This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

I notice that User:NVS(bot) was speedily approved as a clone of Mercury bot. Is there any reason why a single operator should use two identical bots for the same purpose? =Nichalp «Talk»= 06:57, 18 August 2008 (UTC)

I believe that User:NVS(bot) is going to replace User:MercuryBot, especially since User:MercuryBot hasn't edited since January. This, in my opinion, is a request to rename the bot. It was not made at WP:CHU because it would be a waste of resources to rename a bot, especially because bots have a lot of edits that really do not need to be moved over. All that needs to happen for a bot to be renamed is for a new account to be created, it's userspace moved, and for it to be flagged (usually by filing a BRFA, or asking a 'crat). The first two have been completed and the third will be completed shortly when a 'crat flags the bot. As always, if I am wrong or have misunderstood the situation, please let me know.  :) -- Cobi^{( t| c| b)} 08:14, 18 August 2008 (UTC)

If its a rename, in that case I would have to remove the bot flag from Mercury bot. =Nichalp «Talk»= 09:04, 18 August 2008 (UTC)

If they're doing the same thing and it's the same operator, why is the name changing? Agree with Nichalp if this goes through the old bot needs its flag removed. — Rlevse • Talk • 10:11, 18 August 2008 (UTC)

I would venture to guess because Mercury abruptly left at the end of January (last edit 1st of February, to their talk page), and then came back as NonvocalScream, but it would probably be best for NVS to explain why, as I don't know too much about what happened then. As for the bot, the MiszaBot code is proven and I personally see no problem with a redundant bot. This is not the first time this has happened. And, personally, I don't see the problem even if the bot operator were going to operate two clones at the same time. This isn't new either, a quick example of two clones run by the same operator under different bot names I found easily is Tawkerbot2 and Tawkerbot4. I'll drop a note on NonvocalScream's talk page about deflagging MercuryBot, though. -- Cobi^{( t| c| b)} 20:22, 18 August 2008 (UTC)

• To many resources to rename the bot and reattribute the contributions. I don't want to churn the database. Also, I want to follow the naming conventions NonovocalScream controls NVS(bot). I choose this for its simplicity. Please defag one and flag the other. Best, NonvocalScream ( talk) 20:57, 18 August 2008 (UTC)

I note that Rich Farmbrough asked a similar question about process in April. This page contains the terminology:

• When a user has completed a trial run (if needed), their bot has been approved, ...
• Status: In trial
• Bots that have completed the trial period

Lighbot completed its trial for its second request. I can't work out what the three bullets above mean. Is the status of Lightbot second request: approved, in trial or completed trial period? Lightmouse ( talk) 09:46, 17 August 2008 (UTC)

Completed trial. BJ ^Talk 09:48, 17 August 2008 (UTC)

That is what I thought. However, it was recently added to a table on the page with the status In trail. Is that an error? Lightmouse ( talk) 09:52, 17 August 2008 (UTC)

I assume that you are referring to this table? The system which updates this table does not differentiate between bots in trial and those which have completed the trial, as the template that it looks for on the main BRFA page, {{ BRFA}}, only has the parameters 'open' and 'trial'. I would imagine that this is because BRFAs do not usually stay in the trial completed phase for very long, and therefore having a separate parameter and status in he table for it would not be very useful. I may be wrong, however, and adding a 'trial completed' status to the table may be of help; any thoughts? RichardΩ612 Ɣ ɸ 10:05, 17 August 2008 (UTC)

Yes, I was referring to that table. The page has a a section marked Bots in a trial period and another marked Bots that have completed the trial period. The table implies that Lightbot is being treated as if the process thinks it is in the first section, not the second. Lightmouse ( talk) 10:14, 17 August 2008 (UTC)

Perhaps the table should just say "Trial" rather than "In trial"? Either that, or we need a new {{ BotTrialComplete}} template for BAGBot to parse. — Ilmari Karonen ( talk) 13:41, 17 August 2008 (UTC)

The term 'Trial' maintains the ambiguity that caused the confusion. If a 'BotTrialComplete' template could be used, that sounds like a good solution to me. Lightmouse ( talk) 13:51, 17 August 2008 (UTC)

Creating the template itself wouldn't be hard at all, but you'd have to ask ST47 to add support for it to BAGBot. — Ilmari Karonen ( talk) 22:21, 18 August 2008 (UTC)

Heck, I just went and created it. It can always be (substed and) deleted if we decide we don't want it after all. — Ilmari Karonen ( talk) 22:25, 18 August 2008 (UTC)

