Wikipedia:Bots/Requests for approval/Lightbot 3

Add links
From Wikipedia, the free encyclopedia
The following discussion is an archived debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA. The result of the discussion was Revoked. In June 2009, ArbCom restricted Lightmouse from performing any automated tasks. They are now considering lifting this restriction, subject to (re)approval of any tasks by BAG. To avoid any confusion, I am marking Lightbot's old approvals as "Revoked". Anomie⚔ 17:38, 13 July 2010 (UTC)[reply]


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 Approved.

Lightbot 3[edit]

Operator: Lightmouse (talk)

Automatic or Manually Assisted: Automatic

Programming Language(s): AWB, monobook, manual

Function Summary: Janitorial edits mainly to units and dates.

Edit period(s) (e.g. Continuous, daily, one time run): Continuous

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

Function Details: The previous request is at Wikipedia:Bots/Requests for approval/Lightbot 2, which itself was an extension/clarification of Wikipedia:Bots/Requests for approval/Lightbot due mainly to the suggestion by some editors that solitary years are not covered by 'unlinking date fragments'. This further request is an extension/clarification due mainly to the suggestion by an editor that centuries are not covered by 'unlinking date fragments' see comment on my talk page.

  • I would like to make it explicit that I will be editing dates in a variety of forms.
    • A 'date' is any sequence of characters that relates to time, chronology, or calendars. This includes but is not limited to seconds, minutes, hours, days, weeks, fortnights, months, years, decades, centuries, eras, and can be in any sequence or format.
    • Edits may add, remove or modify the sequence or format of dates.
    • Edits may add, remove or modify templates that involve dates.
    • Edits may add, remove or modify autoformatting. For example, where autoformatting is invalid, broken or itself breaks a date for readers. Struck text replaced with:
    • Edits may add or modify autoformatting. Edits may remove autoformatting where it is invalid, broken or itself breaks a date for readers.
    • Edits may add, remove or modify links to dates.
  • I would like to make it explicit that I will be editing units of measure in a variety of forms.
    • A 'unit of measure' is any sequence of characters that relates to measurement of things. This includes but is not limited to units defined by the BIPM SI, the US NIST or any other weights and measures organisation or none at all. This includes but is not limited to time, length, area, volume, mass, speed, power.
    • Edits may add, remove or modify the metric or non-metric units.
    • Edits may add, remove or modify the sequence or format.
    • Edits may add, remove or modify templates that involve units.
    • Edits may add, remove or modify links to units.
  • I would like to make it explicit that I will make other edits.
    • These will usually be minor improvements that are often done by other editors or are part of general MOS guidance.
    • These will usually be incidental to the main motivation for the bot which is units and dates.


Discussion[edit]

  • Lightmouse has proved himself to be an excellent bot-manager in his awareness of editorial sensitivities, his polite interactions with editors who query the workings of the bot, and his prompt and efficient responses to feedback on possible issues with the functions of the bot. He appears to be entirely motivated by a wish to improve the formatting of Wikipedia's article text, and therefore of its appearance and readability; he has applied his detailed knowledge of and respect for the Manual of Style and the other style guides and policy pages that make WP a cohesive force on the Internet. In my opinion, he has already performed an invaluable service to the project in the running of this bot. Please note that this request has been made in response to an editor who explicitly takes no issue with the removal of square brackets around "20th century", inter alia, but has a concern for procedural correctness. I support both that concern and this request for approval. Tony (talk) 12:08, 23 August 2008 (UTC)[reply]

This request seems far too broad. --MZMcBride (talk) 06:32, 1 September 2008 (UTC)[reply]

