largest encyclopaedia from scratch. Sounds like a pretty daunting task to me. But we've done it. Guettarda (talk) 14:26, 2 July 2009 (UTC)

    Not trying to change your opinion or badger you, but it is possible to hide bot edits on your watchlist. Dabomb87 (talk) 15:24, 2 July 2009 (UTC)

25. I had intended to remain neutral but a lot of reading of the above conversations and a lot of thought in the matter keep leading me to the same conclusion : the benefit gained by such a bot does not outweigh the costs. Date linking is a style issue - it is not a content issue, it is not a policy issue, it is simply not an urgent problem. Previous attempts to enforce the MOS via bot have resulted in grief. I concede that the vast majority of the edits made by such a bot will go unnoticed and uncontested, but all it takes is a handful of editors who feel maligned by the bot's action to start yet another firestorm over something that is simply not worth the trouble. Shereth 15:29, 2 July 2009 (UTC)

Neutral

# Although I still see benefits for the autoformating in terms of emitting metadata, and I do not see a clear consensus against all forms of autoformatting, I think this a reasonable proposal. I think it's a little early, but it seems well-considered. I would like to see code, as LightBot claimed to be doing this, but mangled [[1 January]] [[1993 in sports|1993]]. Resolved, sorry. — Arthur Rubin (talk) 13:46, 22 June 2009 (UTC). Changed to Oppose.

1. Support in principle, but I can't see the exclusion list having the desired effect, and there will be a lot of re-adding of those MOSNUM-compliant links after the bot has done its thing. haz (talk) 14:00, 22 June 2009 (UTC)
2. This is like a bad dream, no matter how many times this comes up it always comes back regardless of the outcome. I personnaly don't agree with it however no matter if gets approved or denied the other side will bring it back in a few months so it doesn't matter.--Kumioko (talk) 14:23, 22 June 2009 (UTC)

   Isn't this a little soon? I mean, this subject in the past has been the source of far too much bitterness and contention, has led to the extremely unfortunate issuance of bans, topic bans, admonishings and other such actions against otherwise productive editors and frankly is just not worth the trouble that it seems to perennially generate. Shereth 14:28, 22 June 2009 (UTC)

   • We have to seperate issues of what serves our readers and what affects the editors. In this case we need to serve our readers. While an occasional link here or there is inconsequential, when multiplied over thousands of articles it is a pervasive issue that seems to cause even more over-linking. -- Banjeboi 00:19, 23 June 2009 (UTC)
     • Also, the behavior of the editors involved is the primary reason for the sanctions, topic bans, etc., not the issue itself. Apoc2400 (talk · contribs) has shown that is in fact possible to create reasonable, consensus-based proposals without all the drama. Dabomb87 (talk) 00:37, 23 June 2009 (UTC)
       • Actually no, the issue causes the drama (the user behavior is a symptom of it). Sure, it is possible to have a level-headed discussion about the issue and come to an apparant consensus, and we see that discussion is proceeding nicely here. However, the scope is broad enough that inevitably you encounter editors who are "blindsided" by the bot and will cry foul; perhaps it is just me but I see these kinds of broad, sweeping enforcements of what amounts to a fairly minor problem tend to be magnets for protracted fits of drama. Shereth 14:45, 23 June 2009 (UTC)
         • I don't want to come across as patronizing, but what do you suggest we do? The community has stated multiple times that they are opposed to the linking of dates in most circumstances, so we can't do nothing. Addressing the issue by hand would take much too long. The problem is not limited to "overlinking"; it also has to do with the fact that autoformatting piggybacks on the date linking function, leading to inconsistencies for our readers and users who don't have their preferences set. Dabomb87 (talk) 14:52, 23 June 2009 (UTC)
           • My personal preference would be "do nothing", actually. Yes, the linking generally goes contrary to the manual of style and in the vast majority of cases it should be removed. That said, it is a minor problem that, in my opinion, is simply not worth the the fits of drama that enforcement via bot is likely to engender down the line. Linked dates are unecessary but not harmful. It has been my observation that bots operating on this kind of scope attract drama and often wind up being shut down, either by community consensus or when the operators simply get fed up with dealing with the fallout. Please note that I am merely expressing my opinion/concern and I have not bothered opposing this proposal - I am not involved with the creation, maintenance or operation of the bot and do not intend to involve myself in any of the future "drama" it may cause, therefore I am disinterested in the outcome - merely making a point for consideration based upon my experience for those who do intend to be involved in this. Shereth 15:24, 23 June 2009 (UTC)
             • I understand what you are saying, and respectfully disagree. However, I will not pursue the subject, since this issue has been hashed out many times before and I don't want to clog up this section further. Cheers, Dabomb87 (talk) 15:34, 23 June 2009 (UTC)