I think the new template is an excellent idea, but would we be asking the bot's owner to place the template when the trial is complete, or would we ask BAG members to monitor the request and then place the template? I would rather the former, as it may be hard for BAG members to tell when a trial is over if the user does not 'report back'. If this change gains consensus (and if BAGBot is updated) the main BRFA page should be updated to make this clear. RichardΩ612 Ɣ ɸ 12:13, 19 August 2008 (UTC)

Yes, this would presumably be placed by the bot operator after they've completed the trial. — Ilmari Karonen ( talk) 13:44, 19 August 2008 (UTC)

I am new to the process and I certainly expected that it was my responsibility to say when the trial edits were done. This could invite me to place a template but an option of posting a comment in plain English should also be acceptable. Part of my confusion was that this part of the process was not clear to me. Lightmouse ( talk) 17:05, 19 August 2008 (UTC)

I operate J Milburn Bot ( talk · contribs) which previously replaced 'All Music Guide' with 'Allmusic' after a rebranding. Along with All Music Guide, the smaller All Game Guide and All Movie Guide were rebranded- do I need a new task request to run my bot through the respective 'what links to' lists, or can I just go for it? J Milburn ( talk) 14:59, 2 September 2008 (UTC)

In my experience the new task would be approved in about 2-4 hours. There's probably no harm in filling out the paperwork and it will give you something to reference if anyone asks about the task. Plasticup ^T/ C 16:16, 2 September 2008 (UTC)

Ok, thanks, I've made the request. J Milburn ( talk) 21:51, 2 September 2008 (UTC)

 Done BJ ^Talk 22:07, 2 September 2008 (UTC)

I don't see it in the policy page, and there seems no way but to annotate the approval page; there's no guarantee that anyone will notice such protests. Septentrionalis PMAnderson 23:19, 22 September 2008 (UTC)

I see that the discussion concerning Whobot has already been closed. However I, being the owner, certainly have no objection to it being listed as a dead bot. If I wish to become active in the future, I will re-apply for an active bot flag.

Who ( talk) 01:47, 23 September 2008 (UTC)

Hi,

because User:DOI bot now does more than just DOIs, I've changed its username to User:Citation bot. Can this new user be granted a bot flag, please? Thanks. Martin ( Smith609 –  Talk) 02:37, 21 September 2008 (UTC)

Do you plan to use both concurrently or retire DOI bot? =Nichalp «Talk»= 06:51, 21 September 2008 (UTC)

DOI bot will be "retired". Martin ( Smith609 –  Talk) 02:56, 27 September 2008 (UTC)

 Done User:DOI bot unflagged, and User:Citation bot flagged. =Nichalp «Talk»= 05:11, 27 September 2008 (UTC)

I've just seen that me request was marked as expired. Well I have to object in this case. It is true that I wasn't able to let him run a lot (personal reasons), but I ran it, it just didn't found a lot to do.-- ThorJH ^talk 10:47, 8 October 2008 (UTC)

I have a list of redirects that I want to create using a script. I (human) produced the list and I'm feeding it to the script, which will only run once (again every once in a while when I run it). Does that count as assisted editing, or as bot editing? Note that this script will not edit any existing articles but only create some redirects where none exist, so the risk of harm is minimal to non-existent (redirects are cheap and all that).

I wouldn't be surprised if someone already has a bot that can do what I want. But even if that were the case, I'd rather do it myself as an exercise as I'm interested in creating more sophisticated bots later. -- Itub ( talk) 17:50, 9 October 2008 (UTC)

Its semi-assisted to an extent, but can be listed as automated, as the work is done automatedly after a human has made the list. — Ree dy 18:02, 9 October 2008 (UTC)

Depends on the size of the list really, if it under a few hundred and you space the edits nobody will notice or care. BJ ^Talk 18:17, 9 October 2008 (UTC)

The size of the list will be 3 or 4 thousand, but I'm not in a real hurry. I could space the edits to one per minute and it would be done in less than three days. -- Itub ( talk) 09:14, 10 October 2008 (UTC)

If the pages are all on a related topic (say, The US Civil War) it could be better to get them all done at once and flood people's watchlists just the one time, rather than flooding them over 3 days. Plasticup ^T/ C 15:43, 10 October 2008 (UTC)

Sounds good, but I don't understand how this specific case would flood people's watchlists either way. I'm only creating redirects where none existed, so it is exceedingly unlikely that anyone will be watching most of these non-existent articles. The redirects don't appear at Special:NewPages either. They do appear at Special:RecentChanges, but really they are just a trickle compared with all the other recent changes. Right now the last 50 changes occurred within 18 seconds, so at one edit per minute my script doesn't show up in the last 50 edits most of the time. -- Itub ( talk) 15:50, 10 October 2008 (UTC)

Ah, sorry. I misread your proposal; I thought you were editing existing redirects. Plasticup ^T/ C 15:52, 10 October 2008 (UTC)

