Jump to content

Wikipedia talk:Bot Approvals Group/Archive 7

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1Archive 5Archive 6Archive 7Archive 8Archive 9

BAG bot down?

Seems like BAGBot bot not updating Wikipedia:BAG/Status since 3-4 days. -- Tinu Cherian - 05:10, 4 September 2008 (UTC)

a lack of BAG participation

I think I counted 19 BAG members listed as active. And yet only one responded here. I suggest one of three things: 1)Ask BAG members to comment on BRFAs more often, 2)Move members that don't participate to the inactive section, or 3) get more BAG members. One response on a BRFA (that lasted 10 days) from a BAG member seems kinda of "lacking".--Rockfang (talk) 09:18, 7 September 2008 (UTC)

It is de facto policy to not approve bots you have also approved the trial for, which is why I leave those for other members. In the case of your request I saw no problems but wanted more input from other members before acting. BJTalk 09:49, 7 September 2008 (UTC)
I understand and I thank you for commenting both here, and on the request. FYI, your name didn't show up on your reply above.--Rockfang (talk) 09:41, 7 September 2008 (UTC)

RussBot

I've blocked RussBot because it was being used to do newpage patrol. Although normally I'd be all for this, the bot doesn't specify what criteria it uses to mark pages as "approved" (compare with JVbot, which uses the contents of a whitelist). Please leave it blocked until such time as its criteria are made explicit. DS (talk) 18:53, 9 September 2008 (UTC)

As I have explained on the above user's talk page, this operation is stricly script-assisted, not automated. I manually review each page. --Russ (talk) 19:04, 9 September 2008 (UTC)
Then they should be done using your account, bot accounts are strictly for approved tasks. BJTalk 21:17, 9 September 2008 (UTC)

AnomieBOT

I suggest this bot be taken out of operation until the problem noted at User_talk:AnomieBOT#Need_to_be_kept_under_control is fixed. User talk:AnomieBOT describes 2 cases in successive days where AnomieBOT has caused edit conflicts in articles I've been working on. Anomie has repeatedly reverted a civil and factual reply to one of his responses, and had the gall to post a 3RR warning on my Talk page. His responses fall well short of the "prompt and civil help from the bot's operator if queries arise" required by Wikipedia:Bot#Good_communication. His suggestion that editors should add {{nobots}} at the top of articles is not helpful, as the only template recommended by the "Edit conflict warning" and Help:Edit conflict is {{inuse}}. He could have solved the problem in the time he's spent reverting my comments. -- Philcha (talk) 10:46, 13 September 2008 (UTC)

Unless I've missed something in the last two years no bots are required to avoid edit conflicting users. The code Anomie is talking about is to prevent the bot from fixing refs that were broken by vandalism. BJTalk 11:05, 13 September 2008 (UTC)
I agree that WP:BOT does not specifically require bots to avoid edit conflicting users. However it says:
Because bots are potentially capable of editing far faster than humans can, have a lower level of scrutiny on each edit than a human editor, may cause severe disruption if they malfunction or are misused, and are held to a high standard by the community, high standards are expected before a bot is approved for use on designated tasks.
The speed of bots creates a risk of edit conflicts, as shown by my recent experiences while doing 2 re-structures on successive days. I can't suggest a way of counting edit conflicts caused by bots, but for one user to have 2 edit conflicts with the same bot in 2 days seems rather high to me.
I can't comment on what the code says, as I'm not familiar with Perl - I suspect this applies to the great majority of editors. -- Philcha (talk) 11:54, 13 September 2008 (UTC)

Please do not cross-post complaints. bots cannot detect if it will cause an edit conflict. as it makes the first edit. there is no way that it can detect if someone else is attempting to edit a page. nor do I care if a bot creates one edit conflict per day. how often are users edit conflicted with vandalism or other users? probably the ratio of edit conflicts are 1 bot caused to about 5000 human caused. 0.02% of edit conflicts is not a big deal get over it and stop being disrespectful and harassing. βcommand 14:03, 13 September 2008 (UTC)

Betacommand, you are an experienced editor and I am disappointed by your incivil manner. Please assume good faith: the only comment in this discussion thatI would consider disrespectful or harassing is your own.
Anyway, to address your point, where did 1 in 5000 come from? The last 50 edits the bot made span 8 hours; that's a rate of 300 edits per day. Assuming that only 10% of people affected by an issue complain (which is the standard metric I've heard quoted), that means that 20 of 600 edits caused edit conflicts - over 3%, and enough to be an inconvenience. Further, it is possible that these conflicts are a result of a systematic issue which will affect certain editing patterns more than others - meaning that some editors will be "victimised".
Possible way to avoid edit conflicts are: (1) to avoid editing articles that have been very recently edited; (2) to avoid editing articles with an {{inuse}} tag. (1) may have its problems, but (2) is easily put into use; indeed, it was very promptly effected by User:Smackbot at my request, which avoided a great deal of irritation at my end.
The bot operator's allegedly uncivil activity is a further suggestion that he may not be taking the needs of editors into account. I would suggest that he should be given the chance to either modifies the bot, or justify its current operation; it may be appropriate to disable the bot if he refuses to do either. Martin (Smith609 – Talk) 22:05, 13 September 2008 (UTC)
Martin please dont throw AGF around. I know exactly what Im doing. those numbers 1 in 5000 come from a bot operator who has made over 900,000 edits. take a look at stats, AnomieBOT averages 162.88 edits per day over its last 2696 edits. that means by your own math only 4.8864 edits are affected per day. (I think your numbers are high but I'll let that slide). also bots have zero control over edit conflicts. My comments are not assuming anything. I am taking the disrespectful and attacking manor of yet another person who thinks bots should be perfect. Martin please check your facts before putting your foot in your mouth. βcommand 03:28, 14 September 2008 (UTC)
How is this bot any different than other bots that edit the mainspace? BJTalk 22:20, 13 September 2008 (UTC)
It would probably have been a good idea for you to have actually looked into the issue before some of those comments, Smith609, instead of relying on biased hearsay. To clear up one point, my very first reply to Philcha agreed with the suggestion to watch for {{inuse}}, and it was implemented 9 hours and 20 minutes later (20 hours and 41 minutes before your comment). Also, FWIW, I find made-up numbers unconvincing. Anomie 23:31, 13 September 2008 (UTC)
If you want real numbers, ImageRemovalBot is on the receiving end of an edit conflict once every 3749 edits. --Carnildo (talk) 23:57, 13 September 2008 (UTC)

Offline question for a skilled bot operator

Would someone mind dropping me an email to discuss an idea I had for a bot? I'd pinged a couple of recommended folks, but hadn't heard anything back in some time. The short version is a bot that will trawl through RC, to collect and tabulate pages that are edited by frequency, and then compile the list on a schedule (30 minute increments, perhaps) and post them to a page here, sorted by "most edits". Essentially, it would be a fire alarm, to allow people to see what articles are getting hit "most frequently" in the preceding 15, 30, or perhaps 60 minutes, and to easily find possible edit wars or fast-paced edit vandalism (inspired by the Wheel War RFAR and incredibly overlooked editing volume over Sarah Palin. Might take just a few emails to see if my idea and it's specifics have legs. Thanks! :) rootology (C)(T) 16:13, 21 September 2008 (UTC)

http://www.wikirage.com/ — Carl (CBM · talk) 19:21, 21 September 2008 (UTC)
Well, that would be about it, yeah. Beaten to the punch! rootology (C)(T) 04:59, 22 September 2008 (UTC)

Chris G's BAG membership

I've closed Chris G's BAG membership as successful. I've never closed a BAG membership request, so let me know what needs to be done next. =Nichalp «Talk»= 16:44, 22 September 2008 (UTC)

Lightbot

Lightbot has been diligently removing all links to years. There is no prohibition in the Manual of Style to year-links, and there are many editors (myself included) who feel that a link that enables a user to gain a broader historical perspective is valuable. The precise wording of the relevant provision of MOSNUM is:

  • Dates are not linked unless there is a particular reason to do so.

A bot simply cannot determine whether there is a particular reason to do so; only editors can do that. I believe that having good-faith links deleted by a bot is disruptive to Wikipedia, so I have three questions:

  1. Has Lightbot been specifically approved to remove all date links, including links to years (thereby rendering our year articles orphans)?
  2. If it has not been so approved, what is the procedure for getting it to stop?
  3. If it has been so approved, what is the procedure for getting that portion of the approval reversed?

Thanks. Robert A.West (Talk) 19:59, 22 September 2008 (UTC)

  1. Wikipedia:Bots/Requests for approval/Lightbot 3 is the main approval, although it builds on two earlier BRFAs (1 and 2).
  2. N/A
  3. That's never been very clearly defined. A post to the Bot owners' noticeboard outlining your concerns would be a good place to start.
(also)Happymelon 04:47, 23 September 2008 (UTC)

Clarification for User:The Anomebot2 activities

I've been using this bot for geotagging articles for some time. After discussion at Wikipedia:Wikiproject Geographical coordinates, the bot has, as part of its general article-tagging activities, been adding a new {{coord missing}} geotag to articles which it cannot automatically match. This displays a "coordinate missing" message in the same location and style as normal geotags, in a way modelled on the behavior of empty {{coordinate}} tags on the German-language Wikipedia. All of this was discussed in detail at Wikipedia talk:WikiProject Geographical coordinates prior to the mass-tagging of articles.

Although there have been a few complaints, mostly about a tiny number of mistagged articles, the overwhelming majority of the feedback I have got has been positive. The tagging process seems to have been very successful at its intended purpose, and has resulted in over 2500 articles being hand-tagged by editors in two weeks as a direct result of adding these tags, an activity rate which is likely to rise when the planned outreach activities start.

The bot has now been blocked by User:Davidgothberg, who doesn't like the tags, and believes that this activity was outside the bot's charter. Clearly, I disagree with this, as the tags added are (a) geotags, (b) machine-matched in the same way as before, and (c) added by the bot in the same way as before, all activities which seem to me to be within the range of activities authorized by the original approval. I've devoted a very large amount of effort to this initiative, so I'm quite unhappy about this, as, it seems, are a number of other editors who do not share David's views.

This resulted in a centralized discussion on Wikipedia talk:WikiProject Geographical coordinates, where it seems to me that the overwhelming consensus view is that the activity was within the original intended remit, and that the activity has resulted in significant improvements to articles. A couple of editors, notably David, disagree, but as the discussion progressed they appear to me to have faded away from the discussion, having ceased to make any reasoned arguments to support their case, after their existing arguments were adequately addressed by other editors. However, the bot remains blocked, and they thus appear to have won the argument by default.

Although I could easily unblock it myself, I have no desire to do so. I'd like to get some clarity in this matter, if necessary by submitting a second approval request for {{coord missing}} additions as a separate task, and certainly to get going again on the regular business of auto-matching and reformatting geotags.