It may appear like that but please look at the previous two approvals and the 140,000 bot edits. The main problem has been that people have complained that the first and second approvals were too narrow. The key issue relates to linking of dates that will autoformat (commonly termed 'full date') and dates that will not (commonly termed 'date fragment'). The first approval for Lightbot mentioned 'unlinking date fragments' on the basis that it would unlink dates that will not autoformat. So the 140,000 edits included delinking solitary years and centuries. The second approval was sought because somebody complained that solitary years were not given as an example in the first approval, so rather than debate the Wikipedia meaning of 'date fragment', it was easier just to give solitary years as a specific example. This third approval is being sought because of an almost identical complaint: namely that centuries were not mentioned. So rather than debate the Wikipedia meaning of 'date fragment', it is easier to reword the whole thing.
I did not want to have to come back and get further approval for each type of 'date fragment' just because there is occasional misunderstanding of the term. I decided to abandon the term 'date fragments' and explain it in bullet points with repeated phrasing so that there should be no doubt. I hope that helps. Lightmouse (talk) 08:42, 1 September 2008 (UTC)[reply]
Approving this would be giving you a blank cheque to change anything related to dates and units. BJTalk 10:06, 1 September 2008 (UTC)[reply]
  • I think the whole proposal is too broad, per BJ above. In particular, the date section, which specifically says that it may remove date auto-formatting, which is a particularly contentious issue. My current understanding of MOS:NUM is that auto-formatted dates are neither required nor encouraged, nor are they prohibited. Granting this bot approval allows Lightbot carte-blanche to remove auto-formatted dates, which is a decision best left to the editors of each article rather than to a bot. — Bellhalla (talk) 11:32, 1 September 2008 (UTC)[reply]
  • On another note, my past interactions with Lightmouse (here and here) indicates, to me at least, that this sort of wide-ranging power would not be yielded with the necessary restraint. — Bellhalla (talk) 11:32, 1 September 2008 (UTC)[reply]
    • Bellhala, MOSNUM says that the use of autoformatting is "deprecated". Can you be specific about this feeling that the application of the bot would be "too broad"? In what ways? And please explain why the application of our Manual of Style to articles should be thought with suspicion, rather than as a service to editors as a whole. Tony (talk) 08:49, 2 September 2008 (UTC)[reply]
Lightbot already removes autoformatting where it is invalid (e.g. date ranges), broken (e.g. [[December 25|25]] or breaks a date for readers. You will see plenty of evidence of removing invalid autoformatting or autoformatting error fixing in the existing 140,000 Lightbot edits. This approval does not seek approval for Lightbot to remove valid autoformatted dates. I would be happy to make this clearer by replacing:
  • Edits may add, remove or modify autoformatting. For example, where autoformatting is invalid, broken or itself breaks a date for readers.
with:
  • Edits may add or modify autoformatting. Edits may remove autoformatting where it is invalid, broken or itself breaks a date for readers.
I hope that makes it clearer. Regards Lightmouse (talk) 12:14, 1 September 2008 (UTC)[reply]
Actually, it does not make it clearer because the two sentences could be interpreted as independent instead of modifying each other. Better to state the policy as follows, "Edits may add, modify, or remove autoformatting when it is broken or breaks a date for readers." Tennis expert (talk) 08:27, 9 September 2008 (UTC)[reply]

I support this too, per Tony's reasons. I would support it just as much if it did include removing the autoformatting links, but since it doesn't, that isn't an issue for here.--Kotniski (talk) 12:45, 1 September 2008 (UTC)[reply]

I would invoke Wikipedia:BURO here but the entire bot approval process is entirely too bureaucratic. To not further the problem I dislike these "clarification" requests. We aren't here to argue over semantics, Wikipedia:MOS is the place for that. If an editor feels that the bot isn't operating within the spirit its request they can bring it up on the bot owners noticeboard and consensus can be (not) reached there. BJTalk 12:57, 1 September 2008 (UTC)[reply]
  • Can you give an expanded explanations of the following bullets? I don't want voice my concerns based on any assumptions:
  1. Edits may add, remove or modify the metric or non-metric units.
  2. Edits may add, remove or modify the sequence or format.
Thanks, —MJCdetroit (yak) 16:36, 1 September 2008 (UTC)[reply]
As you can see, I used the term 'add, remove or modify' for all instances of things that Lightbot does.I presume that you are fine with 'add' and 'modify' but worried about 'remove'. Let me state that I have no intention of mass removal of units.
As far as sequence is concerned, I have in the past corrected sequence to put source data first. As far as format is concerned, the convert template often uses a different format to raw text. For example, somebody might write Sq.Ft but the convert template uses 'sq ft'. The word 'remove' makes no sense when it comes to format, the word is just there because I was using a standard phrase.
I think we know each other's editing well enough. There is no hidden agenda here. Tell me your concerns and feel free to suggest modifications of the text. If you want me to remove the word 'remove' from both those lines, I can do that. Lightmouse (talk) 17:25, 1 September 2008 (UTC)[reply]
I think that there are very few times when a conversion should be removed (e.g. in direct quotes), so it may be better to do any removals manually and drop that part of the bullets. As far as 'sequence' is concerned, I think that swapping the order of units (if I am understanding that correctly) is a little tricky and should be left to manual edits. An example of what I am thinking of is when the article is/should otherwise be metric units first (or vise vera) but because of sources U.S. and/or imperial units are listed first. If you swap something like that, an editor's note (<!-- -->) should be left indicating which was the source unit. Also, will your bot be providing conversions from metric units to U.S./imperial units? Other than those concerns, I think you're on the right track and could get on board with this. —MJCdetroit (yak) 01:37, 2 September 2008 (UTC)[reply]
Yes, I agree with you that removal of conversions is rare so I would be happy to drop that option. I simply cut and paste the all-encompassing 'date' section to create an all-encompassing 'unit' section and so I would not have thought about sequence. So I would be happy to drop that option too. As to non-metric units, I have occasionally added them so that is why I want the option but it is not my thing. You will see me doing modifications e.g. 'MAF' -> 'million acre feet' or 'gal' -> 'US gal'. Thanks for your support. — Preceding unsigned comment added by Lightmouse (talk • contribs)
Support.—MJCdetroit (yak) 13:13, 2 September 2008 (UTC)[reply]