I want to run a user requested task with User:John Bot and AWB. Can I just run it? I asked on IRC but was told that IRC okay it would be a bad idea. ("Bot blocked, was approved via IRC" :P). CWii_{( BOO!| Eeek!)} 22:38, 31 October 2008 (UTC)

If a BAG member approved it, then it's approved. You didn't include enough info here for anyone else to approve it, or even comment on it. — Carl ( CBM ·  talk) 22:44, 31 October 2008 (UTC)

User_talk:Bedford#Hi_there.21 My bad CWii_{( BOO!| Eeek!)} 22:52, 31 October 2008 (UTC)

The task is to tag a bunch of images as {{ NoCommons}}. That's a perfectly normal one-time AWB job, so if a BAG member gave you an informal OK and you go at a relatively slow editing rate you should be fine. Things to watch out for include images that might already be tagged NoCommons, and images that have been deleted since they were uploaded. — Carl ( CBM ·  talk) 23:08, 31 October 2008 (UTC)

Okay :) CWii_{( BOO!| Eeek!)} 23:21, 31 October 2008 (UTC)

At the moment my bots are a mixture of C# and AWB, I'm going to convert all of them to use PHP. Do I need to do a BRFA for this change or would that just be WP:BUREAUCRACY? ~ User:Ameliorate! ^{(with the !) ( talk)} 07:40, 4 November 2008 (UTC)

Well, you'd have to assure us that your bot is working (almost!) bug free and as claimed in the tasks it does. More so, if the code is not open. So, I would suggest a shorter 3-day trial where BAG members can watch over the bot functioning to see if it is working properly. =Nichalp «Talk»= 13:24, 4 November 2008 (UTC)

It is totally unenforceable so I don't see any point in requiring it. BJ ^Talk 08:03, 8 November 2008 (UTC)

I notice the request queue is now stretching for a month. I'm curious about this backlog. How do we resolve this? =Nichalp «Talk»= 07:45, 8 November 2008 (UTC)

A lot of it appears to be operators who haven't responded to inquiries or who haven't replied after completing their trial. I just when through and approved, trialed, or responded to every one possible, but short of denying for non-responsiveness, BAG can't shorten the backlog. MBisanz ^talk 07:59, 8 November 2008 (UTC)

A few days ago I went and tagged, with the "Operator assistance required" tag, all the "in-trial" bots that hadn't started the trial and there was no recent discussion. If there is no response after a week from the tagging (tomorrow I think), I'll close them as expired. Mr. Z-man 16:17, 8 November 2008 (UTC)

How would I go about requesting that CommonsDelinker ( talk · contribs) was deflagged? I don't want to write out my whole case here, but it is rather a menace that images get zapped on Commons and silently delinked by a bot. In lots of cases the images would be perfectly acceptable here, and could be undeleted, if only anyone was aware of the problem. So, what's the drill? Angus McLellan (Talk) 23:30, 21 November 2008 (UTC)

If I understand what you're saying, images on Commons shouldn't be deleted without consulting the editors on the English Wikipedia, and because of this, CommonsDelinker should be deflagged? I don't get it. -- Carnildo ( talk) 03:07, 22 November 2008 (UTC)

See WP:RFC/BOT MBisanz ^talk 03:09, 22 November 2008 (UTC)

No, I'm saying they should not be unlinked by a flagged bot, because the whole point of flagging is that it hides the edits. There are no end of images can be validly deleted from Commons but would be kept here. In a perfect world, whoever closed a deletion request at Commons for a transwiki'd image which had come from here would indicate that somewhere on enwiki so we could consider undeletion. But that doesn't happen. Until it does, let's give people the chance to spot it at this end. Angus McLellan (Talk) 12:50, 22 November 2008 (UTC)

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. RichardΩ612 Ɣ ɸ 14:22, 4 January 2009 (UTC)

Returning to do some cleanup work related to {{ wikibookspar}}, my bot complained over not having a bot flag. It seems that the standard practice of granting flags to all approved bots came after its approval. I've been looking around for where I should request for bureaucrat action to take care of this, but have come up short. Anyone able to help? -- Swift ( talk) 08:28, 15 February 2009 (UTC)

Asking at the Bureaucrat's Noticeboard would probably be the best idea, give a link to the approval and someone should be along shortly to flick the switch. Richard ⁰⁶¹² 09:10, 17 February 2009 (UTC)

Thanks! -- Swift ( talk) 11:02, 17 February 2009 (UTC)