I'd greatly appreciate a review of this case, and advice on how to proceed. -- The Anome (talk) 22:52, 14 October 2008 (UTC)

Was there consensus formed for humans adding the template to articles? This doesn't appear to be a bot issue to me. BJTalk 23:02, 14 October 2008 (UTC)
As far I can tell, judging by the discussion there, yes. -- The Anome (talk) 23:04, 14 October 2008 (UTC)
If there is no objection to the template existing and being used I don't see why there would be objections to bots adding it. BJTalk 23:06, 14 October 2008 (UTC)
There are objections, as I described earlier (see the talk page in question for the full discussion), but the overwhelming consensus, after the centralized discussion, seems to be that the activity is OK. My question is: am I correct in thinking that this suffices to allow me to lift the block in good faith (I'd like to think so, and that BJ's comment above confirms it), or is David correct that the bot is not authorized to make these edits, based on his (I think erroneously legalistic) construction of the original bot charter? -- The Anome (talk) 23:10, 14 October 2008 (UTC)
This is bad form, Anome. You bring up an action that I have taken for discussion without informing me? I only found this discussion by accident.
Anyway: The edits of The Anomebot2 started turning up in my watchlist some days ago. It was adding {{coord missing}} to pages such as Igloo who are not located at one place. So I got curios and looked around and the only bot approval I could find was the Wikipedia:Bots/Requests for approval/The Anomebot2, and that only includes adding known coordinates to pages. And I doubted the bot will get bot approval for adding {{coord missing}}. So I blocked the bot, since it was running a questionable task and does not have bot approval for that task. Note that Anome himself has claimed the bot had tagged 102,000+ pages with the {{coord missing}} before I blocked it. Letting his bot tag 100,000+ pages without bot approval for the task is a blatant breach of policy.
Secondly, the only "consensus" that Anome has for this action is from the geodata enthusiasts over at Wikipedia talk:WikiProject Geographical coordinates. I think that before adding a "data wanted tag" to literally hundreds of thousands of pages a broader consensus needs to be achieved. Things like this should be discussed at the Village pumps or similar for some time. Not locally within one WikiProject.
And such "data wanted" tags have been denied before. For instance, some people used to put requests on biography articles for uploading an image of that person. After extensive discussion consensus became that such requests is a bad thing and those requests have been removed from the articles.
If every WikiProject should be allowed to put their specific data wanted tag on hundreds of thousands of articles, then pretty soon the articles will be swamped with such tags. I hope this is not the latest land rush for WikiProject space in the articles? Or should I inform the WikiProjects that I am a member of to better hurry up and reserve a spot in the article pages for their messages?
And even if the {{coord missing}} would be the only data wanted tag allowed in the end, then what right does the WikiProject Geographical coordinates have to waste some reading time for each of the many millions of readers of Wikipedia? That accumulates to a lot of wasted human time. The readers come here since they want information, not since they want to read about what is not in the articles.
--David Göthberg (talk) 17:57, 15 October 2008 (UTC)
Echoing what I said above, I don't think this is a bot matter but rather a debate on {{coord missing}}. If consensus can be established for using it I don't see the purpose of another BRFA. BJTalk 18:25, 15 October 2008 (UTC)

AntiPageBlankBot

I believe this is the correct place to propose the creation of a bot. I was thinking of a bot that would automatically patrol the recent changes list, and the millisecond that it sees "Blanked the page" (or just a blanked page with a different edit summary) within the article namespace, the edit would be reverted. It would also revert a similar thing, an edit that removes more than 700 characters from of an article, with no edit summary (excluding an automatic edit summary), and the IP or username is not found on the talk page (meaning that the content removal was unexplained). This would mean that edits such as "(Replaced content with 'dhlvkjbxcgl.kbjnflkxh bjmlckvnblcnghkldfjlkgjxfkl;gjx this article sucks')", would be reverted. Also, the user would not need to be warned for edits that are reverted in nearly an instant because such edits would not cause any disruption. -- IRP 20:34, 19 October 2008 (UTC)

No thanks. There are a number of anti-vandal bots that already patrol RecentChanges and have far, far less false positives than this bot would. --MZMcBride (talk) 21:26, 19 October 2008 (UTC)
So any page more then 700 characters turned in to a redirect would get slapped here? — xaosflux Talk 01:44, 20 October 2008 (UTC)

Anyway, who here has heard of the abuse filter? — Werdna • talk 01:50, 20 October 2008 (UTC)

Thanks for letting me know about that. It seems like a better idea than mine. -- IRP 23:54, 20 October 2008 (UTC)
It was a good idea, IRP; there are simply systems in place that seem to work better, is all! AGK 18:41, 2 November 2008 (UTC)

Non-BAG members

Just wondering whether Non-BAG members , but technically capable , ask for BRFA trials ? -- Tinu Cherian - 09:43, 23 October 2008 (UTC)

I'm not sure I understand, could you clarify what you're asking, please? All bot operations must be approved through BRFA, so all bot operators "ask for" BRFA approval. BAG members are responsible for overseeing the BRFA process, approving trials, green-lighting bot operations, and recommending bots for flags from the 'crats. Happymelon 10:18, 23 October 2008 (UTC)
Ok, let me try to explain : The first part of BRFA discussion is the Bot trials. My question was whether Non-BAG members CAN ask for bot trials ? -- Tinu Cherian - 11:00, 23 October 2008 (UTC)
Are you asking whether people not in the BAG are allowed to approve a bot for a trial run? I've always assumed no, but I suppose it might be reasonably done (at least) in cases where the trial can be carried out in a manner which does not count under WP:BOT as requiring approval (such as with manual preview of each edit, or "very low in number and frequency, and […] restricted to test pages such as the sandbox"). Anyway, at least per the original concept of the BAG, pretty much anyone with the good judgment and technical skills required to evaluate the need for a trial run ought to be eligible for BAG membership anyway. Certainly it would seem quite unnecessary to create an extra class of people authorized to approve bots for trial but not to issue final approval. —Ilmari Karonen (talk) 11:46, 23 October 2008 (UTC)

CommonsDelinker

On October 2, I noticed some bugs left by CommonsDelinker (talk · contribs · logs · block log).[1] Templates interpret certain whitespace characters as a non-empty field. When the image is removed without removing these other characters, the templates produce garbled infoboxes. I suggested a simple fix - add these characters to the regex that removes the image. The bug was still not fixed by October 29, when I left another note for the user and blocked the bot.[2] Again checked November 3, another note left for user.[3] [4] [5] Plus there was another case where the bot messed up, possibly due to the abundance of pipes and brackets in this case.[6]

Now it's November 16, and the infobox bug is still not fixed.[7] [8] It seems that on any day I can go through about 15-20 CommonsDelinker edits and find one of these cases.

Is there a way to fix all these templates to handle this situation correctly? If not, I intend to indef block CommonsDelinker unless someone can convince me otherwise. I do not believe we would allow an automatic bot on en-wiki to continue operating with these problems. Gimmetrow 01:24, 16 November 2008 (UTC)

Since mw:Extension:StringFunctions is not enabled here, AFAIK you could have the templates ignore certain strings, but not any arbitrary string consisting of the whitespace characters (e.g. you could ignore the strings "1 repetition of U_200e", "2 repetitions of U+200e", and so on, but you can't specify "any number of repetitions of U+200e"). Even that would, of course, require editing every one of those templates. Anomie 04:23, 16 November 2008 (UTC)

Any objections?[9] Gimmetrow 20:24, 18 November 2008 (UTC)

If the bot is breaking articles and its owner will not fix the function that is causing the break, it is permissible to block it under my interpretation of policy. MBisanz talk 20:49, 18 November 2008 (UTC)
The owner's view seems to be that the bot is correctly removing the image, and is not responsible for anything else on the line. Gimmetrow 20:54, 18 November 2008 (UTC)

Is having this error in the infobox worse than having a redlinked image in the infobox? WJBscribe (talk) 20:55, 18 November 2008 (UTC)

I believe Carnildo or ST47 has a bot that would remove the image without breaking the template. MBisanz talk 20:58, 18 November 2008 (UTC)
Those bots watch the local deletion log, not the one at Commons. BJTalk 21:01, 18 November 2008 (UTC)
Not knowing the code, how hard would it be to watch both? MBisanz talk 21:02, 18 November 2008 (UTC)
Fairly easy. --Carnildo (talk) 06:12, 19 November 2008 (UTC)
Ideally the bot owner would fix this and the pages would have neither infobox error nor redlink. This doesn't seem particularly complicated to me. If you all agree this should be done, and the admin community at commons values this bot, commons admins would probably urge the bot op to "add this feature". Gimmetrow 21:04, 18 November 2008 (UTC)

commons:COM:AN#CommonsDelinker. Gimmetrow 07:36, 20 November 2008 (UTC)

Fix claimed.[10] Still doing it though. [11] Gimmetrow 14:39, 26 November 2008 (UTC)

[12] Still doing it. What do we do when a fairly useful bot continues to leave the same problem, and the bot coder doesn't fix it? Gimmetrow 05:25, 5 December 2008 (UTC)

Replace it, then remove the old bot from service. BJTalk 05:37, 5 December 2008 (UTC)
Is Carnildo willing to take on this job and watch the commons deletion logs? Gimmetrow 05:42, 5 December 2008 (UTC)
I can handle the image-removal part of what CommonsDelinker does, but it also replaces renamed images, and I don't have any bot code for dealing with that. --Carnildo (talk) 02:20, 6 December 2008 (UTC)

[13] Still not fixed almost three months after operator was informed of the problem. If CommonsDelinker is indef blocked perhaps it will encourage a more responsible bot writer/operator. Gimmetrow 05:44, 30 December 2008 (UTC)

Give me a couple days to test it, and I'll have a clone of CommonsDelinker ready. --Carnildo (talk) 10:12, 30 December 2008 (UTC)

Re-affirm/clarify

This is perhaps already the case, but clarification would be welcome.

Are bot owners who are approved to depopulate categories due to CfD also considered to be approved to use their bot repopulate categories upon request? (Such as if a category was depopulated just prior to nomination, which is contrary to CfD guidelines.)

Also, are bot owners allowed to help with large WP:BOLD tasks? For example, if a category creator realised that there is a better name for a category they created (due to existing systems), can they request an owner of a bot (which has already been approved for usage in depopulating categories) to help them with the depopulation/repopulation, or does this have to go through BAG?

There are obviously quite a few other applications/examples. (Such as helping with category diffusion.)

So essentially, if a bot is approved to depopulate categories, does this mean that they can also use their bot to populate/repopulate categories, and can they do so beyond just the CFD/working page? - jc37 18:15, 24 November 2008 (UTC)

Normally these bots are approved to "process CFD's". As CFD's can result in rename and merge these bots clearly can add categories as well. If someone wants to rename a category they should put it on CFD, if it is new and uncontroversial it can go on WP:CFDS. It is out of scope for a bot approved to "process CFD" to become a new categorizations bot. The goal of that scope is to ensure that the bot's operations are executing edits that have consensus, one editor who asked an operator to add Category:Pages that start with A to every Mainspace article that begins with the letter "A" would likely not have consensus, and while the bot could easily start working on the task, it failed to show that the bot is using a vetted process to seed it. It should be mainly about consensus, so if a category talk conversation happened that agreeded a category should be subcategorized, there's no good reason a CFD processing bot couldn't do it. — xaosflux Talk 13:58, 6 December 2008 (UTC)
So any consensual discussion, not just one listed at CfD?
And further, are you also saying that such bots cannot help with category diffusion? - jc37 21:55, 10 December 2008 (UTC)

Resignation from BAG

I am no longer active in BAG activities and would thus like to withdraw my name from the group. Thanks. Giggy (talk) 04:28, 9 January 2009 (UTC)

 Done Thank you for your service. MBisanz talk 04:38, 9 January 2009 (UTC)

B'crats in BAG

While B'crats are only technically capable of flagging the bots, should some bcats interested in bots "allowed" into BAG ? or any BAG admin members interested in RFB soon  ? :) -- Tinu Cherian - 05:34, 9 January 2009 (UTC)

User:Andrevan was a crat in the trial period. bibliomaniac15 05:42, 9 January 2009 (UTC)
Using the oft repeated political phrase, "BAG serves at the pleasure of the 'crats". They make the final call on all bot flaggings and could remove the power of BAG on a whim. BJTalk 05:44, 9 January 2009 (UTC)

Idea about categorisation of closed BRFAs (copied from WT:BRFA)

As very few people read WT:BRFA, I thought I'd copy my proposal here to get some eyes on it (after all, the BAG has to do the janitor work at BRFA, and this might help).

The various categories for completed BRFAs have many requests not in the right place due to the final parameter of {{BT}} being left out (and therefore being sorted under 'Wikipedia'). Do you think it would be a good idea to have a bot add the appropriate sort key (taken from the page's title) to [[Category:Wikipedia foo bot requests]]? This will allow requests to be found much more easily, as they will all be under the correct letter. I realise that we have the archives, but they are chronological: using the categories is far easier if you know the name of a bot, but not the date it was approved/denied. Once the backlog has been reduced, this task could be run twice a month (?) or so to keep the categories in order. What do you think? I am willing to do the coding should such a task be deemed necessary. (Directly copied from my post to WT:BRFA, see here for the original) Richard0612 14:15, 17 January 2009 (UTC)

Correct me if my understanding is wrong? You want subcategorize Category:Wikipedia_approved_bot_requests for the approved BRFAs like Category:WikiProject Tagging bot requests, Category:Newsletter bot requests etc ? -- Tinu Cherian - 11:49, 17 February 2009 (UTC)
No, the proposal was about keeping the BRFA-specific categories organised. I filed a BRFA in relation to this task (which was approved), more details can be found there. Richard0612 12:15, 17 February 2009 (UTC)
Ok.-- Tinu Cherian - 13:19, 17 February 2009 (UTC)

Coordinators' working group

Hi! I'd like to draw your attention to a new working group, an effort to bring both official and unofficial coordinators together so that projects can more easily develop consensus and collaborate. This group has been created after discussion regarding possible changes to the A-Class review system, and that may be one of the first things discussed by interested coordinators. We are also planning a better coordination among all projects and centralizing.