General comments

• This is the end... wait, is it, really? I can't believe the linkwars are over. Perhaps another RFC is already up, yet not on CENT. So while I definitely support grooming the articles for the sake of consistent presentation, I'd wait a few days before voting. NVO (talk) 13:10, 22 June 2009 (UTC)
  • This is hopefully the final RfC AFAIK. This RfC is different in that—aside from the fact that it is more about implementation than general opinion on linking dates—it operates under assumptions supported by consensus that previous RfCs could not. Dabomb87 (talk) 13:13, 22 June 2009 (UTC)
  • I am not aware of any other RFC about to launch. User:Ryan Postlethwaite talked about holding something on date autoformatting, but he got busy offline and is on a break. I sure don't think this will end all discussion about dates and links, but I think it will solve the mayor issue. Then the edge cases can be discussed case-by-case on article talk pages. --Apoc2400 (talk) 13:25, 22 June 2009 (UTC)
• Leaning to support but I would like to see two additional functionalities added along with one implementation detail:
  1. The ability of the bot to ignore dates wrapped in a yet-to-be-made template for dates that page editors believe should remain linked. Likely the template would demand the date in a format that the bot will not recognize so this isn't a change to the bot.
  2. The ability to recongize the presence of one of two templates (I believe they exist) that implicitly describe which date formats are to be used for an article, or barring that, examining the first date in an article to determiningg which date format to unlink dates to.
  3. In implementation, giving all editors two weeks to place the above where they feel it necessary before the bot is run loose. --MASEM (t) 13:53, 22 June 2009 (UTC)

• 1) Absolutely, if there is a template for wrapping dates that should remain linked, then the bot will respect it. I don't want to make non-trivial changes to the proposal when some have already voiced their opinion, but the bot will certainly check for a template. These is some related discussion in the archive of this page. 2) I don't know of these templates. The bot would just unlink dates to the format they are in. I think trying to automatically standardize the format on each article is too prone to errors. There is also quotations and perhaps lists and references that should sometimes have a different format from the rest of the article. It will fix completely uncontroversial errors, such as delinking [[August 5]][[2008]] to August 5, 2008 rather than August52008. 3) There will be one month for adding article to the exclusion list. This can be done at the same time. --Apoc2400 (talk) 15:13, 22 June 2009 (UTC)
• FYI, the templates for describing date formats of an article are {{dmy}}, {{mdy}}, and {{Iso}}. Dabomb87 (talk) 15:16, 22 June 2009 (UTC)

• The goal can be achieved by manually delinking the dates where the links are not appropriate, and then either adding the article to the exclusion list, or adding an appropriate form of the {{nobots}} template, which will be honored by the delinking bot. --Jc3s5h (talk) 16:19, 22 June 2009 (UTC)

• I've no major problem with an article-wide exclusion list, but per the discussion on when to link dates, I think that the bot needs to respect a per-instance exclusion more importantly than a per-article, though if it does both, great. Per-article makes sense on the chrono-related pages, while most other pages will opt (if any) for one or two selected dates - the rest should still be de-marked and remain this way. --MASEM (t) 15:07, 24 June 2009 (UTC)

The relatively few pages on the exclusion list will be under a great deal of scrutiny and it will clearly be easy to spot the dates therein that need to be delinked and those that should remain. There really is no need to complicate the bot by having it examine a page on its exclusion list in order to do a per-instance delinking. On the other hand, I'd worry about the pages using just a {{nobots}} template - perhaps we should encourage editors who place that tag to also include the article on the exclusion list? --RexxS (talk) 16:40, 24 June 2009 (UTC)

• One of the December RFCs is linked in the footnotes on the main page, but I think it would be worthwhile to link somewhere here to at least the most recent RFC, in order to provide some context. Mlaffs (talk) 14:42, 22 June 2009 (UTC)
  • Good idea. What do you think? Dabomb87 (talk) 14:46, 22 June 2009 (UTC)
    • The most recent RFC is linked in the first bullet point of the proposal, but I don't mind the link Dabomb87 added to the footnote. --Apoc2400 (talk) 15:00, 22 June 2009 (UTC)
      • Sorry — I completely missed that link in the first bullet. Apologies... Mlaffs (talk) 15:45, 22 June 2009 (UTC)