Lightmouse knows exactly what he's doing and what he's doing is improving Wikipedia. He is very responsive to requests and suggestions. I'm confident that he's not about to go setting his bot off to do work not supported by consensus and common sense. Calls of "blank cheque" and "too broad" should be weighted against Lightbot's contribution history: Lightmouse has earnt that broad blank cheque and will spend it wisely. JIMp talk·cont 18:04, 1 September 2008 (UTC)[reply]

I have had to repair about 5 to 10 edits by lightbot that affected articles on my watchlist. I believe the modification of incorrect autoformatted dates should be approved, but the modification of units should be a manually aided process rather than an automatic process. --Gerry Ashton (talk) 19:09, 1 September 2008 (UTC)[reply]

Comment: I have been using the script and I must say it has been quite helpful/useful, but I am still getting the hang of things. I think the bot is a good idea, but I am just not so super experienced with it all yet to strongly support, though I certainly do not object. Cirt (talk) 20:58, 1 September 2008 (UTC)[reply]

What's happening with piped links like 1939? Hiding T 10:42, 2 September 2008 (UTC)[reply]

Piped years frequently break autoformatting. I believe that linking piped years is a side-effect of linking solitary years and linking solitary years is a side-effect of autoformatting full dates. I can already see a decline in both. Piped/hidden/camouflaged/easter-egg links to years are now beginning to be more noticeable and some users/projects say that they should not be used (e.g. music project, film project - possibly). Others still like them. Your example was [[1939 in comics|1939]], Lightbot specifically avoids 'years-in-comics' and some others. I hope that clarifies it for you. Lightmouse (talk) 11:26, 2 September 2008 (UTC)[reply]

What noises are the film project making? Do you have a pointer? Hiding T 21:56, 2 September 2008 (UTC)[reply]
The topic was raised at Wikipedia talk:WikiProject_Films/Style_guidelines#Film_years. Regards Lightmouse (talk) 22:03, 2 September 2008 (UTC)[reply]
Removing so-called "piped/hidden/camouflaged/easter-egg links" should not be within Lightbot's mandate and should be removed from this proposal. Tennis expert (talk) 08:27, 9 September 2008 (UTC)[reply]
  • Neither of these beliefs is always true; for example, years of creation are uniformly linked in articles on peerages, apparently on the grounds that context will add value to the article, by giving the reader information on what King, and what Government, created the title - and their political circumstances. Septentrionalis PMAnderson 16:42, 2 September 2008 (UTC)[reply]

Is this intended to automatically delink years before 1582? Septentrionalis PMAnderson 16:42, 2 September 2008 (UTC)[reply]

I think we are getting into interesting detail but I am not sure if it is too much detail. I do not understand what you mean about articles on peerages. I looked at Thomas Thynne, 1st Marquess of Bath, Marquess of Bath, John Carteret, 2nd Earl Granville, Earl Granville, Duke of Hamilton, and Baron Carteret and all the years links looked like solitary years. I might have missed whatever 'year-in-something' you are talking about. Can you be more specific?
I don't know why you ask about delinking years before 1582. Can you be more specific?
Lightmouse (talk) 17:40, 2 September 2008 (UTC)[reply]
  • I do mean solitary years; if your generalizations are not intended to apply to them, I misunderstand what you say.
  • I bring up 1582 because there has been a suggestion to delink all dates before 1582, as a way of solving the problem of what happens if they are auroformatted into ISO. I oppose that, as unnecessary and likely to cause trouble; I do not want to see it included in the vague language of this proposal. Septentrionalis PMAnderson 17:53, 2 September 2008 (UTC)[reply]