All designated project coordinators are invited to join this working group. If your project hasn't formally designated any editors as coordinators, but you are someone who regularly deals with coordination tasks in the project, please feel free to join as well. — Delievered by §hepTalk on behalf of the WikiProject coordinators' working group at 20:42, 28 February 2009 (UTC)

I'd like to place a request that at least a few members act as a link between the BAG and bot owners and that page. The present discussion could generate a lot of botwork.Headbomb {ταλκκοντριβς – WP Physics} 02:38, 2 March 2009 (UTC)

New Page?

Just mainly because I'm pedantic when I want to be I was wondering if there should be a place for possible controversial bots/bot ideas to be vetted before they're taken to BRFA, mainly to act as a central discussion point away from the general bot approval work and to keep it less cluttered. Main case in point is/was WebCiteBot. Q T C 01:08, 13 March 2009 (UTC)

I'd say WP:VPR should be used for proposals. Anomie 01:14, 13 March 2009 (UTC)
Same as above. VPR, with a link on {{cent}} should do it. §hepTalk 04:22, 13 March 2009 (UTC)

(Moved to WP:BOTREQ.) - Jarry1250 (t, c) 16:42, 28 March 2009 (UTC)

Un-retired BAGger

Greetings. I was an active member of BAG from August 6 to October 29, 2008, when I retired from Wikipedia. (Real life intervened, and I needed to get myself to focus on non-Internet-related concerns.) My life has happily calmed down quite a bit since then, and I un-retired on February 26 of this year. Now I'd like to go back to helping again with BAG. So my question is: do I need to reapply? Am I a member of BAG who was previously on hiatus, or a former member who should reapply before closing or approving requests? All the best, – Quadell (talk) 19:45, 7 April 2009 (UTC)

Nope, just move your name up from the Former/Retired section, you didn't do anything heinous since your retirement, so I see no reason to make you re-apply. MBisanz talk 20:02, 7 April 2009 (UTC)
Concur, your confirmation was in the 'new style', complete with 'crat close, and was less than a year ago; so I see no reason to discount it. Welcome back! Happymelon 09:29, 8 April 2009 (UTC)

Oh good. Glad to be back! – Quadell (talk) 12:19, 8 April 2009 (UTC)

Welcome back , Quadell! -- Tinu Cherian - 13:02, 27 April 2009 (UTC)

I would like to suggest one of two things. Either that more attention is paid to potential bot names by bot approvers, or the policy is changed to be looser in standards. The reason I bring this up is because of this and this.--Rockfang (talk) 22:18, 8 April 2009 (UTC)

I would say it depends on what the bot is doing. A bot editing articles or highly visible pages should clearly be marked as such, but the 2 referenced above are only going to be editing a few projectspace pages that no one else even has a reason to edit and their operator's userspace. Mr.Z-man 23:19, 8 April 2009 (UTC)
I for one would like to see the naming policy followed, but it does not seem to be my decision to make. Anomie 00:40, 9 April 2009 (UTC)
Well, you could just not approve them. On the other hand, bots that only edit their own userspace usually don't need approval. — Carl (CBM · talk) 01:48, 9 April 2009 (UTC)
I won't approve them, but others will in some cases. The "own userspace only" exemption doesn't apply in the cases Rockfang mentioned above, BTW. Anomie 01:59, 9 April 2009 (UTC)

Seems like an odd requirement to me. User:Rambotron? User:Combotime? User:Oh_Bother? Just sayin'. – Quadell (talk) 02:11, 9 April 2009 (UTC)

For what its worth, I think bots should follow the naming policy. --ThaddeusB (talk) 20:48, 22 April 2009 (UTC)

As long as the bots include an indication in their edit summary or something similar, I'd agree with Quadell. --MZMcBride (talk) 06:17, 24 April 2009 (UTC)
I think the naming policy should apply to bots that edit in the mainspace, or any talk pages (i.e. anti-vandal bots, image tagging bots, etc.) Database reporting bots aren't going to be doing anything controversial. –xeno talk 05:28, 27 April 2009 (UTC)
Eh, I still see little appreciable benefit. User pages are the primary way of providing information about an account. And it's not as though we have a ban against non-bots using "bot" in their name, so to try to rely on it as a metric is entirely useless (e.g., "KPBotany," "Botteville," "Lordkinbote," "Robotman1974," etc.). With unified login, it becomes even more difficult to regulate bot names cross-project (and cross-language!). --MZMcBride (talk) 05:43, 27 April 2009 (UTC)
The reason I say anti-vandal bots, tagging bots, and the like, should have "bot" in the name is because they leave messages on user talk pages. Newer users might not realize it's an automated process. –xeno talk 05:53, 27 April 2009 (UTC)
Eh, maybe. Though whether the name is "FooBot" or "ClownCar," they're still probably going to be confused about what an automated process is in general. And if they're like most users, they'll click the user / user talk page link to figure out who posted to their page (where they'll see that the account is a bot). So... meh. I somewhat agree with the "outward-facing" bots using "bot" in their name, to a point. Though I still think "should" is better than "must." --MZMcBride (talk) 06:06, 27 April 2009 (UTC)
Fine by me, I'm a fan of "should". –xeno talk 06:10, 27 April 2009 (UTC)
The word "incorporating" in the (former) policy was perhaps unclear, and certainly did not match common practice. Common practice is that the name should end with the word "bot" (possibly followed by some serial indicator, e.g. the roman numerals as used by User:ClueBot II-X), and WP:USERNAME does reserve this style of naming for bots. Anomie 12:06, 27 April 2009 (UTC)
MZMcBride statement about not having a ban against non-bots using "bot" in their name is incorrect (or at least incomplete). The username policy does limit the use of usernames ending in "bot" to bots. It was intended against obvious bot names. It was not intended to prohibit any occurrences of those characters. I'm not sure what the current practice is, but it was common practice to block usernames that appeared to be bots, but were not. -- JLaTondre (talk) 12:17, 27 April 2009 (UTC)
  • I like MZMcBride's policy change, but I suggest some minor corrections. The new wording "account's name should ... make the nature of the account clear" is incongruous with the next sentence, which has nothing to do with the account name. Also, I would say that the bot must identify its operator and make its function clear, through some combination of its username, edit summaries, and user page. Wronkiew (talk) 06:40, 27 April 2009 (UTC)
  • If a bot leaves a message at a user page I want to be able to see that it is a bot message from a look at the signature. With bots that make other changes (rv vandalism, assist in processes etc.) where I look at diffs anyway it might be OK to only mark them in the edit summary, but personally I'd certainly prefer them always being clearly named as a bot, i.e. ending with "bot" (plus possibly some identifier, e.g. "MiszaBot III"). I'd say that current practise is that bots are clearly named like that. From the recently approved ones and all high-profile bots, besides the two from MZMcBride I see only CataBotTsirel and Robotic Garden where I don't immediately recognize the status from the name.
    I also don't really see a reason not to stick to that, except maybe that it's boring? Amalthea 10:19, 27 April 2009 (UTC)
  • I see no good reason to change the policy in this manner. With bots having usernames ending in "bot", it is easy to see in any discussion where the name is mentioned that the account is a bot, without having to go to the bot's user page to figure that out. To borrow the example above, if "ClownCar" is making bad edits, there's a fair chance that people will uselessly ask it to stop, give it warnings, report it at WP:AIV, and so on (because who cares about a vandal's userpage, it's usually just a sentence along the lines of "Hi, I'm {{USERNAME}}" to avoid the red-linked username heuristic) instead of just having the nearest administrator block it as a malfunctioning bot.
    I don't buy the argument that "it doesn't matter because new users won't know what a bot is anyway"; new users may similarly not know what WP:N, RFA, AFD, ArbCom, and such are either, but it's easy to learn and a helpful shorthand. While an exception for "database reporting bots" wouldn't really matter, I don't see any reason for such an arbitrary exception either. "It's boring" is certainly no reason for an exception.
    I think we should not accept this change, and instead clarify the wording to match common practice and WP:USERNAME. Anomie 12:06, 27 April 2009 (UTC)

As I do not see any sort of consensus for the change, I have boldly reverted it; the proposed change may be seen in this diff. I have also added {{underdiscussion}} and advertised this at WP:VPP, WP:VPT, and WT:USERNAME. Anomie 12:21, 27 April 2009 (UTC)

I think it is important to be able to tell bots from humans by simple glancing at the username. Chillum 13:04, 27 April 2009 (UTC)
Regardless of what it says or not in the edit summary, IMO, Botnames "should" be prefix/suffixed with the keyword Bot -- Tinu Cherian - 13:06, 27 April 2009 (UTC)
Yes, I would like to echo tinu here. Chillum 13:54, 27 April 2009 (UTC)
  • As long as we disallow (or "discourage", which is practically the same around here) non-bot usernames that look like bot-usernames (and "User:Oh_Bother" doesn't look like a bot-username, last I checked), we should require bots to have bot-usernames, too. --Conti| 14:20, 27 April 2009 (UTC)
  • There are good reasons to require "bot" to be part of the username of a bot. I've seen nothing in the discussion above about why such a requirement has any negative aspects. Sure, the bot-owner won't have complete naming freedom, but really, is that important? -- John Broughton (♫♫) 01:24, 28 April 2009 (UTC)
I agree with John on this one, there are compelling reasons to keep "bot" in the username. As for must include "bot", I'm not sure, but I'm all for "strongly encouraged". Useight (talk) 02:36, 28 April 2009 (UTC)
John, the same argument could be made for requiring the word "person" be used in all non-bot account names. With the same counter-arguments. – Quadell (talk) 02:38, 28 April 2009 (UTC)
Except that "person" is the default assumption, so such a requirement would be redundant. Anomie 02:52, 28 April 2009 (UTC)
Yes, bots need to look like bots. No Skin jobs. Chillum 02:40, 28 April 2009 (UTC)
  • I agree with most others here; bot names absolutely ought to include "bot" or some similar discriminator that makes their nature clear just from the username. This makes it trivial to identify bot edits visually, which is fairly important. The two examples Rockfang links above concerned me as well - frankly, I regard the approvals as defective, though fixable with a name change. The username requirement shouldn't obviate the requirement for clear identification on the bot's userpage, but identification in userspace isn't sufficient on it's own, since that would require actively checking userpages to find out whether a user is a bot. Moreover, as John Broughton notes, there isn't any negative to the policy, unless you count the "negative" of not being able to run a bot with a confusing name. With regard to Quadell - respectfully, no, there is no useful way to make that argument. We presumptively allow human editors, and don't seriously consider bureaucratic policies that would discourage them from editing with no fault on their part. We presumptively disallow operation of autonomous bot accounts - they are a privelege extended to those who the community trusts to behave properly with them. Proper behavior includes acceding to community requirements to make bot editing clear, such as by the username, and not looking to complain or evade the requirements. Gavia immer (talk) 02:54, 28 April 2009 (UTC)
Several of the bots without "Bot" in their name predate the Bot Approvals Group, and it tends to be asserted that they aren't subject to policy, or at least not in the same way. I'm not big on that concept, but it's not a worthwhile fight to have, either. On the other hand, any bot without that claimed exemption should be identified in the name. As far as SUL accounts go - this is English Wikipedia, and while we allow human editors more freedom now, it's reasonable to require bot operators to use our standards, as already mentioned. At most, this means the bot operator uses one more account name; hardly unreasonable. Gavia immer (talk) 04:14, 28 April 2009 (UTC)
"...and while we allow human editors more freedom now, it's reasonable to require bot operators to use our standards..." We bot editors are human too. ;} – Quadell (talk) 10:30, 28 April 2009 (UTC)
Okay, so in the case of Whip Dip and Slide... this is a bot that states it is a bot in every edit summary, linking back to the bot policy and the approval for running it. But the word "Bot" isn't in the name itself. If I were a new Wikipedian, I think it would be clearer to me that WD&S is an automated task than, say, a bot called "Robotic Chicken". – Quadell (talk) 10:30, 28 April 2009 (UTC)
True, but then again, there are no bots out there that have a name like "Robotic Chicken", right? The point is not that the three letters "b", "o" and "t" (in that order) have to be in the name of the account, the point is that the name of the account should make clear that it is a bot. All these arguments about usernames like "Combotime" and "Robotic Chicken" seem rather silly to me, quite frankly. --Conti| 10:50, 29 April 2009 (UTC)
Nope. User:Robotic Garden was approved last month without incident, and is active. – Quadell (talk) 10:58, 29 April 2009 (UTC)
I stand corrected, then. :) I'm not sure if I'd like that bot editing the mainspace, tho. So what about we require bots that edit the mainspace to be clearly identifiable by their name, at least? Think of the childrennewbies! --Conti| 11:12, 29 April 2009 (UTC)
  • I have to agree with others here - bots should be instantly recognisable from their usernames. There is no compelling reason why such usernames shouldn't be prefixed/suffixed with 'Bot', other than for fun. TalkIslander 07:54, 29 April 2009 (UTC)