Greetings. Polbot went through a lengthy approval process, with three separate trials, at Wikipedia:Bots/Requests for approval/Polbot 10. It was approved this morning, and I began running the bot. Polbot was then blocked by User:Docu, who didn't approve of the bot's function of creating temporary subpages. What should I do? Thanks, – Quadell ^{( talk)} 13:57, 4 March 2009 (UTC)

Docu is correct about the guideline, although perhaps someone who is going to habitually ignore other guidelines should be more polite about it. The easiest way to make everyone happy might be to just create the pages elsewhere, for example Wikipedia:WikiProject United States courts and judges/United States federal judges/fjc/J. Random Judge. They can still be merged, moved, or deleted from there easily enough. Anomie ⚔ 15:11, 4 March 2009 (UTC)

I could do that. But which would be better? Wikipedia:WikiProject United States courts and judges/United States federal judges/Henry Baldwin or Talk:Henry Baldwin/fjc? – Quadell ^{( talk)} 15:36, 4 March 2009 (UTC)

My previous comments are at User_talk:Polbot#Subpages_in_article_namespace.

Personally, I have no preferences where the pages are created, if they are created outside article namespace. An advantage of Anomie's suggestion is that they can all be accessed with Special:PrefixIndex/Wikipedia:WikiProject United States courts and judges/United States federal judges/.

Shorter pages titles would probably be preferred. It could also be at Special:PrefixIndex/User:Polbot/fjc/. -- User:Docu

I like that last idea the best. I'll change the code, unblock the bot, and re-run her. – Quadell ^{( talk)} 15:51, 4 March 2009 (UTC)

Ok, I deblocked the bot. -- User:Docu

That solution seems satisfactory to all - I will keep an eye on those project subpages and get them moved to the mainspace as quickly as possible. Cheers! bd2412 T 17:36, 4 March 2009 (UTC)

Should we have another section for when a trial is concluded and awaiting BAG feedback? Rich Farmbrough, 13:19 27 April 2007 (GMT).

I'd appreciate input here. Dorftrottel ( talk) 09:04,  March 28, 2008

Moved to Wikipedia:Bot owners' noticeboard#TinucherianBot

I have a list of my GA reviews, and it would be helpful if I had an automated tool to add to that list whenever I pass/fail an article. If a bot I make only edits my userpages, do I need an approval? Noble Story ( talk • contributions) 02:35, 21 March 2009 (UTC)

How often would it be making edits? §hep ^Talk 02:37, 21 March 2009 (UTC)

As I said, only when I pass an article, which would mean, at most, 2 or 3 times a day. Noble Story ( talk • contributions) 04:07, 21 March 2009 (UTC)

I wouldn't think you'd need to get approval for that. 98.31.12.146 ( talk) 02:54, 1 April 2009 (UTC)

I would like to appeal a decision by someone to deny the use of assisted editing scripts. What is the actual procedure to go about it? Would I simple put in another identical request and then reply to it asserting the reasons why it should be reconsidered? Should I reply to the original request, overriding the bold red "Please do not modify it" request? Should I put a subsection for appeals? Where are the previous appeals anyways? Int21h ( talk) 01:58, 31 March 2009 (UTC)

According to WP:BOT#Appeals and reexamination of approvals, you post here. We'll need details. Anomie ⚔ 02:09, 31 March 2009 (UTC)

Well, on a more liberal reading of the policy, my request ( Wikipedia:Bots/Requests for approval/Int21hBot) fits more of the assisted editing class of uses along the lines of WP:HG, and as such "are not usually considered to constitute use of a bot." I plan on using PyWikipedia to do various random, custom, one-time tasks that are tedious and difficult otherwise, and according to Bjweeks that is not allowed.

By the time I got through some of the past requests and all the bureaucratic nonsense surrounding the use of "bots" and figured out I shouldn't have needed to request anything in the first place, he had already closed the discussion after 1 other comment (and seems to have deleted another one) and zero revisions and denied it (less than two hours later.) So I'm going to go ahead and ignore the denying authority, Bjweeks, and instead appeal to WP:IAR. What I'm planning to do will not even be recognizable as automated, much less interfere with anything on Wikipedia, and will help me make Wikipedia better, which is the point of all this. What I'm afraid of is the multitude of Admins that will wield the banhammer over me using scripts because of misunderstandings of the policy, so I will be using the declared bot account. Someone should flag the account as a bot if it is customary. Int21h ( talk) 05:14, 31 March 2009 (UTC)

That BRFA was correctly denied: we don't approve bots to do "whatever I feel like, I promise I'll be good". If you edit under the bot account without approval, it will be blocked. If you perform script-assisted edits under your own account, you will be fine. If you perform edits that are clearly bot edits under your own account, you may be warned and/or blocked.