Now it is my turn to be confused. As I understand it, the '1582' suggestion only applies to autoformatted dates because autoformatting can turn full Julian dates into non-compliant ISO. I am not aware that anybody had suggested applying the '1582' suggestion to anything other than autoformatted full dates. Can you give me a link to where this was suggested? Lightmouse (talk) 20:50, 2 September 2008 (UTC)[reply]

  • Rewrite to clarify, with examples, exactly what you want to do, and why you need a bot to do it. Septentrionalis PMAnderson 22:29, 2 September 2008 (UTC)[reply]

I did some experimenting with date and time preferences. A day and month will not, and cannot, be transformed into ISO 8601 format, because that format must have a year. In principle, a month and year could be transformed back and forth to ISO 8601 format, but that does not seem to happen. So if I did my experiments right, and if the code is stable, there is no harm in using autoformatting on a month and day (such as July 4), nor is there any harm in using it on July 1776 ( but in the latter case, it doesn't do much good, either). --Gerry Ashton (talk) 22:52, 2 September 2008 (UTC)[reply]

Thanks. Lightbot will delink solitary years and other date fragments because that is its job. Look at any of the 140,000 edits and the three approval applications. It does not delink valid autoformatted dates. I know that there are lots of people talking about the year 1582 at MOSNUM and elsewhere but that is for them, it simply is not relevant to Lightbot. Lightmouse (talk) 23:34, 2 September 2008 (UTC)[reply]
Then what are you asking for, if you're going to run the bot anyway? Septentrionalis PMAnderson 19:28, 3 September 2008 (UTC)[reply]
I am not very familiar with the system, but this looks like an extension–formalisation of the robot's mandate. Personally, I agree with Kotniski: I'd prefer to see the date links gone as well, but even without this function I am perfectly happy to support this 'bot. It's doing good work that improves, in simple little steps, article style and presentation on a wider scale than human editors ever could; with a 2,500,000-article encyclopaedia, we sure need it. Waltham, The Duke of 20:29, 3 September 2008 (UTC)[reply]
  • I would like to endorse and support the efforts of Lightmouse, and accept this Lightbot3 proposal. This bot will be a welcome addition to the automated tools at the community's disposal, taking out the drudge in compliance with WP:MOS. I believe that there are sufficient safeguards in the two Lightbots, and the third will no doubt have the same in addition to Lightmouse's commitment to rectify any systematic faults which are brought to his attention. Ohconfucius (talk) 05:01, 4 September 2008 (UTC)[reply]
  • I support this request. Teemu Leisti (talk) 01:15, 5 September 2008 (UTC)[reply]
    • I don't think this is an RFA. Do you have an argument to make, or reasoning to share with us? SQLQuery me! 05:33, 7 September 2008 (UTC)[reply]
  • I would urge Lightmouse to be much more specific in his proposal. It is generally undesirable to force enforcement of anything which is not a policy, and the MOS is a guideline which has less than full consensus, most specially because there are too few general editors who have contributed to it. The problem with such a bot is that it can end up trampling on anyone who doesn't agree. I think that a good example of this recently is lightbot's conversion of acres to km^2 (instead of hectares). In this case Lightmouse strongly defended his right to do so, which I agree with in manual editing cases, but not when a rather arbitrary decision is being bot-enforced. I would firstly urge Lightmouse to stop considering the MOS as a policy, and secondly to be conservative in the editing tasks selected.AKAF (talk) 15:43, 5 September 2008 (UTC)[reply]
Can you give a link to the edit that you are thinking of? Lightmouse (talk) 15:47, 5 September 2008 (UTC)[reply]
Here [1] is a static link to your talk page before archiving of the section titled "Overwriting convert defaults - from (acres) to (m2) instead of (ha or km2)". The thing which makes me uneasy is the assumption that lightbot will enforce a decision for which there is no consensus. I have a strong opinion on the subject, but it is the enforcement itself which I find problematic, not the direction of the decision. AKAF (talk) 09:14, 8 September 2008 (UTC)[reply]
  • Comment: Although date autolinking has been deprecated on MOSNUM, it is NOT (at this time) the case that there has been consensus approval for the mass deletion of existing links. Accordingly, it should not be so used until there is such consensus. Askari Mark (Talk) 00:38, 6 September 2008 (UTC)[reply]
To Askari Mark: This approval does not seek approval for Lightbot to remove valid autoformatted dates. You will see on this page that I have said so more than once. I am happy to say it again. Lightmouse (talk) 08:32, 6 September 2008 (UTC)[reply]
The wording "Edits may remove autoformatting where it is invalid, ..." do not make this clear, particularly when there are editors who maintain that since DA is deprecated, existing links should be removed in an automated fashion as they are no longer "valid". Hence my comment. Askari Mark (Talk) 03:21, 11 September 2008 (UTC)[reply]
There are a few examples of where autoformatting will work but should not have been applied. There are *lots* of things like: [[April 17]]-[[19|April 19]]. In that case autoformatting causes an error and will look like '17 April-19. That is what I mean by invalid. There are so many weird and wonderful ways in which editors link to dates and only a small number are valid. Listing all the invalid forms would be a full time job in itself. I would be happy to reword it but I can't think of a better form of words. Suggestions would be welcome. Lightmouse (talk) 10:55, 11 September 2008 (UTC)[reply]
Perhaps you meant "[[April 17]]-[[April 19|19]]"? In any case, I agree with Lightmouse's point; date links can be used in five or six valid ways. Have a look at the following for some wrong ones:
Brace yourselves...