IMO, bots absolutely must be identifiable as such from their username. Although I recognise that most people bringing up the point are playing Devil's Advocate, a rigid and literal interpretation of a rule "must include 'bot' in the name" is not appropriate on Wikipedia. We all recognise that a User:Robotic Chicken or User:Oh Bother are not "identified" as bots, despite having the phrase in their username; equally, a User:Automated process probably is identified as such, despite not having the phrase (although this is perhaps more dubious). As with all policies, we need the application of common sense, not wikilawyering, in the application of this rule. Happymelon 09:33, 29 April 2009 (UTC)

Well, I agree with applying common sense. But how does a user come to see a bot -- how might confusion arise? A user would have to see an edit in the edit history (or recent changes), right? This is why I feel an edit summary saying the user is a bot is sufficient. That's why realistically, no one is going to be confused about whether WD&S is a bot or not. Also, no one has complained about User:Robotic Garden. Because, really, it's not a problem. – Quadell (talk) 11:06, 29 April 2009 (UTC)
A signature on a talk page? A name used in a discussion? An ANI report? The sig-on-talk-page scenario is particularly important, as users sometimes won't get the newmessages banner, and will be left wondering "where the hell did that come from?". I agree that an edit summary would be 'legalistically' "sufficient", but why is it sensible to do it that way? Is there any relevant reason why the bot 'needs' to not have "Bot" in its username? Happymelon 16:07, 29 April 2009 (UTC)

Proposal

There seems to be consensus here that the policy as written -- the letters b-o-t always need to be in bot account names -- isn't exactly right, but there also seems to be consensus that the instruction shouldn't be done away with either. I'd recommend the following change. Instead of:

The account's name should identify the operator or bot function, and must make the nature of the account clear by incorporating the word "bot", with exception for bots already active on other wikis.

...we change it too...

The account's name should identify the operator or bot function. Additionally, it should be immediately clear that its edits are made by an automated account; this is usually accomplished by including the word "Bot" at the beginning or end of the username. (Bots active on other wikis may need to use other means to indicate this.)

This keeps the guidance, but allows enough wiggle-room to apply common sense in unusual situations. Would this satisfy everybody? – Quadell (talk) 12:23, 29 April 2009 (UTC)

  • Support -- Tinu Cherian - 13:44, 29 April 2009 (UTC)
  • Looks good. --Conti| 14:44, 29 April 2009 (UTC)
  • Definitely. Happymelon 16:07, 29 April 2009 (UTC)
  • Agreed. - Jarry1250 (t, c) 16:14, 29 April 2009 (UTC)
  • We allow bots from other wikis, with operators who may not understand English. This changes our policy for these accounts, directing them to use other means to identify the bot. They may be unable to do this because of language differences. What other means did you have in mind? Perhaps the bot edit flag is sufficient in this case? Is the flag sufficient to satisfy the "immediately clear" requirement for any bot? Wronkiew (talk) 16:37, 29 April 2009 (UTC)
    No because the flag is not visible in signatures or on the bot's userpage, is not given on diffs, page histories or contributions, and is only retained for 30 days (it's only stored in the recentchanges table, not the main revision table). The bot flag is a flag (computing, not a flag.
    If a bot operated by a non-english-speaking editor causes problems, then BAG are likely to be informed sooner or later; at which point one of them can create or modify the bot's userpage to indicate its nature if that's not already apparent. Such a message should A) explain the nature of the account, B) explain that its operator does not speak english, and C) direct future viewers to BAG for problem resolution. That should be sufficient, IMO. BAG can then take the decision as to whether the bot can be fixed by discussion with its owner, or whether it needs to be blocked, which is the role they generally take anyway. Happymelon 17:00, 29 April 2009 (UTC)

I've made the change. – Quadell (talk) 17:57, 29 April 2009 (UTC)

  • The only concern I have is that "this is usually accomplished by" may give a little too much wiggle room. It looks like the consensus in the section above is "You must name your new bot this way, except in exceptional cases", not "You should name your new bot this way, unless you don't want to" as the new wording implies. Anomie 18:12, 29 April 2009 (UTC)
  • Seems entirely reasonable, per Anomie. Stifle (talk) 13:47, 30 April 2009 (UTC)

CommonsDelinker at ANI

Resolved

This ANI report may be of interest.

After seeing that complaint, I checked on a bug I reported months ago. At the time, after repeated promptings, the operator said it was fixed. It never was. The bug still exists. [14] Carnildo said a replacement could be ready soon back in December 2008. What's the status on that? Gimmetrow 01:04, 29 April 2009 (UTC)

I see no merit in that ANI report (though I may take up bd2412's suggestion, if I can think of a good way to do it). However, I am inclined to recommend a block based on your still-unfixed bug, especially if Carnildo has a replacement ready. The operator of that bot has had more than enough time to fix that bug. Anomie 01:44, 29 April 2009 (UTC)
Many at AN/I are saying it's an error with the template, not with the bot, and I'm inclined to agree. (My, that's a lively discussion there, isn't it?) – Quadell (talk) 01:48, 29 April 2009 (UTC)
The bot should be able to detect when it is leaving a spurious left-to-right mark character and correct the situation, IMO. Anomie 01:52, 29 April 2009 (UTC)
The script isn't creating the LRM. It's already there in the field. It should be easy to catch most cases of this with some modification to a regex. (I wouldn't necessarily expect a regex capable of catching every odd combination of line breaks, spaces and this unicode characters.) Gimmetrow 01:59, 29 April 2009 (UTC)
I know, but it's still leaving it. Such a regex wouldn't be particularly hard in a PCRE, just [\s\x{2002}-\x{200f}\x{202a}-\x{202f}] will match normal whitespace plus most of the interesting spaces in Space (punctuation)#Table of spaces plus the various bi-directional text marks. Adding additional characters is simply a matter of adding them to the bracket expression. Anomie 02:14, 29 April 2009 (UTC)
That's a pretty exasperating bug. All in all, I agree, it's the bot's fault -- and although the bot-op can hardly be blamed for missing it, he can be blamed for not fixing it. – Quadell (talk) 02:34, 29 April 2009 (UTC)
P.S. I have contacted the bot operator at Commons:User talk:Siebrand#CommonsDelinker problems on the English Wikipedia.Quadell (talk) 12:56, 29 April 2009 (UTC)
User:Bryan is the bot developer. Some conversations: [15] [16] [17] [18] Gimmetrow 13:10, 29 April 2009 (UTC)

Some are claiming this issue is a problem with template coding. There is a grain of truth to that, although I think the bot could and should be written to avoid the specific issue with most invisible unicode character. The grain, however, is that this bug only appears in templates, and I think it appears only when the template code adds image syntax around the text in the field (eg: | image = someimage.svg). A template which expects image fields to include all the image syntax (the double brackets: | image = [[File:someimage.svg|200px]]) may be relatively immune to this issue, as I think the usual template coding for the latter field would at worst produce a little extra space if the field contained invisible unicode. Perhaps there ought to be some sort of "standard syntax guidelines" so people can develop common expectations for all templates. This would make it somewhat easier for editors to use templates, since editors wouldn't need to guess which syntax is used in any particular template. A good set of guidelines would probably encourage the syntax forms that favor flexibility and fault-tolerance, too. (Though there are certainly situations where we don't want flexibility in template fields...) Bot policy isn't the right place to discuss this, but I'm not sure where that place would be. Thoughts? Gimmetrow 16:30, 29 April 2009 (UTC)

Some infoboxes probably use the presence of an image to determine whether to display the caption, and it's plausible that an infobox could add a placeholder image (e.g. File:Male no free image yet.png), a default image (see {{Infobox Court Case}} for example), or a "needs image" tracking category when no image is specified. Also, part of the reason many infoboxes take just the image name (with or without the File: prefix) could be so they can enforce a particular thumbnail size and such.
WP:VPR would probably be the place to start your standardization discussion. Anomie 16:52, 29 April 2009 (UTC)
Or perhaps Wikipedia:WikiProject TemplatesQuadell (talk) 17:00, 29 April 2009 (UTC)
FWIW, there is {{Image}} which accepts all kinds of input. Amalthea 17:04, 29 April 2009 (UTC)

While it would be preferable if CommonsDelinker fixed the Unicode controls directly, I don't think it matters that much as these articles now show up on WikiProject Check Wikipedia's Template with Unicode control characters report, one or two days after being edited. They usually get fixed soon afterwards. -- User:Docu

Eh, the first rule of bots is that they do no harm.... --MZMcBride (talk) 01:35, 30 April 2009 (UTC)
Hm .. the image was already deleted. If it's delinker.py he is using, the regex around line 188 should include (\u200E|\uFEFF|\u200B|\200C|) in one way or the other. -- User:Docu 01:56, 30 April 2009 (UTC)


With regards to blocking CommonsDelinker, I've got a bot ready to take over if needed. --Carnildo (talk) 01:26, 30 April 2009 (UTC)

Commonsdelinker is part of the pywikipedia framework. In the future please file bugs at our bug tracker. This way most developers notice it. I changed the regex to also remove the unicode chars directly in front and after the filename when used in a template. before and after. multichill (talk) 12:53, 30 April 2009 (UTC)

Thanks. From what I have seen at Check Wikipedia, generally it's just after the extension [19]. -- User:Docu
Hooray! It's good to see this taken care of. All the best, – Quadell (talk) 13:57, 30 April 2009 (UTC)

Bot question.

Asking for more input here, as was suggested. I previously ran a bot, SteveBot (talk · contribs · tasks · flag log · actions log · block log · other logs · count), which ran quite well, and has also ran at Simple English Wikipedia with good results. However, I was banned in August last year, and because of a loss of trust, my bot was shut down. I'm asking here if the community feels I can resume my bot tasks, as I'd rather do something with the approval of the community, than without it. Best, Steve Crossin Talk/Help us mediate! 04:36, 6 May 2009 (UTC)

It seems fine to me. The bot is approved for fairly non-controversial tasks, so there shouldn't be any issue. –xeno talk 04:52, 6 May 2009 (UTC)
Seems fine to me as well. --MZMcBride (talk) 05:01, 6 May 2009 (UTC)
I dont think it is an issue. -- Tinu Cherian - 05:44, 6 May 2009 (UTC)
I read through the gigantic AN discussion that seems to be the only on-wiki discussion of the ban, but there is a lot of heat and not much light there. However, I don't see any reason the two tasks previously approved (for the record, 1 and 2) could not be resumed. The ban apparently had to do with sharing admin accounts and possibly off-wiki chat log sharing. While there was a call at one point in that discussion for limiting Steve to only one account, I personally see nothing to indicate that clearly identified alternate accounts (i.e. a bot account) or any actual editing was at issue in the ban. I do caution Steve to err on the side of not making bot edits if consensus is not clear. Anomie 13:07, 6 May 2009 (UTC)

Consensus statement

Since the Bot Approvals Group is likely to be advised and instructed in an Arbcom decision in the near future, I've constructed a draft consensus statement to assist the ArbCom. If all BAG members would indicate whether or not they approve of the various statements, it would be very helpful. Any comments by BAGgers or others would be helpful. Wikipedia:Bot Approvals Group/Draft consensus statement on date delinkingQuadell (talk) 17:33, 14 May 2009 (UTC)

Functionaries?