• Commons Please be aware that a similar situation is also happening at Commons, where user:Slobot is auto-converting everything to ISO 8601 numeric Y-M-D. Andy Dingley (talk) 15:36, 22 June 2009 (UTC)
  • Is the bot also unlinking dates? Dabomb87 (talk) 15:45, 22 June 2009 (UTC)
    • It appears that Commons is also translating the dates in the information box to the local language and this required ISO format input. --Apoc2400 (talk) 16:01, 22 June 2009 (UTC)
      • Commons has an obvious need for easily- and universally- (both bot and human) comprehensible dates, hence the messing I believe. - Jarry1250 ^{(t, c, rfa)} 17:06, 22 June 2009 (UTC)
    • That would depend on what "linking dates" is taken to mean. "Dates as wls" (linking content) has little relevance to Commons, however it never had a great deal of relevance to Wikipedia for that matter, and existed largely as a way of making format localization work for logged-in users. This is what Commons is now moving towards (which I personally favour), although I believe it's only(?) applied to dates within the {{Information}} template. Wikipedia has however raised objection to any sort of auto-formatting like this. Andy Dingley (talk) 17:24, 22 June 2009 (UTC)
• Question This has come up previously in similar discussions, but I think it's worth asking again with new players involved (and thank god for that!). What will the bot do in cases where points 1 and 2 of the proposal intersect? In other words, how would the bot handle [[March 15]], [[1983 in film|1983]]? I'm assuming that it would remove the first link but leave the second alone, but I'm not a big fan of assuming. Mlaffs (talk) 16:58, 22 June 2009 (UTC)

• My hunch is to agree with you and delink March 15 but keep [[1983 in film|1983]]. Or, just leave it if it is hard to detect these cases. --Apoc2400 (talk) 17:48, 22 June 2009 (UTC)

Given the background to the use of bots in this area, I would strongly urge taking the most conservative approach and leaving alone anything that is not clearly a date-triple of the three types quoted in #1 (plus the extension given in #8). If you look at it from a programming point of view, the bot will be looking for patterns that match those three forms, and I really can't see that complicating it by checking for [[March 15]], [[1983 in film|1983]] patterns would be of any benefit. Best to leave that sort of work for human editors, imho. --RexxS (talk) 23:06, 23 June 2009 (UTC)

Sure, these cases should be sufficiently few that they can be handle manually. --Apoc2400 (talk) 13:26, 24 June 2009 (UTC)

To clarify: such dates will not be touched. I think [[March 15]] is the example above should not be linked, but it is better to handle it manually. --Apoc2400 (talk) 11:39, 25 June 2009 (UTC)

• Comment The statement "The arbitration committee has recently determined that the process of mass delinking should be approved by the community." is very poorly worded, leaving the impression that arbcom is giving instruction to the community, contrary to what they actually ruled. The ruling was almost the exact opposite of this: "1.3) All mass date delinking is restricted for six months. For six months, no mass date delinking should be done until the Arbitration Committee is notified of a Community approved process for the mass delinking." LeadSongDog come howl 19:57, 22 June 2009 (UTC)
  • That's true but not the full truth. The first arbitrator vote for this remedy said: "This gives the Community a chance to implement a plan and then notify ArbCom. We anticipate that BAG will make the determination about bots and that will be the way that it will be done if implemented." The fifth says: "This is a good proposal. It creates a preventative restriction that automatically lifts upon notice of a community approved process." These were essentially the only comments. [1] Arbcom's understanding that there will be unlinking bots is less explicit in the final decision, but even there we have "2.1) Date delinking bots will perform in a manner approved by the Bot Approvals Group." Hans Adler 15:00, 23 June 2009 (UTC)
    • My point was (and remains) the the wording of this RFC implies that Arbcomm ruled in favour of mass unlinking when in fact they ruled against it. Voters should not be mislead by the RFC.LeadSongDog come howl 22:18, 23 June 2009 (UTC)
      • It was never intended to imply that. I now changed it to The arbitration committee has recently determined that any process of mass delinking should first be approved by the community. (bold shows changes) --Apoc2400 (talk) 22:33, 23 June 2009 (UTC)
        • Much better. Thank you.LeadSongDog come howl 06:56, 24 June 2009 (UTC)

• Comment In line with my comment in the baby and bathwater section, it is worth asking whether some editors who sincerely believe that readers need to veer off into a date article have taken the step of adding the date article to the "See also" section, or mentioning such a link using the {{seealso}} template? If so, the "See also" section should be added to the list of things for the bot to avoid (the template links articles without the need for brackets). Chris the speller (talk) 18:04, 24 June 2009 (UTC)

