Wikipedia talk:Bot policy/Archive 22

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia


Not the BAG nomination system

WT:BOT has been stuck discussing the BAG nomination system. There were quite a few other changes to WP:BOT. I'm listing the changes I noticed in this version of the live draft. Some rather significant new processes or policies are:

I'm not even sure what the "extended trial" is supposed to mean or how it might apply. Is the multi-user idea acceptable to everyone? I think the user talk opt-out could go under configuration tips, but does it need to be a requirement?

Other possibly notable changes: a line saying "Bot operators may wish to create a separate bot account for each task" was removed. A statement was added about Wikipedia:Bot_policy#User_scripts, and there is a new Wikipedia:Bot_policy#The_.27bot.27_flag section.

Do we agree with all of these? Do we no longer want to even mention that it can be OK to have different bot accounts for different tasks? The bot-flag section is mostly descriptive except that it implies BAG will no longer approve any bots to run without the bot-bit. Regarding "User scripts" in particular, if a javascript locks up a user's account for a long time making the user unresponsive to the queries on the operation of this javascript, should bot policy apply? Should WP:BOT say clearly yes or no, or should this be left ambiguous?

Also, I've been thinking about a model where BAG functions rather like a mentor. If an approved bot has technical issues later, then a BAG member (or more in exceptional cases) could take responsibility to watch over it until issues are resolved. Gimmetrow 04:15, 7 May 2008 (UTC)

I would like the Wikipedia:Bot_policy#Appeals_and_reexamination_of_approvals to include something on the proposed WP:RFC/BOT process, since that would give discussions greater community visibility than WT:BOT or WT:BAG would. MBisanz talk 06:18, 12 May 2008 (UTC)
Yes, the opt out needs to be a requirement. —Locke Cole • t • c 06:38, 15 May 2008 (UTC)
We can require it in specific cases. AWB is {{bots}} compliant, I am told, so many bots have a built-in opt-out. I think we're fairly clear on encouraging opt-out, but requiring it could be a problem for legacy bots. Gimmetrow 06:56, 15 May 2008 (UTC)
I'm not positive, but I'd like to think this requirement would apply to such a small subset of existing bots that it really shouldn't be that big of a deal to implement it. Besides, I'm not sure it's wise to give a pass on this either to existing bots or to new bots (by not making it a requirement now). —Locke Cole • t • c 07:30, 15 May 2008 (UTC)
It wasn't a requirement before, so if it's going to be it now there should be a consensus for it, don't you think? Is there? Gimmetrow 07:55, 15 May 2008 (UTC)
Not every minor change to policy requires a poll to enact (especially when the benefits are obvious and the requirement mostly painless). But yes, I do believe there is consensus for this to be a requirement. —Locke Cole • t • c 08:15, 15 May 2008 (UTC)
No, but it needs a discussion where people agree to it. If two people at some obscure page changed something and wrote into "policy" a "requirement" that affected you, would you feel bound by those changes? Gimmetrow 08:37, 15 May 2008 (UTC)
By that logic, we can put an end to the BAG right now because it was never subject to a community wide discussion which resulted in consensus for it. FWIW, this was a topic that received some attention at the last Betacommand ArbCom case, and my impression was that there was support for this (beyond "two people"). —Locke Cole • t • c 08:54, 15 May 2008 (UTC)
Let's say, hypothetically, there was no discussion whatsoever for the formation of a BAG. There was an existing requirement that a bot be approved, and there was in fact an editor doing the approving. (BAG was formed shortly after that editor retired, as I recall.) BAG's formation did not change a requirement for anyone, and it didn't really change the process from a bot operator's perspective. Can you not see how adding a new requirement is slightly different? But in any event, point me to the appropriate subsection of that arb case, if you think it shows consensus. Gimmetrow 09:05, 15 May 2008 (UTC)
{{bots}} has been rejected as policy several times. its a very very bad system which most operators will not follow. βcommand 13:28, 15 May 2008 (UTC)
I know that, Beta. But I think LC is proposing opt-out in general, and not a specific technique. Gimmetrow 19:08, 15 May 2008 (UTC)
This is exactly what I'm proposing. BTW, {{bots}} was rejected by you and other bot operators, I don't think the community cared one way or the other how it was done, just that there was a way. —Locke Cole • t • c 00:50, 16 May 2008 (UTC)