Regarding this change, I don't believe we're Wikipedia:Functionaries, are we? We don't have any specialized technical access. – Quadell (talk) 14:15, 24 May 2009 (UTC)

Yeah that's what I thought, hence why I removed it   «l| Ψrometheăn ™|l»  (talk) 14:17, 24 May 2009 (UTC)
Ha! I was looking at that backwards. Well, then, I approve. :) – Quadell (talk) 14:56, 24 May 2009 (UTC)
It depends on the definition of "functionary". If you go by the definition used at the moment at WP:Functionary, we're not because we have no special userrights. But if you go by the definition at the moment on Category:Wikipedia functionaries, we are because we hold a position of trust with special responsibilities. I asked at WT:Functionary#Definition? for the people there to pick one definition and stick to it, whichever way it goes. Anomie 18:18, 24 May 2009 (UTC)
WP:Functionary doesn't mention the BAG and no BAG member its on the functionaries mailing list for being a BAG member. Hence there is no explicit or implied inclusion on the BAG as a functionary.   «l| Ψrometheăn ™|l»  (talk) 06:28, 25 May 2009 (UTC)

Are some bots approved very fast?

I think that approving DefaultsortBot after only 4 days of discussion in Wikipedia:Bots/Requests for approval/DefaultsortBot it was really hasty. Moreover, we were dealing with a Bot that would make tenths of thousands of edits and the the trial edits were... 50! I think we have to reconsider how fast can some bots approved. -- Magioladitis (talk) 08:46, 30 May 2009 (UTC)

In my experience, the amount of time it takes for bots to be approved is inversely proportional to the perceived controversy that the task might cause, not necessarily the number of pages that might be edited. In this case, the task was not generally uncontroversial; if human editors hadn't independently imposed their will on Arabic names I doubt there would have been many complaints. Four days seems kind of reasonable in that circumstance. - Jarry1250 (t, c) 08:54, 30 May 2009 (UTC)
The bot right now is adding DEFAULTSORT to unprintworthy redirects, to articles that DEFAULTSORT=articlename, to Chinese people, etc. Moreover, if we are about to affect thousands of articles we have to be sure that we don't have to do it 2 and 3 times. Sometimes combining jobs is a good idea. Maybe, another bot that is dealing with thousands of articles could do the job. I don't say that I have something in mind but maybe. -- Magioladitis (talk) 09:05, 30 May 2009 (UTC)

Questions

I've been looking through some of the BAG-related complaints on Wikipedia talk:Requests for arbitration/Date delinking/Proposed decision, and I decided to bring some of the points here for discussion. Anomie 18:19, 6 June 2009 (UTC)

Do we really pay insufficient attention to community consensus?

Some people claim that we ignore "community objections" in approving bots. Some of these claims are really "You're ignoring my objections, never mind WP:NOTUNANIMOUS!" and/or "I hate bots!", and sometimes the community objections arise only after the bot begins its activities, but some may be accurate (or may have been in the past).

For my part, I've been trying to watch for potentially-controversial BRFAs and tell the interested parties to go ask for wider community input. On the other hand, I have little patience for people who have no good reason for opposing a bot doing something that has consensus for humans to do, is easily automated, and is tedious to do manually (e.g. the recent complaint over {{R from other capitalisation}}); in those cases I tell them to go try changing the consensus for the thing to be done at all. Anomie 18:19, 6 June 2009 (UTC)

Yes. Sometimes bots are looked at from only a technical perspective and the broader "Does anybody want this?" or "How does this improve the encyclopedia?" questions are left answered or ignored. Usually someone comes along and points this out, though there have been lapses. Also, I've seen a tendency to focus strictly on the bot operator's way of doing things rather than looking for easier or less resource-intensive solutions. --MZMcBride (talk) 05:13, 7 June 2009 (UTC)
The problem is, as always, lack of community input in bot matters. People generally don't care about bots until one does something they don't like. So its up to BAG, the operators, and the few other people who comment on BRFAs to determine whether or not a task is going to be controversial. In some cases its obvious, but in others, people tend to act rather irrationally. While people active in bot matters will generally know that making ref formats in articles consistent is considered a controversial task, how many of us actually believe it should be? Mr.Z-man 17:30, 7 June 2009 (UTC)
As I've commented below, I think particular bots should be more widely advertised. Community input isn't really needed, or desirable for newsletterbots, and there are a limited number of community members who watch the BRFA closely enough to catch the rare one which is not something minor. AKAF (talk) 13:00, 8 June 2009 (UTC)
I don't see this as a general problem, but I'm not sure I can judge whether my own judgment is lacking or not. :) – Quadell (talk) 14:33, 8 June 2009 (UTC)

Do we give approvals that are too broad?

Some claim we habitually give approvals that are too broad, with Lightbot 3 often cited as an example. Personally, I think Lightbot 3 (as stated in the revised "Function details") was only over-broad in the final date bullet and the unspecified "other edits" section. But the discussion there showed sufficient lack of consensus that IMO it should probably not have been approved, and the subsequent actions (particularly in pushing the limits of the approval) are a good argument for a better review/revocation process.

As a more general matter, I don't recall seeing approvals lately that were so broad they shouldn't be given to anyone. However, I do wonder if we should be more careful about giving approval for unspecified "general fixes" or tasks requiring determination of consensus elsewhere to inexperienced bot ops. Anomie 18:19, 6 June 2009 (UTC)

Not usually. Bot scope is generally well-defined. There are some bots that have been approved (some intentionally) for "generic maintenance / tasks" or similar in the past. I wouldn't worry too much about old requests, though. It is however important to not make those same type of mistakes going forward. --MZMcBride (talk) 05:13, 7 June 2009 (UTC)
Depends. Sometimes. I think its not habitual, but more of a problem with some types of bots than others. Lightbot1's approval was also a bit broad for my taste. I don't see any way to codify this well though. In my opinion, bots which are making a lot of edits need to be much more specifically defined than bots which make only a few. AKAF (talk) 13:00, 8 June 2009 (UTC)
I haven't seen this to be much of a problem. – Quadell (talk) 14:34, 8 June 2009 (UTC)
I think we've been tending away from broad approvals lately, with the exception of proven botops, which seems to be working well. I don't know how unstable MOSNUM was at the time of the Lightbot approvals, so I don't know whether the approval was too broad then. Mr.Z-man 20:06, 8 June 2009 (UTC)

Should approvals require more than one BAG member?

Someone suggested that a bot approval should require more than one BAG member to give the go-ahead. This might not be a bad idea, IMO, but there's no need for a full-fledged vote; if anything, just two or three BAG sign-offs should be the furthest we should go if we would do that. Anomie 18:19, 6 June 2009 (UTC)

This wouldn't work, bag generally only has one or two active bag members at a time. The time it would take for bots to be approved would be just ludicrous. --Chris 01:56, 7 June 2009 (UTC)
No. One sufficiently-qualified member of the Approvals Group should be enough. (Though it may be good to explicitly state somewhere that the approving member needs to ensure they don't have a conflict of interest in the approval.) --MZMcBride (talk) 05:13, 7 June 2009 (UTC)
While it is always good to have opinions ( read comments at BRFA) from other BAG members, it is not practical to enforce approval from multiple BAG approval everytime. -- Tinu Cherian - 04:51, 8 June 2009 (UTC)
For simple statisticbots or newsletterbots, definitely not. For bots which will make a large number of edits and have some people from the community complaining, yes, absolutely. These requests often run longer, and if we can't find 3 BAG members over a month who approve, then the bot shouldn't be running. These bots are the exception rather than the rule, as far as I can see.AKAF (talk) 13:00, 8 June 2009 (UTC)


Should BAG members be somehow responsible for bots they approve?

I can't see how that would be done, besides encouraging BAG members to comment on any appropriately-advertised discussions about the bots they approved. Anomie 18:19, 6 June 2009 (UTC)

No. Bot ops are responsible for the edits and actions their bots make. BAG can only be blamed for approving bots too hastily or approving bots that a reasonable person would have known were a bad idea. (Spelling-checking bots, for example.) --MZMcBride (talk) 05:13, 7 June 2009 (UTC)
Seriously no! This is similar to Wikipedia being made responsibile for the 'factual' errors in its articles. -- Tinu Cherian - 04:48, 8 June 2009 (UTC)
I have clarified (hopefully) below. Hopefully the approver is informed and impartial about the bot. Definitely a good first port of call if the operator is not responding. I absolutely do not agree with MZMcBride's comment above. If that is the case then we should immediately dissolve the BAG, since the users bear all responsibility anyway. AKAF (talk) 13:00, 8 June 2009 (UTC)
If I nominate an editor to be an admin, I'm very careful to check his/her credentials extremely well, because if he/she turns out to be a bad admin it will reflect badly on me. But I'm not in any sense "liable" for mistakes the new admin makes. It's the same with bots. – Quadell (talk) 15:00, 8 June 2009 (UTC)
Hmmm, I never blamed or thought worse of WJB for nominating Archtransit for adminship. Users are responsible for their own actions and edits (whether through their main account, an alternate account, or a bot account). --MZMcBride (talk) 17:51, 8 June 2009 (UTC)

Should bots be more clear in their activities?

Should it be required that bots' user pages link to the BRFAs for tasks they are currently doing? Should there be more BAG oversight of things like edit summaries and use of the bot and minor edit flags? Or is all that needless bureaucracy?

I personally would like to see more linking to the BRFAs from bot user pages, but I don't think we need to worry about edit summaries and the like in most cases. We should address complaints about excessively vague or misleading summaries, though, if necessary by amending the original approval. Anomie 18:19, 6 June 2009 (UTC)

Agreed with your analysis, Anomie. On a somewhat-related note, I will say that shouting "BOT" and linking to the user page is completely unnecessary and should be stopped immediately. --MZMcBride (talk) 05:13, 7 June 2009 (UTC)
I agree that it is unnecessary and should not be enforced (bar our usual policies on human editors leaving representative edit summaries). However, I have a personal preference for leaving useful edit summaries which do link to a user page; I would not like to see these eliminated. Any bot-userpage using {{Bot}} will have a link to its original BRFA, by the way. - Jarry1250 (t, c) 10:34, 7 June 2009 (UTC)
OTOH, the first BRFA may not be particularly relevant for a bot that has moved on to seven other tasks. AnomieBOT's first BRFA is still relevant, but it also has 13 other relevant BRFAs (3 of which add additional functions to the first BRFA's task). Anomie 15:21, 7 June 2009 (UTC)
Bots not following our current naming conventions should continue to shout "BOT" in the edit summary, but the userpage link isn't really useful. But if someone wants their bot to include "BOT" with a link, I see no reason to forbid them. Anomie 15:21, 7 June 2009 (UTC)
It would only make sense to include a page to the user page if it weren't already included. But in any diff, watchlist entry, recent changes entry, history entry, etc., there's already a link to the user page and user talk page. For example:
(diff) (hist) . . mb User talk:FooBar; 11:09 . . (-3,577) . . FooBot (talk | contribs) (BOT: zomg making an automated edit)
You have the bold bot flag (b), you have the bot name (linked to the bot user page), and the talk page link. How is shouting BOT with a link to the user page helpful to anyone? You may as well include the target page and timestamp in the edit summary. It's nonsensical. If anything, it makes far more sense to link to the BRFA. --MZMcBride (talk) 15:32, 7 June 2009 (UTC)
Which is exactly why I said the userpage link isn't really useful. But some bot edits do not use the b flag, and the b flag is not saved in the actual page history entry (only in the recentchanges table) and is never displayed in the page history (only in those special pages querying the recentchanges table), so the stating of "BOT" by those bots not named clearly is still potentially useful. Anomie 17:21, 7 June 2009 (UTC)
Sure, but it isn't necessary to shout it. Unless it's a secret initialism I'm unfamiliar with. ;-) --MZMcBride (talk) 01:18, 8 June 2009 (UTC)
I think adding a simple (Task X) to the edit summary would be useful for bots with a large number of tasks. It can be quite difficult to find, otherwise. AKAF (talk) 13:00, 8 June 2009 (UTC)

Do we need a better process for reviewing/revoking approvals?

Right now, the process is just "Start a section on WT:BRFA". The problem is that this tends to result in one or two initial comments, followed by nothing happening unless the bot op does something voluntarily.

Whether we change the policy to officially endorse WP:RFC/BOT, or create WP:Bots/Requests for review, or something else, I do think we need to figure out something so people have any hope of getting real results. Nothing needs to actually change in the Bot Policy besides perhaps the venue (according to the policy, we already have all the ability we might need to modify or revoke any bot's approval), it just needs to be made clear how we intend to implement it. IMO, the following points need to be addressed:

  • Should the community discussion be structured as a discussion or a !vote?
  • Can any BAG member unilaterally tell the bot op to stop the bot while the discussion is ongoing, or should it require more discussion/sign-offs/!voting?
  • Can any BAG member unilaterally impose a modification to the approval, or should it require more discussion/sign-offs/!voting?
  • Can any BAG member unilaterally close the discussion, or should it require more discussion/sign-offs/!voting?