• Comment: There seems to be some confusion about whether removing square brackets from full dates will somehow make it more difficult for a future system of date sutformatting to be implemented. Surely it cannot. This bot has two levels of identification to make: text in brackets; and when that text is a date. Any future bot implementing the markup for a future DA system will need just the second part: to identify a date. In fact, one of the results of this bot running would be to regularise dates into one of three formats, by consolidating different variants such as "[[June 30]],[[2009]]", "[[June 30]] [[2009]]" and "[[June 30]][[2009]]" into June 30, 2009. This would simplify future bots' work in identifying dates, not complicate it. --RexxS (talk) 15:28, 30 June 2009 (UTC)

While the first bot may simplify the job of a second bot, this doesn't mean the the entire process is simpler. A one-pass fix seems inherently simpler than a two-pass solution. Also, the searching for bracketed dates reduces the amount of text that is possible to be a date, although I'm not sure how this affects the efficiency of parsing with regular expressions. —Ost (talk) 17:43, 30 June 2009 (UTC)

On the other hand, what are the concerns? That the bot, running on someone's pc, takes x millisec to find a bracketed date, rather than y millisec to find a date? Whether x>y or x<y isn't much of a hill of beans, is it? I agree that the total impact on the servers would be greater for two bots running, since each page has to be read (and sometimes written back) for each bot. Nevertheless, I am convinced that the developers will soon let us know if the pace of a bot has to be throttled to lower any server stress. As this can always be done, I don't see a convincing technical argument that would imply we had to incorporate DateDelinkingBot and DateConsistencyBot into one. The advantages of not doing so are outlined at #Consistency below. --RexxS (talk) 21:07, 30 June 2009 (UTC)

I consider having to make more than one edit needlessly disrupting when there is WP:NODEADLINE to fix the dates; we can do it right in one-pass. Two bots would make all articles have to be fixed twice and flood watchlists with these minor edits. But, if the consensus is that multiple edits are acceptable, I can understand your point (which I addressed a bit more above in #Oppose). —Ost (talk) 22:16, 30 June 2009 (UTC)

Thank you, the point about disruption through flooding watchlists is well-made. I actually don't think a bot could be made to make dates consistent in the way that WP:Mosnum#Full date formatting requires, so I'm less worried about disruption from a bot doing that job. --RexxS (talk) 03:22, 1 July 2009 (UTC)

Thank you too, for explaining your position here and the thread above. Your point about ensuring MOS date consistency is well taken. With the intentionally limited range of dates within the scope of this RfC and the apparent consensus against the current date autoformatting, I do see the logic in waiting for a bot run that would alter date formatting—whether that be to a static or dynamic format. Rather than targeting only a subset of dates, the date consistency bot could actually cleanup all dates in an entire article. And since such a bot or script may indeed have its own technical hurdles, taking time to gather seperate community consensus does seem prudent. —Ost (talk) 13:32, 1 July 2009 (UTC)

Ost316 wrote "Rather than targeting only a subset of dates, the date consistency bot could actually cleanup all dates in an entire article." No, it can't. The format of dates within direct quotes, examples, source code, and the like should not be changed. Due to the many ways of formatting such text, a bot will not be able to distinguish dates that should be reformatted from those that should not. --Jc3s5h (talk) 13:40, 1 July 2009 (UTC)

That's true, but to be fair, Ost316 later qualified as "such a bot or script". IIRC, Lightmouse had a script that would probably go a long way towards finding articles with inconsistent dates and highlighting them for a human to make decisions on. But that's a project for the future. --RexxS (talk) 17:23, 2 July 2009 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Proposed exceptions

Please see User:Full-date unlinking bot#Proposed exceptions if you are interested in excluding a page or multiple pages from the bot's operation. —harej (talk) (cool!) 08:10, 9 July 2009 (UTC)

The above link to ...#Proposed exceptions seems to be a dead link. Apparently, the list that used to be there, is now at User:Full-date_unlinking_bot#Exceptions. --Mike Schwartz (talk) 09:05, 23 March 2010 (UTC)

Delinking dates

I was asked to post this notice here and several other places. I have started a discussion at Wikipedia talk:Manual of Style/Dates and numbers#Delinking dates on delinking dates using an AWB bot. Input is always appreciated. Hmains (talk) 19:55, 17 June 2012 (UTC)