The general definition of "script-assisted edit" is that you need to be manually approving each edit after seeing a diff or otherwise verifying that the edit is actually correct. Anomie ⚔ 11:33, 31 March 2009 (UTC)

All this sounds amazingly like Lord of the Flies. But I will use the dedicated account (that regretabbly now has the "bot" word in its name), so as to make any messups on my part with using scripts easier to revert. Blocking this account for editing using scripts ^[1] is against consensus and I strongly urge you or anyone else to refrain from doing so without good reason. Int21h ( talk) 22:05, 31 March 2009 (UTC)

So, which are you claiming doesn't have consensus? Wikipedia:Username policy where it says "Your username should not give the impression that your account has permissions which it does not have. Thus it should not [...] end with 'bot', which is used to identify bot accounts", or Wikipedia:Bot policy where it says "Accounts performing automated tasks without prior approval may be summarily blocked by any administrator"? Or both? In any case, you'd do best to bring up the matter at WP:VPP. Anomie ⚔ 01:40, 1 April 2009 (UTC)

So, OK, I can agree with that. But I would like a separate account from my own that both clearly established who the author is (Int21h) and it's purpose (scripted editing). How else do I go about that then? Adding an "Auto" on the username? Int21h ( talk) 05:07, 3 April 2009 (UTC)

I don't think there's any policy or common convention for that sort of thing, just tag it with {{ User Alternate Acc}} or one of the other templates in Category:Alternate Wikipedia account templates (or construct a similar message yourself). Anomie ⚔ 12:10, 3 April 2009 (UTC)

You seem confused by the definitions of bot and script (which don't make a great deal of sense). A script is a piece of software (generally user JavaScript, but some externals applications as well e.g AWB) that requires a user to confirm the edit at the time it is made ("clicking the save button"). A "bot" is anything else. Bots can be manually assisted, that is have their input reviewed by a human and then fed to them (e.g. image moving bots) but they remain bots. Scripts are considered an extension of the user and all edits are their direct responsibility.

Bots are required to be on separate accounts and have all major tasks approved (established bot operators get some leeway with running simple jobs that are related to the bots main function). Making any edits on a bot account that lacks approval is going to lead to a block, as will running bots on your main account. If you intend to write pywikipedia programs that meet the definition of script (pywikipedia contains a diff printing method for manual conformation) you can run them on your main account with no problems. Otherwise you should seek specific approval.

With that said the bot procedures can be silly at times and are enforced very liberally for bots that nobody has a problem with. However coming here to invoke WP:IAR isn't a great idea. BJ ^Talk 01:44, 2 April 2009 (UTC)

I noticed a thread about a bot that someone wanted shutdown until it was approved (that might not me exactly the right wording, but it gets the point across). The thread is/was here. Since I know absolutely nothing about bots, I thought I'd come ask you folks. Is there a process in place to turn bots off, or block bots? - which would be the exact opposite of what you "BAG" folks do. If there was, I figured that you peeps would at least be able to point me in a good learning direction. Thanks, and Cheers. — Ched :  ?  02:09, 23 April 2009 (UTC)

If there is clear consensus a bot is misbehaving an admin should just block the bot indefinitely, until the operator addressed the concerns. – xeno ^talk 02:12, 23 April 2009 (UTC)

In this case, the bot was approved to do a certain task, and a few details were fuzzy. One person doesn't like the way the bot chooses to do those details, and claims there is no consensus to do it the way the bot is doing it. So far I only see one person disagreeing, and a lot of people saying it's not a big deal either way. – Quadell ^{( talk)} 02:23, 23 April 2009 (UTC)

(ec)Well, actually in this case it's an admin running the bot, and an admin wanting it stopped. I didn't know if there was a general procedure in place or not - but given the situation. I think I'll just stick to the "fly on the wall" routine, and see what happens. Thanks Xeno, appreciate the input. ;) — Ched :  ?  02:25, 23 April 2009 (UTC)

Oh .. and you too Quadell ;) — Ched :  ?  02:26, 23 April 2009 (UTC)

Well there's no rule against an an admin blocking another admin's bot, but in this case the specifics of the task are indeed fuzzy. I left a message at WikiProject Good Articles so that should help to determine whether the bot should be doing that. – xeno ^talk 02:37, 23 April 2009 (UTC)

I have blocked a bot that was approved yesterday, MauritsBot, because it was breaking thousands of pages by putting interwikis on templates outside of noincludes. Hesperian 03:58, 29 April 2009 (UTC)

Looks like all the bad edits have been reverted by Hesperian. The stray interwikis probably broke thousands of pages, but MauritsBot had only made a little over a hundred edits before it was blocked. Wronkiew ( talk) 05:17, 29 April 2009 (UTC)