My opinions: Community comments should be a discussion, not a poll; "stop while discussing" should be able to be unilaterally applied by any BAG member; and any modifications, revocations, or closing of the discssion should normally be proposed by a BAG member, given some time for comment by other BAG members, and consensus among whoever cares to comment determines if it "passes". And IMO, the BAG member who originally approved the bot should feel obligated to join the discussion but it shouldn't actually be required. Anomie 18:19, 6 June 2009 (UTC)

Discussion is always good when dealing with contentious bots. There are a lot of lessons to be learned from the Betacommand incidents. --MZMcBride (talk) 05:13, 7 June 2009 (UTC)
I agree with all of the above (discussion rather than vote, for example). Personally, I have an RFC ongoing, and I think that whilst not the perfect solution, it does show how an appallingly small number of people comment on BRFAs (an issue I am seeking to help resolve by writing a few up for The Signpost) vs. complain after a bot has started its operation. I am led to believe (but please challenge me here) that during the Betacommand incidents the bot was not stopped during discussion, something which I personally feel is important. So on the four points: Discussion - Convention if not policy - Yes (within reasonable bounds) - No preference. - Jarry1250 (t, c) 10:41, 7 June 2009 (UTC)
As far as I can see (please show me otherwise), the WT:BRFA method has not resulted in any removals of bot privileges since January 2007. Is is realistic to expect users to use a forum which never removes privileges? AKAF (talk) 13:00, 8 June 2009 (UTC)
You're preaching to the choir as far as I'm concerned. Do you have any comments on my rough proposal for a replacement? Anomie 17:54, 8 June 2009 (UTC)
The hardest part is putting the replacement page in a place which the users will easily find. I'm thinking about a bot equivalent of those signs on trucks "Want to complain about my driving? Call 555-TRUCKS". A link to the "bot comment" page should be clearly displayed on the user page of every bot. Expecting the inexperienced user to navigate through WP:BOT is unrealistic, especially when this policy is in constant flux. It should, if possible, encourage both positive and negative comments. As far as removing/suspending approval is concerned: I would think that this should be as simple as giving approval. I think that the best method would probably be 1. "suspend" - any BAG member at any time 2. "re-approval" at RFBA, just like any other normal approval. The BAG members are experienced enough to see when a suspension is necessary. I would suggest that a re-approval should run for at least 7 days, to allow comments, and should be advertised (maybe at RFA/AN, like for the requests for BAG membership). AKAF (talk) 18:29, 8 June 2009 (UTC)
I think a defined process would be good, but its not something that comes up often, so we don't need to go overboard. If we did have such a process, it would need to be game-proof and drama-resistant - BAG would need to have the final say and it should be a discussion, not a poll or vote. Perhaps a mandatory cool-down period between the time of the incident and the start of the process during which the bot can't run, so that people can cool down and discuss rationally, and so that the process doesn't become a place to forum shop. Mr.Z-man 20:13, 8 June 2009 (UTC)
It may come up more once people feel like they have a chance of getting a result. Is it game-proof enough for the whole thing to boil down to "One or more BAG members looks over the complaints and decides what (if anything) needs to happen bot-wise"? I think drama is unavoidable, though... Anomie 21:23, 8 June 2009 (UTC)

General replies

I know that I have been quite vocal on these points Wikipedia_talk:Requests_for_arbitration/Date_delinking/Proposed_decision#AKAF, so I will write here, and you can all feel free to vent your frustrations on a centralised point :-)

In general I see a lot of the problems as related to poor communication. The BAG pages are, to be frank, an unnavigable mess, if you want to find something specific. Discussions seem to be randomly distributed over the talk pages of 4 or more pages. I think if you put a request to clean-up/reorganise the project pages in signpost there would be more than a few takers. I am very willing to help, but I have no wish to step on anyone's toes.

A lot of the problems I see with bots is that the process is arcane and hard to navigate for someone outside the system. This means that often the BAG members never hear about a problem before it escalates greatly. For instance, take a look at the user page for Lightbot. This was a pretty standard bot, even one of the better ones, as far as the documentation on the user page is concerned. Now think: you are a normal user, coming to ask about an edit. Ask yourself the following questions:

  1. Where can I complain about this bot? In this case the user page is clearly linked. Lightmouse just ignored the complaints.
  2. If that doesn't work, then what? (nothing)
  3. Which task was the edit to which I object related? Task numbers are almost never named in the edit summary.

Now generally if you have a problem you go looking for an administrator who is informed, but not involved. This involves WP:AN/I at the moment, which is a broad brush, since most admins there will be uninformed, and simply note that the bot is performing as approved. If you look at lightbot's user page there is no link to WT:BRFA. Actually I'm far from a complete novice, and this page is the first I have hard of this method of disapproval, so that leaves maybe less than 1% of all complainants who will be helped at that page. I have previously suggested that a BAG member be responsible for the bot which they approve, but I really think it would be sufficient to be an official point of contact, since they are informed about the bot. This would help in (say) the first 2 weeks if a problem arises which was not immediately obvious in the trial. Call it a probation period.

So I have a number of concrete suggestions:

  1. At the bot approval, categorise the bot request. Some bots can be quickly approved, others not so much. You probably neither need nor want community input on a lot of the lower level bots, so its senseless to tag all bots the same. Maybe there need to be different rules for different bots. A bot which collates user editing statistics is obviously much less likely to be a problem than a spellbot, for instance. There are far fewer of the higher level bots, so it may be practical to advertise just these to the community. I don't know how to do this best, but I would categorise them as to the difficulty of undoing them after 6 months:
    1. Server load only/statistics (no edits outside of a subpage of the bot)
    2. Newsletter (one edit max on user talk pages)
    3. Tagging (one edit max per article talk)
    4. Single edit (one edit max per article page)
    5. Multiple edits on article talk pages (eg sinebot)
    6. Multiple edits on article pages
    7. Multiple passes of multiple edits (I would think that at least this category should attract 3 BAG members)
  2. Require that the bot give enough information in the edit summary to identify the task number.
  3. Create a template for bot pages, which summarises all of the useful contact information for a newbie user. For instance:
    1. Links to the approved tasks
    2. Links to the operator
    3. If there is an urgent problem with this bot, what should you do?
    4. If this bot is operating correctly, but doing something wrong what should I do?
    5. This bot was approved by user:XXXX. If there are problems with the bot up to (Approval + 2 weeks) please contact.
    6. A general description of bots and the bot process?

AKAF (talk) 06:36, 8 June 2009 (UTC)

Many of the points above were taken from your post, actually. At the moment, the best thing you can do to help is to join in the discussion above and make (concise) points to try to convince everyone; feel free to add new sections for each proposal as I did above so each discussion doesn't get lost in the clutter of the others.
I don't see that discussion is really spread "randomly": WT:BRFA is for discussion about bot approvals, WT:BOTPOL is for discussions about the Bot policy, WT:BAG is for discussion by and about the Bot Approvals Group (this discussion is here, BTW, because I wanted to discuss the points with the other BAG members before trying to take any action elsewhere), and WP:BON is more or less the bot op equivalent of WP:AN. The current "review" process is clearly enough stated in the Bot policy, it's not like it's particularly hidden.
My opinions:
  1. This seems like it would be excessive bureaucracy, at this point there aren't so many bot requests that categorization is needed. There's also the fact that some don't particularly fall into one category, or don't come without a history. For example, my bot is currently approved for a few fairly generic tasks (e.g. fixing template transclusions once consensus is reached), but that came only after a fair number of "do it in one specific case" tasks were completed without issue.
  2. That discussion is already underway at #Should bots be more clear in their activities? above.
  3. We do already have {{bot}}, which is already widely used.
    1. AnomieBOT has 25 approved tasks at the moment (14 of which have been completed), plus two that do not need approval as they edit only pages in my or the bot's userspace. A template trying to handle all that would probably end up being excessive.
    2. Already taken care of by {{bot}}.
    3. Already taken care of by {{bot}}: Block the bot. Many bot ops also use {{Emergency-bot-shutoff}}.
    4. At the moment, there's nothing to say: "Post on the bot's/op's talk page" is obvious, and "Post on WT:BRFA" isn't terribly helpful. If we get a better review process in place a clear link to that from {{bot}} would be appropriate.
    5. The underlying idea is currently being rejected above, you should comment on that there. Although, would this be moot if we had a functional review process?
    6. You want every bot userpage to summarize WP:BOT? We already have {{bot}}, which already takes care of 3.2, 3.3, and 3.6 as far as it should go IMO. the template should be sufficient as far as templates go.
Again, please feel free to join the discussion. Anomie 11:56, 8 June 2009 (UTC)
None of the pages are hidden. But to be honest, none are particularly well advertised. Until recently people with a problem were advised to go to AN, BN or start an RFC. Can you give me a link to somewhere where a request for reconsideration of a bot was made on WT:BRFA, and the approval was withdrawn? I look over it semi-regularly, and currently there's one where the reporter's request is refused. There was Jogersbot in Jan 2007, but since then nothing, I think. I think that practice shows that the idea of posting at WT:BRFA is a red herring, since problem bots appear never to be looked at due to those postings. AKAF (talk) 12:37, 8 June 2009 (UTC)
I can't give an example, and I share your opinion that "Comment on WT:BRFA" as a review mechanism is completely useless. I've actually been considering proposing a reform (as above) for some time now. Anomie 17:52, 8 June 2009 (UTC)
I think that {{bot}} could be improved. As it is, it isn't very descriptive of the bot process for a relatively inexperienced user. I was thinking bore of just 5-6 links with an explanation as to what they are: WP:ANI, WT:BRFA, and WP:BAG might, for example, be on there with a one sentence suggestion about when each should be used. As far as point one is concerned, I think for Anomiebot you could template the link Special:PrefixIndex/Wikipedia:Bots/Requests for approval/AnomieBOT, or is there a reason why that won't work? AKAF (talk) 18:49, 8 June 2009 (UTC)

Anybot

Anybot has screwed up in the grandest possible fashion; see Wikipedia talk:WikiProject Plants#Algae articles AnyBot writing nonsense. This is not the forum to discuss that. However I am of the firm opinion that this situation must be taken into account should the BAG be asked to approve other bots operated by the bot owner. How does one inject such information into the BAG's corporate memory? Should I leave a brief note at the bottom of Wikipedia:Bots/Requests for approval/anybot? Hesperian 23:25, 16 June 2009 (UTC)

Among other things, the manner in which "this situation must be taken into account" depends on whether the bad edits are solely due to the unauthorized operation since April, how the unauthorized operation was able to occur, and whether any April-or-earlier errors are due to errors in the bot or errors in the source data (which was apparently approved by WP:PLANTS). The bot is currently blocked until that is determined and the operator is busy until October (and thus is unlikely to be requesting more bot approvals), so there seems to be no rush. Anomie 01:21, 17 June 2009 (UTC)
Indeed. I'm not telling you guys what any future decision must be. I'm telling you guys that you need to take this into account in making any future decision.
My question is where and how to add a note that you guys would find when the time comes.
I'm trying to help you guys out, unless you actually prefer to deliberate in ignorance. So how about winding back the defensiveness a wee bit?
Hesperian 01:49, 17 June 2009 (UTC)
Thanks for the bot info and feedback, Hesp, though to be honest I'm hearing more defensiveness from your end of the court at the moment. It's bizarre that this bot was running, and the bot-op did the right thing to block it. He's also recommended rolling back the bot's operations since April, though I'm not sure whether that's warranted or not. Any thoughts? – Quadell (talk) 02:53, 17 June 2009 (UTC)
You're welcome, Quadell.
Bizarre indeed. Someone compromises an account in order to keep an erroneous bot running? What is the motive?
I think discussion on how to clean up the mess should be centralised elsewhere. I came here to ask a different question. Any chance I might get an answer to it any time soon? Hesperian 03:45, 17 June 2009 (UTC)
If I can speak frankly, I think the issue here is that bots are a back-corner of the Wiki that get very little traffic outside of coders creating things and people bring tasks from Wikiprojects here for approval. In this case the bot request was open 6 months with lots of input from projects, more than most actually. Out of the 995 approved bots reqs on record, about 5-10 have been as problematic after the fact as this one. I'm really not sure how BAG can improve this batting average since it is checking for technical soundness and community consensus, not actual content soundness. MBisanz talk 04:18, 17 June 2009 (UTC)
I am not criticising the BAG for approving this bot. Hesperian 04:36, 17 June 2009 (UTC)