Examples which:

...render properly but should not be used
  • [[7 April]], [[2008]] and [[April 7]] [[2008]]
  • [[7 April]][[2008]]
  • [[07 April]] [[2008]]
...do not render properly
  • [[7 April 2008]]
  • [[7]] [[April]] [[2008]]
  • [[April 7]], 2008
  • 7 [[April 2008]]
  • [[7 April]] [[2008 in film|2008]]
  • [[7 April]] [[2008 AD]]
  • [[7th April]] [[2008]]
  • [[7 Apr]] [[2008]]
  • [[2008]] [[April 7]]
  • [[2008-4-7]]
  • [[2008]][[04-07]]
  • [[04-07-2008]]
  • [[07-04-2008]]
  • [[7/4/2008]]
  • [[7 April]]–[[9 April]] [[2008]]
  • [[April 7-9]] [[2008]]
  • [[April 7|7]]-[[9 April]] [[2008]]
  • 7–9 [[April]] [[2008]]
Other wrong practices
  • (only first instance of each date linked)
  • (only first instance of each year linked)
Off the top of my head. And now think of the combinations. Waltham, The Duke of 18:40, 11 September 2008 (UTC)[reply]
  • I agree with AKAF but it appears that I am too late because Lightbot is already doing its thing, going rapidly down the alphabet (at "C" now). When did this discussion end so that Lightbot could be turned loose? Also, it appears that Lightmouse himself is removing autoformatted dates (even if Lightbot is not). Tennis expert (talk) 00:27, 7 September 2008 (UTC)[reply]

Lightbot has two approvals for its work and has done over 140,000 edits. You are being invited to comment on this third approval request. This does not seek approval for Lightbot to remove valid autoformatted dates. You will see on this page that I have said so more than once. I am happy to say it again. Lightmouse (talk) 01:33, 7 September 2008 (UTC)[reply]

    • Perhaps Lightbot won't do that, but you are more than willing to do it yourself. And because you run Lightbot and turn it on or off without much explanation, I believe Lightbot has too much power. Just the "over 140,000" edits is scary. You seem to believe that's a badge of honor or something. Tennis expert (talk) 06:57, 7 September 2008 (UTC)[reply]
      • Well, since the vast majority of those edits have improved Wikipedia, I think it is something for Lightmouse to be proud of. And if he is removing "autoformatted" dates manually, then that's also to his credit - that style is now deprecated.--Kotniski (talk) 09:23, 7 September 2008 (UTC)[reply]
  • Support --John (talk) 01:51, 7 September 2008 (UTC)[reply]
    • This is not an RFA, please come back with a coherent argument. Frankly, I feel as many above, that this request is a little too broad, and, would politely request, that the requesting user refine this (into multiple parts if needed for clarity), into what you actually plan on doing. SQLQuery me! 05:30, 7 September 2008 (UTC)[reply]
      • Thanks for reminding me that this is not an RfA. I agree with Tony at the top of this discussion section, so my support may be seen as endorsing Tony's coherent argument. Hope that makes it clearer. --John (talk) 02:52, 8 September 2008 (UTC)[reply]
  • Oppose Lightbot has proven that it is a bulldozer that runs over everything in its path and puts human editors in the almost impossible position of undoing its thousands of edits whenever it, intentionally or not, ignores consensus. And the ability for any editor to "Stop" Lightbot is meaningless, too, because Lightmouse just undoes the "Stop" within minutes of it being added. Aside from all those problems, I agree with SQL completely. Tennis expert (talk) 06:14, 7 September 2008 (UTC)[reply]
    • We don't approve requests by counting (unlike RfA *cough*), the support/oppose prefix isn't necessary. BJTalk 06:16, 7 September 2008 (UTC)[reply]