Mauritz has indicated to me that s/he will fix the problem before running the bot again. I am satisfied with this, but would prefer that the decision whether to unblock be taken by a BAG member. Hesperian 09:12, 29 April 2009 (UTC)

Wow, that's too bad. The trial didn't show any such problems, when I looked through the edits. Thanks for taking care of it, Hesp; I've unblocked the bot. – Quadell ^{( talk)} 12:09, 29 April 2009 (UTC)

It may be a good idea to remind Interwiki.py users not to work in the templatespace; this is a fairly common occurance. I wonder why they haven't just written something into the code to stop it working there without some kind of forced override. – xeno ^talk 12:36, 29 April 2009 (UTC)

Hi, I just whipped up a hacky tool in my toolserver folder that can easily generate the title and text for a new BRFA. This is designed for bots that have a gazillion tasks, and you always forget which number you're on. X clamation point 23:45, 3 May 2009 (UTC)

Examples:

• New SoxBot task: http://toolserver.org/~soxred93/new_brfa.php
• New AnomieBOT task: http://toolserver.org/~soxred93/new_brfa.php?bot=AnomieBOT
• New ListasBot task: http://toolserver.org/~soxred93/new_brfa.php?bot=ListasBot
• New Sambot task: http://toolserver.org/~soxred93/new_brfa.php?bot=Sambot

Could someone please tell me what the process is (if indeed there is a process) for getting the approval of a bot task overturned? ListasBot's third task was approved after a discussion involving only two editors (the bot owner and one BAG member), and I feel the the decision may have to be reassessed. Thanks, — Martin ( MSGJ ·  talk) 15:05, 9 May 2009 (UTC)

Block it, and then create discussion somewhere - an RFC, talk page, or VPM would be fine. - Jarry1250 ^{( t, c)} 16:08, 9 May 2009 (UTC)

Actually, the official "process" at this time is to begin a discussion on this talk page. Immediate blocking would IMO not be appropriate in many cases; that should only be done when the bot is actively causing harm or acting against consensus and not just because any random admin wants to start a discussion. Anomie ⚔ 21:18, 9 May 2009 (UTC)

Sorry, I was assuming harm was being done. No, you're quite right, better to discuss first. - Jarry1250 ^{( t, c)} 21:27, 9 May 2009 (UTC)

FYI, I've moved this discussion here. Thanks, Matt ( talk) 03:25, 12 May 2009 (UTC)

Unfortunately it was decided not to approve my Bot to its main purpose which is making categories out of lists. I would still need a bot flag for the task of putting Hebrew Interwikis so the bot can operate correctly in the Hebrew Wikipedia. Tomer A. 13:58, 15 May 2009 (UTC)

You may file another BRFA as "Tomerbot 2" if you just want approval to do IW. – xeno ^talk 14:05, 15 May 2009 (UTC)

Yeah, that should be quick and uncontroversial. – Quadell ^{( talk)} 14:45, 15 May 2009 (UTC)

10x. Tomer A. 20:08, 17 May 2009 (UTC)

A number of bots rely on the {{ DEFAULTSORT}} or {{ Persondata}} parameters, or the listas parameter in biography templates. The data encoded in these templates may be reliable for individuals whose names are in the European naming style of inherited surnames.

Unfortunately, over the years, well-meaning, but ill-advised volunteers have mistakenly assumed Arabic and Chinese names should also be shoehorned into the European naming scheme. This is a problem. And bots that treat this data as reliable are compounding an already serious problem.

I suggest no bot that relies on this unreliable data should be approved. Geo Swan ( talk) 08:17, 27 May 2009 (UTC)

People reading this should probably look at WP:Administrator's noticeboard/Incidents#A rogue bot ( most recent diff) and User_talk:Mikaey#It is a big mistake to try to shoehorn traditional Arabic names into the European naming scheme of inherited surnames... to get an idea of what's going on. Matt ( talk) 08:26, 27 May 2009 (UTC)

The Defaultsortbot is only one of the bots recently approved that compounds this mistake. The recently approved listasbot also relies on the unreliable data in these parameters. Various robot-assisted editing tools rely on the unreliable data in these parameters. Geo Swan ( talk) 09:14, 27 May 2009 (UTC)

As you have been told several times already in the linked discussions, the bot is not adding DEFAULTSORT de novo; it is only copying what a human editor has already placed in a listas parameter, persondata template, or the like. There is not much that can be done about GIGO, so unless you can provide evidence that this problem is widespread enough that a significant fraction of the bot's edits will be erroneous I can't see that anything needs to be done here. One possibility, if you can supply a sufficiently accurate heuristic, would be for the bot to ignore or log pages with seemingly-Arabic names for human review.