I don't see why we need to go out of our way to record this somewhere. We don't do that for normal editing issues. Mr.Z-man 05:47, 17 June 2009 (UTC)

I think leaving a note at Wikipedia:Bots/Requests for approval/anybot would be a good idea, Hesp. – Quadell (talk) 12:09, 17 June 2009 (UTC)
I suppose I have a question. If the bot op is a admin and a continuing bot op, is there an actual fear they will fail to reference this situation in seeking future approvals? I am sort of hoping that a user already in positions of trust would do that. MBisanz talk 13:03, 17 June 2009 (UTC)

Has this bot been approved for running again? I thought it was stopped until writers decided what they want to do, but it's editing 100s maybe thousands of articles. And it's getting them wrong, again! --69.226.103.13 (talk) 22:52, 24 June 2009 (UTC)

Yes, Martin unblocked the bot and set it running "correcting" errors. Given the context (discussion proceeding on the understanding that the bot was blocked; concerns about unauthorised access; the BAG notified that the bot was blocked, but not notified of the unblocking; recommendations that the bot be blocked and/or deapproved; expressed opposition to the bot being deployed to fix its own errors; no BAG approval for the new task; apparently no test run) I think the unilateral unblock was highly inappropriate. And then there is the problem that it is introducing errors again. I have reblocked and asked Martin to talk a BAG member into unblocking. Hesperian 23:46, 24 June 2009 (UTC)
Following on from the bot's latest run, during which it replaced hundreds of error-filled, bot generated articles with hundreds of bot-generated articles filled with different errors (see Wikipedia_talk:Articles_for_deletion/Anybot's_algae_articles#New_and_different_errors) and considering Martin's decision to run the 'fixed' version of the bot without making any test edits and apparently failing to address the majority of the concerns raised in the AfD discussion, I feel that it has now become time to seriously start considering whether it has become time to revoke Anybot's approval and bot flag. --Kurt Shaped Box (talk) 10:48, 25 June 2009 (UTC)
Indeed. Most disturbing is that this last run performed novel edits, not reverts, on hundreds of articles, in which many new errors were introduced; yet every edit was flagged as minor, and given the edit summary "Restore article to last good version." That is seriously fucked up shit. Hesperian 11:02, 25 June 2009 (UTC)
Martin also asked me to verify that it was impossible to run the bot without authorization (which is one of the factors that contributed to this fiasco in the first place), then set the bot in motion seven minutes later *without* waiting for my response to confirm that the 'hole' had been closed. To be frank, I've lost all confidence in this bot to perform its designated task as intended. As I've previously mentioned elsewhere, I'm sure that Martin is trying his best here - but this just isn't working out. --Kurt Shaped Box (talk) 11:24, 25 June 2009 (UTC)

RFC results: date unlinking bot

Dear colleagues,

The Full-date unlinking bot RFC has closed, with strong endorsement of the proposal to develop and run a narrow-task bot to remove the syntax from triple-component dates (74.3% to 23.8% to 1.9%).

Remedy 1.3 of the Date Delinking judgment by ArbCom, specifies a community-approved process for mass delinking, and Remedy 2.1 that "Date delinking bots will perform in a manner approved by the Bot Approvals Group".

The bot is currently being developed by User:harej at User:Full-date unlinking bot. Given that this point has been reached after a long debate in the community and involves a degree of technicality, I request that BAG participates in producing and running the bot, with a view to its approval. Some comments in the RFC stated that this should involve a careful testing process and, at least initially when it is up and running, regular stops to ensure the bot is running as intended.

I have notified ArbCom of the RFC results and of this request. I look forward to your response. --Apoc2400 (talk) 16:47, 9 July 2009 (UTC)

What stops malicious bots

I was referred from general help desk to here. So I will copy paste what I put there in hope I will get a better response then some generic "do not worry about it". This got me curious :)

I am interested in bots. I have read a bit on them and the ways to make them. This got me curious. What is stopping people from creating malicious bots to just randomly run and destroy parts of wikipedia? Has this ever been attempted before? I know that there is a process to get bots approved but in theory couldnt a user just create a bot for malicious activity and let it run until it gets caught then rince and repeat? Obvious bot would be obvious. But, what about a bot that would change how links directed on articles containing words x,y,z for example? Thanks Ivtv (talk) 05:25, 15 July 2009 (UTC)

Wikipedia has Captchas which stop a lot of the spam bots. A most of these bots also run via open proxies which are now blocked by ProcseeBot. Since most vandal bots follow a specific MO they can also be stopped with the Abuse filter and block bots (a recent (well in the last year or so) example I can think of was the 'Hi, Good Site' spambot and the 'anontalk' bot). meta:vandalbot has a bit of info on them as well (although it focuses on cleaning up after them) --Chris 08:46, 20 July 2009 (UTC)

BAG mentor?

Can a member of BAG volunteer to mentor me with some bot creation and approval tasks? I have two bots in development- one is very minor, the other is small in scope but a little more detailed. I just want someone who doesn't mind me bothering them, can review the technical concepts, tell me when I'm ready, point me in the right direction, etc. tedder (talk) 06:20, 15 August 2009 (UTC)

My talk page is always open. --MZMcBride (talk) 06:24, 15 August 2009 (UTC)

Wikipedia:Bots/Requests for approval

It is over two weeks, according to the stats, since a BAG member has edited at Bots request for approval. --69.225.3.119 (talk) 00:11, 24 September 2009 (UTC)

28 seconds and no reading: problem solved, everything possible approved, all lingering issues ignored. --69.225.3.119 (talk) 01:27, 24 September 2009 (UTC)
and you dont know how much time he spent reviewing those edits. please stop being so condescending. BAG is elected by the commumity just like admins, arbcom, checkusers and others. βcommand 01:30, 24 September 2009 (UTC)
I'm not being condescending, I'm being angry. --69.225.3.119 (talk) 01:42, 24 September 2009 (UTC)
why are you angry? because there is a minor backlog? if there is a specific BRFA that you want reviewed let us know and we can take a closer look. for the most part BAG attempts to let the community decide if they want the bot. βcommand 01:45, 24 September 2009 (UTC)
Because all of my comments about the content creation bot were 100% completely dismissed. There is NO community consensus for this bot. ThaddeusB and Abyssal said they didn't need any, and then offered that Abyssal was the only community of concern. The admit they don't have the expertise to monitor the bot's contributions. They ignored concerns I raised about another contribution of Abyssal's. They don't have the correct bibliographic information in the articles-an issue I already informed them of. They attacked me personally. They dismissed all of my concerns.
Then, a BAG member comes by and wants a trial run. What part of the community decided they want this bot? Its operator? If that's the approval criterion BAG is superfluous. --69.225.3.119 (talk) 01:49, 24 September 2009 (UTC)
As I said elsewhere (how many other places have you brought this up?), if I wanted a trial, I would have approved it for a trial, not asked a question. Mr.Z-man 02:01, 24 September 2009 (UTC)
And, if you wanted a summary, you could have asked for a summary, as you did for another bot. But you didn't. You asked if it was ready for a trial. --69.225.3.119 (talk) 02:15, 24 September 2009 (UTC)

Proposed Arbcom Motion re date delinking

Your attention is brought to a motion currently being considered by the Arbitration Committee:
Wikipedia:Arbitration/Requests/Motions#Motion to amend Wikipedia:Requests for arbitration/Date delinking.

At the time this notice was posted the text of the motion read:

This wording may have since changed; please see the above link for the current wording.

On behalf of the Arbitration Committee, Manning (talk) 09:45, 12 October 2009 (UTC)

The idea of removing restrictions on the use of "bot" in usernames is being discussed. Cross-posting here just in case there is some strong objection from this quarter that has not been considered. Beeblebrox (talk) 09:06, 6 December 2009 (UTC)

ArbCom election reminder: voting closes 14 December

Dear colleagues

This is a reminder that voting is open until 23:59 UTC next Monday 14 December to elect new members of the Arbitration Committee. It is an opportunity for all editors with at least 150 mainspace edits on or before 1 November 2009 to shape the composition of the peak judicial body on the English Wikipedia.

On behalf of the election coordinators. Tony (talk) 09:26, 8 December 2009 (UTC)

Locating a bot writer for an innovation

Although the repetition of a word can add to the cohesion of a text, in many cases it is poor style (I catch myself doing it, especially when tired).

For example, I've just pointed this one out at FAC:

"Activated for service in World War I, the division saw brief service in the conflict, but never fought as an entire division."

But the phenomenon is not just within a sentence; some words are prominent enough to be avoided even when paragraphs apart. I am keen to locate a Wikipedian who might be prepared to develop a bot or script as an editing/gnoming aid that would highlight repeated words in an article, thus flagging to editors where rewording may be required. It would need to avoid a white-list of a few hundred common grammatical words, of course (e.g., the, to, been, am). A more sophisticated version would do better, by matching the frequency of occurrence of each word in English (there's a publicly available list) with the distance between specific occurrences in a text; a simple algorithm would be required to balance commonness against distance, which would need to be tested). The more sophisticated version would clearly be preferable, but might be a challenge to develop and might require too much grunt.

Don't broadcast it off-wiki, because no one has ever had this idea and developed it into an automated tool; WP would be the first.

Anyone know of people who might be interested and capable? Tony (talk) 14:10, 11 December 2009 (UTC)

Probably best discussed at WP:BOTREQ. –xenotalk 14:13, 11 December 2009 (UTC)
Thanks Xeno; done. Tony (talk) 15:24, 11 December 2009 (UTC)
I admire your dedication to high-level writing. I suspect this will only be useful for FAs and other already-identified good articles. However, it could be very useful for these articles, and running the bot on a well-written FA before it is on the main page would be good timing, imo. --IP69.226.103.13 (talk) 17:37, 11 December 2009 (UTC)
I just caught myself writing "some" twice in a sentence. Occasionally this is required (lists, parallel constructions), but it's remarkable how easy it is to get into a repetition groove. I think it would be useful on any article one is editing. It would flag, not automatically edit (that would be up to the human). Wikipedia:BOTREQ#Locating_a_bot_writer_for_an_innovation Tony (talk) 12:53, 12 December 2009 (UTC)

Last chance to vote in the Arbitration Committee Elections

This is a brief reminder to all interested editors that today is the final day to vote in the December 2009 elections to elect new members to the Arbitration Committee. The voting period opened at 00:01 on UTC 1 December 2009 and will close at 23:59 UTC on 14 December 2009 as initially planned. Updated 21:08, 13 December 2009 (UTC).

The voting this year is by secret ballot using the SecurePoll extension. All unblocked editors who had at least 150 mainspace edits on or before 1 November 2009 are eligible to vote (check your account). Prospective voters are invited to review the candidate statements and the candidates' individual questions pages. Although voting is by secret ballot, and only votes submitted in this way will be counted, you are invited to leave brief comments on the candidates' comment pages and discuss candidates at length on the attached talkpages. If you have any questions or difficulties with the voting setup, please ask at the election talkpage. For live discussion, join #wikipedia-en-ace on freenode.