Readers should note that Wikiproject_Tennis, according to SMCandlish, is a hornet's nest involved in a dispute about naming conventions. It appears to be stinging everyone that comes near tennis articles even on unrelated issues. User:Tennis_expert is involved in that project. Some editors have asked if Wikiproject_Tennis can justify an opt out from the Manual of Style, see Wikipedia talk:Manual_of_Style#Wikiproject_Opt-outs_.28and_date-linking.29. The exception to MOS compliance is defended on the basis of preserving the 'status quo' and 'pre-existing consensus'. But the many date defects and inconsistencies in tennis articles are defended on the basis that the articles are 'work-in-progress'. Those are two mutually exclusive positions that may interpreted as a symptom of ownership. Readers may also want look in on Wikipedia talk:WikiProject_Tennis#Wikilinking_years. Lightmouse (talk) 10:31, 7 September 2008 (UTC)[reply]

May I add that User Tennis expert (whose expertise in that field I appreciate) appears to be conducting a one-person freak-out about keeping all date fragments bright blue in the tennis empire. This is a pity for the readers and the tennis editors who genuinely want to improve the appearance and readability of our strong tennis categories (something to be proud of, actually), and who want the high-value links in those articles undiluted by useless ones. It's a great pity, but no one's going to press improvements where they're not wanted, and no one's going to edit war or cause instability; that's the last thing we want. As for the bulldozer comment, well, anyone's free to revert Lightbot's good work. No hard feelings. I ask participants here to take Tennis expert's comments in that context. Tony (talk) 12:41, 7 September 2008 (UTC)[reply]
Thanks for being incivil, Tony, and saying that I'm conducting a "one-person freak-out". But it's nice to see that you don't want instability in our articles and that you're apparently conceding that Lightbot can do undesirable things. Too bad that the owner of this bot doesn't have a similar attitude. Tennis expert (talk) 18:49, 8 September 2008 (UTC)[reply]

I think this bot so far has done a very good job in the articles I have seen it perform. These are mainly tennis bio articles, where there is a multitude of these meaningless date links. By the way, everybody at the Tennis Project appear to be against the linking of dates. One editor, however, in my personal optic, is against removing links until doing so is in accordance with consensus. How and when this is formed is, however, not clear to me.--HJensen, talk 18:05, 7 September 2008 (UTC)[reply]

I'd agree with HJensen and Tony. Lightbot has made a really positive overall impact on Wikipedia and Lightmouse is a willing and diligent editor who has shown time and again that he is prepared to work with other editors to get things right. The tennis project is anomalous and its bizarre understanding of what constitutes the development of half-decent articles is borne out by the fact that, according to its own assessment subproject, it has 2 GAs (nothing featured) from nearly 6,000 articles. The Rambling Man (talk) 21:29, 8 September 2008 (UTC)[reply]

Motion to deny[edit]

  1. The bot is currently operating without issue, no editor has brought complaints regarding the bot's mandate to the communities attention via the noticeboards. Wikipedia:IAR comes into play here as well, we should not be debating over minor wording of requests.
  2. This request is extremely broad and would allow the bot to make almost any changes and still be in the bounds of the request. In my personal opinion the approval would not be valid, as it doesn't adequately describe the function of the bot.
  3. Lightmouse has canvased this request, and it has devolved into a debate of Lightmouses' character and RfA-esque voting. BJTalk 05:44, 9 September 2008 (UTC)[reply]
I support the motion to deny, although I strongly disagree with the assertion that the bot is "currently operating without issue". See, for example, this and this. Tennis expert (talk) 06:04, 9 September 2008 (UTC)[reply]
To be clear, even if this is denied Lightbot can continue operating unless consensus is established to stop it. BJTalk 17:34, 9 September 2008 (UTC)[reply]
Yes, I understand that. But at least an expansion of Lightbot's activities would not occur. Tennis expert (talk) 20:15, 11 September 2008 (UTC)[reply]


Resolution[edit]

I'd like very much to resolve this request. To that end, here is a summary of Lightbot's current approval:

  • Converting units using {{convert}}, simplify that template, correct errors with its use. This includes adding and modifying the convert template.
  • Correct any dates or other measures of time to use autoformatting properly, delinking and reformatting where appropriate. This includes ranges of dates.
  • Delink any date fragments - months, decades, weeks, etc.