BTW, Geo Swan, I notice in your contribs that you recently went through a number of articles on people with Arabic names and removed the listas parameter from {{ WPBiography}}. It would have made much more sense to replace it with the "correct" value instead, possibly also with a comment pointing out why that is the correct value. Otherwise, well-meaning but misinformed editors are likely to repeat the same mistake. Anomie ⚔ 12:26, 27 May 2009 (UTC)

Let be clear about impatience my raising of the corrupt information the bots are relying on has generated. Don't shoot the messenger. This situation is an enormous mess, and it is a mess I do not believe I played any role in generating. I regard this impatience as inappropriate. I would appreciate those expressing it to reserve any expressions of impatience until after we have discussed the various suggestions I have been getting about what I did wrong.

I know I have been told that the various bots and humans assisted by robot editing tools are not naively shoehorning bogus surnames into articles about non-Europeans "de novo". And I know this is incorrect. Here is an instance:

If you look at the edit summary that includes "AWB", it means that it was actually Wikipedia:AutoWikiBrowser that suggested that, in the attempt to add the "listas" section to the template. I didn't know it was incorrect, so I let the tool do its thing.

User:Raven1977 was responding to a question I asked him about his or her addition of a bogus listas to Talk:Abbas Habid Rumi Al Naely. A few days earlier another contributor had added a bogus listas to Talk:Abbas Habid Rumi Al Naely. I asked him or her why. And that first contributor told me their edit was based on the (bogus IMO) value in the article's {{ DEFAULTSORT}}. I subsequently removed the article's {{ DEFAULTSORT}}. So this second guy, with the assistance of AWB, was naively adding these bogus values "de novo". I corrected the articles this second guy was naively munged, that were on my watchlist. When I looked at his most recent contribution history I saw he or she had recently added listas fields to SEVERAL HUNDRED other articles on individuals with Arabic names.

Their comment, "...so I let the tool do its thing." As you stated above, a human should be in the loop. A human should be making an informed choice as to whether to follow, or over-ride, the bot's suggestion. Alarmingly Raven1977 was abdicating responsibility to provide the informed decision-making. I wrote the author of AWB. He told me that AWB only suggested inserting a default sort, or listas, when a human had already made an informed choice that one was appropriate. So, how come Raven1977's use AWB was suggesting bogus values "de novo"? I dunno.

I am afraid you are mistaken. Early in the history of placing DEFAULTSORT and listas parameters there were bots that went around placing wild guesses at what individuals surnames were, without consideration to the billions of individuals who had names that didn't fit in that scheme. User:Jim Cubb acknowledged this on my talk page, and I have come across records of these bot's operations myself. So, no offense, I do believe a "significant fraction" of the bot's operation will be erroneous.

As I tried to point out to Raven1977, when he or she was compounding the problem by relying on unreliable advice from AWB, they were affecting 3 articles per minute. In my limited attempts to clean up after Raven1977 I was averaging about one article every three minutes. See the problem? The clean-up is much more time-consuming than compounding the error in the first place.

Anomie, implicit in your comment is the assumption that treating individuals with European style inherited lastname-surnames should be considered the standard case. I think that is backwards. Billions of individuals have names best sorted starting with the first character, proceeding smoothly to the last character. Arabic influenced names aren't the only ones. 1.3 billion Chinese people have surnames -- but their surnames begin their name, not end it. It is European style names that need to be treated as the exception -- not the default. In my opinion these bots should be leaving alone all biographical articles unless there is some reason to believe the individual's name uses the European style.

Various other people have suggested that, instead of removing the bogus listas and DEFAULTSORT values, I should have replaced them with "correct" values. No one has asked if I think this is a good idea. Actually, I considered doing this. I considered it, and decided it was a really bad idea. I have about 700 1000 articles about individuals with Arabic names on my watchlist. The more references I come across to these individuals, the more choices I see to transliterations and renderings we would need to chose among to pick a base name the other names should be redirected to. Most of those 1000 names have been moved, at least once. When I choose, or participate in the choice of which alternative transliteration should be at the base name, the choice is largely arbitrary. And it is likely to be reversed, or superceded.

It is safe assumption that articles about individuals with Arabic names are likely to be renamed. So, adding in the current article name, in the listas field, and in the DEFAULTSORT parameter, is simply an invitation to an additional maintenance burden. In my opinion, it is a totally unnecessary one. In my opinion it should be the European-styled names that should be considered the exception. Non-European-styled names should be able exist without a mutable, troublesome defaultsort, when the sort key for them is identical to whatever the current name of the article is, not the name that was current when the bogus Europeanized surname was shoehorned in.