Follow this link to cast your vote

For the coordinators,  Skomorokh  12:57, 13 December 2009 (UTC)

Discussion regarding AWB's general fixes

There is a discussion regarding the use of AWB's general fixes by bots. See Wikipedia:Village pump (technical)#Perform AWB's general fixes from a dedicated standalone bot. All are welcome to comment. The participation of BAG members would be particularly helpful. Thank you. Equazcion (talk) 17:48, 31 Dec 2009 (UTC)

Thread that may be of interest

Please see Wikipedia:Administrators' noticeboard/Incidents#Rich Farmbrough using main account to run bot tasks and comment if appropriate (especially with respect to "SLAs" on bot approvals). Thanks, –xenotalk 20:04, 31 December 2009 (UTC)

I've had enough to do with that person for a while, so I'm going to stay out of it besides this comment: The "let him do whatever he wants until someone complains" proposal there is one of the worst ideas I've seen around here in a while, both in the general sense that no one really needs that sort of permission and in the specific sense that this particular user has recently shown neither the near-perfect level of good judgment necessary for such "approval" nor the willingness to stop and work out complaints rather than dismissing them and plowing ahead. Anomie 03:21, 1 January 2010 (UTC)

Wrong way of the close a BRFA

I belive that the bot request Wikipedia:Bots/Requests for approval/Ganeshbot 4 was approved [20] in an incompatible way with User:SQL/How to close a BRFA. Everything would be fine, but the Bot Approvals Group member approved the bot with serious limitation and closed the discussion. When approving a bot with limitation, then it is not good to immediately close the discussion, because other wikipedians have no chance how to comment this. (I think that reopening the discussion or approving without limitation would be fine.) --Snek01 (talk) 15:28, 25 March 2010 (UTC)

Looking back, I think I agree with you. Perhaps it was a bit harsh to add such a serious limitation without letting there be a comment period. I am unsure of what to do here, it is a bit of an awkward situatation, so I'll talk to some other members of the BAG and see what can be done. Sorry for the hasty actions, Tim1357 (talk) 22:53, 25 March 2010 (UTC).
I don't really see the need for a limitation. I think the bot should be able to create the articles all at once, or at least over the course of a few days; waiting half a year to create 560 pages is a bit overkill (and would probably be slower than their current rate)! If the project has trouble with the flood of new stubs, which I think is unlikely, the bot operator can have the bot create a smaller number of pages (on their own accord) and wait for those to be checked before moving on, but I don't think we as BAG members should set that number for them and prevent them from being able to run at a faster rate.
It would probably be best to either 1) unarchive and reopen the request, allowing the discussion to continue and to be closed it later, or 2) figure out something here quickly (such as my proposal above – normal approval with operator freedom to make pages at the rate they choose), and then modify the approval on the main BRFA page. — The Earwig (talk) 00:19, 26 March 2010 (UTC)
I would like to get started with the bot runs. It would be great if the limitations were removed and the bot is approved to run under the project's supervision. Thanks. Ganeshk (talk) 03:53, 27 March 2010 (UTC)
I'd be happy to let you run the task without limitations, but I'm hoping Tim1357 (or another BAG member, perhaps) would comment on my proposal above. — The Earwig (talk) 04:03, 27 March 2010 (UTC)
As long as Tim doesn't vigorously object, I think the removal of the limitation should be alright without having to go re-open and close things. If anyone later disputes it, they can look at this discussion. MBisanz talk 04:48, 27 March 2010 (UTC)
Okay. I will take that as approved without limitations. Thanks to all! Ganeshk (talk) 13:14, 27 March 2010 (UTC)

Expert technical opinions sought

Forgive me for posting off-topic: at MoS talk and the Village Pump there are discussions about setting up a shortcut for non-breaking spaces. WP desperately needs one, and previous attempts have failed for lack of technical support. The advice of BAG members with programming/html expertise would be much appreciated, since the avoidance of "literals" (I think they're called) and other technical issues will be required. Tony (talk) 07:22, 17 April 2010 (UTC)

Proposal to have BAG handle request for messaging/recruitment by researchers

See Wikipedia talk:Research#Proposal to let BAG handle individual requests. –xenotalk 18:22, 11 June 2010 (UTC)

Proposal to let BAG handle recruitment requests by researchers

It is proposed here, input from BAG members would be appreciated. Cenarium (talk) 18:24, 11 June 2010 (UTC)

Alert Wikipedia Administrators Notice Board/Incidents Bot creating 100s-1000s of factually incorrect species articles

[21] —Preceding unsigned comment added by JaRoad (talkcontribs) 00:27, 16 August 2010 (UTC)

Bots & long term inactivity

See Wikipedia talk:Bot Approvals Group/Archive 5#Various inactive Bots with inactive owners

I think it's time to have another review of bots that have been inactive for a long time. They make it hard to use special:listusers to track bot edits, they are to some extent a security risk, and where a bot hasn't been used in years I think it makes sense for BAG to double check that the task is still a desirable one before it is run again.

Assuming BAG remain of the same opinion as last time this came up (see link above), my recollection is that the approach taken was:

  1. Where both the bot and the bot owner have not edited for a long time (say 18 months) - deflag the bots with a note to the owner to contact BAG on their return
  2. Where the bot has not edited for a long time but its owner is active - drop a note on the owner's page checking whether they still plan to use the bot / need the flag

That still seems a sensible approach to me. Could someone create a list of all accounts with the bot flag that have made no edits/logged actions in the last 18 months? WJBscribe (talk) 11:14, 8 August 2010 (UTC)

-- Cobi(t|c|b) 13:35, 9 August 2010 (UTC)
This list isn't accurate. Kyle the bot last edited 16 March 2010. That's just the first one I randomly checked on so don't know how many other problems it might have. -- JLaTondre (talk) 21:18, 9 August 2010 (UTC)
My attempt:
Note that bots that make few or no non-deleted edits (e.g. User:ChrisSalij Bot or User:ProcseeBot) will be erroneously listed. HTH. Anomie 06:11, 12 August 2010 (UTC)
IMO we should be a bit more aggressive in deflagging adminbots (of both the bot and admin flags). EyeEightDestroyerBot's last log action was on March 4th, and AntiAbuseBot's last was July 25th (one month is probably too soon). Shubinator (talk) 02:54, 18 August 2010 (UTC)
I think a period of at least 3-6 months of inactivity is long enough for deflagging. No point in keeping them flagged if they aren't going to be used. ···日本穣? · 投稿 · Talk to Nihonjoe · Join WikiProject Japan! 15:40, 18 August 2010 (UTC)
I think that deflagging after a year of inactivity is fine. 3 months is too short, 6 months still seems a bit short - no need to force bot operators to jump through hoops to get reflagged if they come back after a medium-term wikibreak - inactive bots are not causing any particular problems. As for re-flagging, I think they could just visit WP:BN and bureaucrats could reflag similar to WP:RESYSOP (if bureaucrats are unsure if the task(s) is/are still desirable, they could defer to BAG). –xenotalk 15:44, 18 August 2010 (UTC)
If they have a legit reason (such as their owner being on a wikibreak), then I'm fine with leaving them flagged. However, if they just suddenly stop doing anything, I think 6 months is plenty of time to determine if they are active or not. ···日本穣? · 投稿 · Talk to Nihonjoe · Join WikiProject Japan! 15:28, 19 August 2010 (UTC)
That means we need to start getting into investigations as to why they left, etc., etc. Much easier just to use 1 year as a demarcation point and deflag any bots inactive for 365 days or more. Inactive flagged bots aren't causing any particular problem such that another 6 months of inactivity will make or break anything (as evidenced by the fact that flagged bots exists without edits since 2005!). Using 1 year cuts the number of flagged bots approximately in half. –xenotalk 15:32, 19 August 2010 (UTC)
Well, many of the bots on the list have been inactive for several years and also have inactive owners. I suggest starting with these. If someone who is active has a bot they haven't run for a while but might run again and the task looks OK, I'd be OK leaving the flag. 6 months seems a bit short to me too. WJBscribe (talk) 17:38, 21 August 2010 (UTC)

Need to read WoRMS ID

Hello BAG Team: I have a requirement to retrieve the WoRMS ID number from about 15,000 articles so that I can compare them against WoRMS database to check for validity. The Aphia ID number is a 6 character number that is entered into the {{WRMS species}} template as the first parameter (see Galeodea beui for example). I don't want to download the database dump and deal with it's size since my requirement is very specific to Gastropods. The bot policy forbids downloading substantial portions of Wikipedia's content by requesting many individual pages. Is it okay for me to read pages directly from the website for this requirement? or are there better ways to do this? Will this need a bot approval? There are no edits made, just reads. Please advise. Thanks. Ganeshk (talk) 03:11, 23 August 2010 (UTC)

Since this is a read only request, a BRFA is not required, however using Special:Export would be your best solution, and exporting multiple pages per request. ΔT The only constant 03:13, 23 August 2010 (UTC)
Thanks Δ! I will look into Special:Export. Ganeshk (talk) 03:17, 23 August 2010 (UTC)

Can I run a bot in user space and move articles manually w/o approval?

Hey everyone. I have a simple bot script put together to build "List of companies of CountryX" articles based on the relevant category ("Companies of CountryX"). As the amount of articles to create is minimal (under 150 I think) and is simply to fill the space where a category exists but no article my plan is to get the bot to create the articles under it's user space and then manually double check and move articles to main space. This is pretty much what I have already done by hand with around 5 list of articles and the bot is mostly just to automate the slog-work.

The question I have is; do I need Bot approval to do this? As the bot only touches its user space and I am moving articles manually I believe it will be fine - but if someone could actually clarify this prior to me firing it off :) that would be great. (example of its work: User:Listbot/tests/Cuba) --Errant [tmorton166] (chat!) 21:30, 26 August 2010 (UTC)

Should be ok: "any bot or automated editing process that affects only the operators', or their own, user and talk pages (or subpages thereof), and which are not otherwise disruptive, may be run without prior approval." (as long as you review the pages manually before moving them) –xenotalk 21:38, 26 August 2010 (UTC)
Thanks, I was 99% sure that was the case. But it always pays to check (rather than have to explain it afterwards :)). And, yes, to confirm all pages will be checked (that is a trivial exercise compared to hand compiling the links in the first place :P). Cheers Xeno. --Errant [tmorton166] (chat!) 21:47, 26 August 2010 (UTC)

Ganeshbot request closed by not BAG member

Snek01 has announced that Ganeshbot has approval to run [22]. It may be an excess of administratium on my part, but don't requests for approval have to be closed by a BAG member? Apologies if he is a member, but I can't see him on your table, and also as he is very enthusiastic for Ganeshbot, it might be better if a less involved member summed up. Also there was still a discussion on the page when Snek01 closed it. --Elen of the Roads (talk) 17:13, 29 August 2010 (UTC)

As far as I see he didn't close it, he simply summarized the process that the bot will run. - EdoDodo talk 17:30, 29 August 2010 (UTC)
I don't think he closed it so much as he dismissed it, "I would like to thank all for their comments," is what you say when closing a discussion at a meeting, though. JaRoad (talk) 17:53, 29 August 2010 (UTC)

Apologies. JaRoad is right, it was the way Snek01 phrased it that made it appear that he was trying to close the discussion, and EdoDodo is probably right that this wasn't his intention. Sorry for the interruption. --Elen of the Roads (talk) 22:21, 29 August 2010 (UTC)

Polite request in respect of outstanding applications

I realise it isn't normal to post here concerning ongoing applications, however two are slightly unusual in Arbcom and BAG have both had roles in defining the scope. I'm influenced by the phrase: "Sometimes you just have to give us a little prod". Please can we move forward with the following two applications that were created following suggestions at Arbcom:

Arbcom also made the following statement:

  • " Nonwithstanding remedies #7.1 and #8, Lightmouse (talk · contribs) is permitted to use his Lightbot (talk · contribs) account for a single automation task authorized by the Bot Approvals Group. "Automation" is to be interpreted broadly to refer to any automated or semi-automated tools whatsoever."

As notified at Arbcom, I've now done non-bot testing. The AWB code used in the latest tests is available at User:Lightmouse/AWB/scripts/units. This process (the BAG part) was started on 13 July and I think we're all ready to move on to the next stage. Would it be ok if I 'gave a little prod' (as suggested) so we can move to the 50 edit bot trial? Thanks. Lightmouse (talk) 16:06, 6 September 2010 (UTC)

Taking Over Bot

I have spoke with the user currently Operating the bot "Article Alert Bot" and I got an OK from him on taking over the bot. However the toolserver user would like a confirmation from the BAG. I am simply fixing the API to return a second time. I plan on operating the bot on the existing account. I know the language that most of the bot is coded in as a second language. I would like an OK from the BAG about taking over the bot. Joe Gazz84usertalkcontribsEditor Review 20:03, 12 September 2010 (UTC) P.S. Do I need an approval agian if I am just fixing the API?

I think this is probably fine, and no new approval would be needed if there was no change in the function details. –xenotalk 15:36, 13 September 2010 (UTC)
While the code will not be an issue (if it remains the same), there is also the matter of whether the prospective operator has the knowledge and experience to do the job effectively. Anomie 21:45, 15 September 2010 (UTC)
I can do it effectively, I am just waiting for the toolserver to be approved. I can file a BRFA if you like. Joe Gazz84usertalkcontribsEditor Review 23:35, 15 September 2010 (UTC)