To those points, this bot is, and has been, Approved.. This bot also seeks approval for these tasks: (my comments in italics)

  • I would like to make it explicit that I will be editing dates in a variety of forms.
    • A 'date' is any sequence of characters that relates to time, chronology, or calendars. This includes but is not limited to seconds, minutes, hours, days, weeks, fortnights, months, years, decades, centuries, eras, and can be in any sequence or format.
    • Edits may add, remove or modify the sequence or format of dates.
    • Edits may add, remove or modify templates that involve dates.
    • Edits may add or modify autoformatting. Edits may remove autoformatting where it is invalid, broken or itself breaks a date for readers.
    • Edits may add, remove or modify links to dates.
  • The above are all already Approved. per previous approval requests.
  • I would like to make it explicit that I will be editing units of measure in a variety of forms.
    • A 'unit of measure' is any sequence of characters that relates to measurement of things. This includes but is not limited to units defined by the BIPM SI, the US NIST or any other weights and measures organisation or none at all. This includes but is not limited to time, length, area, volume, mass, speed, power.
    • Edits may add or modify metric or non-metric units.
    • Edits may modify the format.
    • Edits may add, remove or modify templates that involve units.
    • Edits may add, remove or modify links to units.
  • Already Approved. with regards to the convert template.  Approved. with respect to any units that need to be linked or unlinked per MOS.
  • I would like to make it explicit that I will make other edits.
    • These will usually be minor improvements that are often done by other editors or are part of general MOS guidance.
    • These will usually be incidental to the main motivation for the bot which is units and dates.
  •  Approved. for MOS issues related to numbers, dates, units, and the like.
Are there any further issues that need to be resolved? --uǝʌǝsʎʇɹoɟʇs(st47) 23:44, 11 September 2008 (UTC)[reply]
Of course there are unresolved issues! See everything that has been written before on this project page. Tennis expert (talk) 06:43, 16 September 2008 (UTC)[reply]
Since I just approved a number of tasks, it should be obvious tht I'm looking for further requests for approvals, not complaints about the bot's operations (which belong on WT:BAG or WT:BRFA...). Since no such requests have surfaced,  Approved. per my previous clarification in this section. --uǝʌǝsʎʇɹoɟʇs(st47) 23:36, 21 September 2008 (UTC)[reply]
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.


Reconsider[edit]

I move the approval of this bot be revoked until

  1. A test suite is developed which will be run after every single change to Lightbot, no matter how minor
  2. A plan is presented to roll back all edits performed while the bot was malfunctioning.— Preceding unsigned comment added by Gerry Ashton (talk • contribs)

I second this; it was never clear that the intent here was to unlink all year dates whatsoever. There is an ongoing discussion as to what year links are valuable; there is no consensus that none are. Septentrionalis PMAnderson 23:16, 22 September 2008 (UTC)[reply]

I third this. I have read the above request trying to work out what was approved or not, and gave up after the third or fourth set of strike through bits. If people can't work out what was approved, a request can't be considered valid. Carcharoth (talk) 23:44, 22 September 2008 (UTC)[reply]

Fourth. The approval did not give the bot the right to remove all linked years. Corvus cornixtalk 23:45, 22 September 2008 (UTC)[reply]