Let me be frank. Whoever first thought up the idea of automating sorting, made a mistake when they didn't foresee that the billions of individuals who don't have European style inherited lastname-surnames would require a tag, or some other mechanism, to tell bots this individual should not be treated as if they had a European style inherited surname. There should have been a template with a name like {{ NoDefaultSortOk}} or {{ NoListasOk}}. This should have been the default. And, I suggest, even though it is rather late in the day, it would make sense to establish a convention that the listasbot, the defaultsortbot, AWB, Kingbotk, and all other bots leave alone artcles with a tag that tells them the article is not someone with a European style name.

Candidly, Geo Swan ( talk) 17:00, 28 May 2009 (UTC)

WP:TLDR. A long rant, examples of human editors screwing up, and assertions without evidence will not help your case. If you want something done, you'll have to provide a clear, concise suggestion of a positive course of action. Random assertions about the population of China mean absolutely nothing; more useful would be the number of articles on the English Wikipedia with "European" versus "non-European" names. Anomie ⚔ 19:37, 28 May 2009 (UTC)

It's true that it's a problem. I hope that we can recruit people knowledgeable in Arabic (and Persian) conventions to confirm the sorting of these names. – Quadell ^{( talk)} 13:04, 27 May 2009 (UTC)

There sure seem to be a lot more pages in Category:Open Wikipedia bot requests for approval than are listed as still open on this page. -- Pascal ⁶⁶⁶ 05:44, 29 May 2009 (UTC)

I went through and removed those that had been approved/denied, but someone else needs to have a look see which if any have slipped through the net. - Jarry1250 ^{( t, c)} 06:38, 29 May 2009 (UTC)

There are only 15 there at the moment, and they all appear to still be open; nothing looks to have slipped through the cracks. Useight ( talk) 03:38, 3 June 2009 (UTC)

I added "Estimated number of pages affected" to Wikipedia:Bots/Requests for approval/InputInit. It's something I've been meaning to do for ages. Let me know if there are any issues. -- MZMcBride ( talk) 22:29, 7 June 2009 (UTC)

What should be entered for any task that's intended to be an ongoing task? For example, User:WebCiteBOT, User:ClueBot, and so on. Anomie ⚔ 23:47, 7 June 2009 (UTC)

n/a? Or perhaps 1 page per day or whatever. Obviously this won't apply to all bots, but it is something that comes frequently and the question is rarely asked enough. If it doesn't apply, bot ops can skip over it or write "n/a". Though I'd imagine even bots like ClueBot have an estimate of how many pages they edit per day. -- MZMcBride ( talk) 01:00, 8 June 2009 (UTC)

I've gone ahead and clarified that in the comment. Although really estimating for some bots would require a good deal of study to determine beforehand. Anomie ⚔ 02:31, 8 June 2009 (UTC)

Looks good to me. :-) -- MZMcBride ( talk) 02:47, 8 June 2009 (UTC)

With the completion of the date delinking arbitration case, I was just wondering what next steps, if any, needed to be taken by BAG. Lightbot ( talk · contribs) hasn't been banned directly, but that's the effective result of the remedies related to Lightmouse ( talk · contribs). Just wondering if it's appropriate for BAG to revoke any task approvals for Lightbot that still exist, and ask a 'crat to remove the account's bot flag and block the account in accordance with the remedies? I also note that Lightbot isn't listed anywhere on the status page. Mlaffs ( talk) 15:43, 15 June 2009 (UTC)

Yes, the 'crats should remove the bot flag and an admin should block the account, as Lightmouse is indefinitely prohibited from using any sort of automation. As far as I'm concerned, Lightmouse must (1) successfully appeal to ArbCom to have that prohibition revoked, and then (2) request (re-)approval before running any sort of bot in the future. Anomie ⚔ 02:01, 16 June 2009 (UTC)

User:fl blocked it, but a crat still needs to deflag. – Quadell ^{( talk)} 12:51, 16 June 2009 (UTC)

Requested at WP:BN#Lightbot. Anomie ⚔ 14:12, 16 June 2009 (UTC)

Done by Dweller Anomie ⚔ 00:24, 17 June 2009 (UTC)

The anybot owner says it didn't change the articles this time with the same bot but with "a different bot." [5] I don't see the bot approval discussion in Martin's edit history. [6]

Please explain the basics of this to me in relation to this group. It seems each bot requires approval by the board on the project page, then flagging by a bureaucrat to operate. And that this must be done for each new task. If this is "a different bot" shouldn't it have a request for approval?

There are a lot of subpages for this project. This post is about "requests for approval," so I think this is the correct place to ask this. Was this bot approved? Did it require approval? Or was I mistaken in thinking that because the bot was blocked the issue could safely be discussed without additional problems arising? -- 69.226.103.13 ( talk) 07:23, 25 June 2009 (UTC)