"Years" discussion started at Wikipedia talk:Manual of Style (dates and numbers)#Year articles and wikilinks to year articles. Carcharoth (talk) 00:22, 23 September 2008 (UTC)[reply]
  • Bandwagon oppose to this being approved per my objections above. BJTalk 01:15, 23 September 2008 (UTC)[reply]
  • Anti-bandwagon support for the approval. There is long-standing consensus that solitary years and the like are not to be linked; this bot has long been removing such links, greatly to the good of the project. A few people's half-thought-out objections should not be allowed to obstruct this beneficial work.--Kotniski (talk) 13:38, 23 September 2008 (UTC)[reply]
    • There is no consensus that linked dates are to be delinked. And your comment about "half-thought-out objections" is itself objectionable. Corvus cornixtalk 18:32, 23 September 2008 (UTC)[reply]
      • There has long been consensus for soiltary dates being unlinked; more recent consensus for deprecating the autoformatting-type links. Sorry if anyone was offended by my remark, but I haven't seen any substantial objections beyond "I don't like this" (you will doubtless point me to some if there are).--Kotniski (talk) 09:41, 24 September 2008 (UTC)[reply]
  • Oppose as BJweeks and my comments above. I generally disapprove of bots being used to win edit wars. AKAF (talk) 06:48, 24 September 2008 (UTC)[reply]
    • What has any of this got to do with edit wars? If bots bring articles into line with the guidelines, we're all winners.--Kotniski (talk) 09:41, 24 September 2008 (UTC)[reply]
      • By its nature, a guideline such as the MOS tends only to have consensus within the small group which actually edits the manual of style. Normally the wiki-wide consensus of the changes suggested by the MOS is gauged by whether those edits are reverted by other editors. Lightbot attempts to turn particular sections of the MOS into a policy by winning all of the following edit wars. Note that I say suggested changes, not mandated changes, since MOS is a guideline rather than a policy. Even if the MOS were a policy, there is still considerable doubt about Lightmouse's interpretation. In many of these cases, Lightmouse is using Lightbot to win the edit war about whether his interpretation is correct. Additionally, this approval is so broad that it essentially allows Lightmouse to write the MOS as he pleases, and enforce his own interpretation on wikipedia. As Lightbot is currently implemented, I think probably at least 60% of its edits are not controversial, but this approval is far broader than the current implementation. AKAF (talk) 10:44, 24 September 2008 (UTC)[reply]
        • We have to consider benefits against costs. I would say about 99% rather than 60% (I've never noticed it get anything wrong on my watchlist, though I know people have reported isolated cases where it's made mistakes - bots, like humans, are not infallible). Even if it makes the occasional mistake (like we humans do), they can be put right, and that should not be used as a reason to stop the bot making the huge number of improvements it's making to WP articles all the time. As far as I am aware Lightmouse is careful to follow the MOS, which isn't really open to much misinterpretation, and has been pretty stable in this matter for a long time now (apart from the debunking of the autoformatting nonsense, which was achieved after a very long and thorough discussion).--Kotniski (talk) 12:15, 24 September 2008 (UTC)[reply]
          • When I read Wikipedia talk:Manual of Style (dates and numbers) I don't see the kind of consensus for Lightbot's currently programmed tasks which you apparently believe is there. Further, the scope of this approval is such that Lightmouse can edit units and dates as the whim takes him. If there is less than full consensus for his current tasks, I simply don't see how there can be a consensus for the entire scope of this approval.AKAF (talk) 12:38, 24 September 2008 (UTC)[reply]

Kotinski's post assumes that every article is on someone's watchlist, and that those watchers will notice and fix any errors that Lightbot makes. I have roughly 150 articles on my watchlist, and I'd say Lightbot has made errors on around 5 of them. Articles about topics that happened before the year 1000 seem to be more prone to trouble. So I think the error rate is too high. Also, there is no procedure to avoid making the same mistake even after it has been reported. There is no procedure to undo edits that were made during a period when the bot is known to have been broken. I don't think Lightmouse has the correct attitude towards quality control to be running a bot. --Gerry Ashton (talk) 17:31, 24 September 2008 (UTC)[reply]

The scope of this approval seems insanely broad. I see that this very issue was raised by several of the commenters in the original discussion, but never addressed. There are many MOS related edits that require human judgment and should not be made by bots. -Chunky Rice (talk) 23:33, 24 September 2008 (UTC)[reply]

  • I can see very little benefit wikilinking every date in every article. It obscures the focus. As a human editor I prefer to stay in charge of what is linked and what not. Inwind (talk) 07:04, 26 September 2008 (UTC)[reply]
    • But that's not what's at issue here. A bot which did that would be as mischeivous as Lightbot is, but we don't have one.

I think bot has done 340,000 edits over a period of 5 months. It is inevitable that many people have seen its edits. This debate looks to me like a debate about the Manual of Style. I am not sure if this is the place where the MoS can be redebated but this is what it says:

  • Dates (years, months, day and month, full dates) should not be linked, unless there is a particular reason to do so
  • The linking of dates purely for the purpose of autoformatting is now deprecated.

Discussion of the meaning of those words is best undertaken at wt:mosnum. Lightbot (talk) 11:00, 26 September 2008 (UTC)[reply]

  • No, it is principally a debate over when the MoS should be enforced by bot; which is only when there is unconditional acceptance that the rule should be enforced universally and without exception. There is none such here. If Lightbot is restarted, I will request that it be blocked. Septentrionalis PMAnderson 18:22, 26 September 2008 (UTC)[reply]
The above discussion is preserved as an archive of the debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA.
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia:Bots/Requests_for_approval/Lightbot_3&oldid=1143528065"