Bot Approvals Group

I edited this section to reflect the lack on consensus but at the same time this should not imply that the current membership is frozen and that there is no way of joining. Users should have a means of requesting membership if desired. Dbiel (Talk) 01:35, 16 May 2008 (UTC)

Just because one user keeps shouting that there is no consensus does not make it so. Just because the RFA method did not acheive consensus does not mean the presently used method has none. I would ask you to return the policy to reflect what's actually in use, please (6 closed today, I think, 5 passing, and added, btw) SQLQuery me! 01:44, 16 May 2008 (UTC)
As I said above: when there's no consensus for a new system, we use whatever is in place. We don't just say "there's no consensus for anything" and then have nothing. If we did that everywhere, we would never have any stable policies or processes. Mr.Z-man 02:11, 16 May 2008 (UTC)
Consensus can change, though. It's not helpful for a group of editors to stonewall any changes to the process they created by disagreeing with all changes and then pointing out that there's no consensus. The BAG needs to help in coming up with a serious way to reform themselves. rspeer / ɹəədsɹ 05:20, 16 May 2008 (UTC)
As I said above, making this into an Us vs. Them battle is really unhelpful. There were over 30 users on the Adminship poll who opposed putting BAG membership on RFA. So far, the only change that has been proposed is moving BAG membership to RFA. I opposed it, I gave a reason. I believe I was dismissed as "momentum" by people trying to force through the change despite all the opposition. I gave up trying to argue. The poor attempts at "discussion" are on both sides of the dispute. Mr.Z-man 05:46, 16 May 2008 (UTC)
It doesn't have to be an "us vs. them" battle. Help come up with another way. rspeer / ɹəədsɹ 06:47, 16 May 2008 (UTC)

I totally agree. And it is exactly what I attempted to do. The only thing I took out was the 10 day reference and who would close the discussion. A fixed 10 day window is stupid. If discussion is continuing, it should be allowed to finalize, not be locked into a specific time frame. Dbiel (Talk) 04:26, 16 May 2008 (UTC)

I must have missed something then, no problems here with no specific time limits. SQLQuery me! 08:55, 16 May 2008 (UTC)

Crat closures

Is it possible for us to agree on anything? If it is, I would expect that this would be it. Yesterday Kingturtle closed six RfBAGs on WT:BAG, making five promotions. I don't have a fault to find with his judgement in these evaluations, and it took him less than half an hour. Even in a nomination as obviously successful as MBisanz', I am much more comfortable with seeing it judged by an impartial outsider; doubly so for the less unanimous ones. I will freely admit that I do not hold some of the past RfBAGs to be particularly good textbook examples of "consensus"; remembering that consensus is not the same as unanimity. While I don't envy the closers of some of the less well-supported RfBAGs the job of balancing the need not to let a request languish, with the need to establish consensus, that makes me more certain that the job should fall to someone who we have directly mandated (usually with a massive and near-unanimous consensus) to perform precisely that circus trick. Is it possible for us to agree that having all RfBAG nominations, wherever and however they are conducted, closed by an impartial bureaucrat, is a Good Idea™?? Happy‑melon 08:39, 16 May 2008 (UTC)

Wait, who's opposing 'crat closures? --Chris 08:43, 16 May 2008 (UTC)
I really hope no one, but you can never be sure on this page :D Happy‑melon 09:01, 16 May 2008 (UTC)
I changed the policy to this some months back and the crats indicated they weren't interested in closing unless it was a controversial close, so I changed it back to permit BAG to close clear-cut matters. Kingturtle was just being nice I think closing all of them. MBisanz talk 08:48, 16 May 2008 (UTC)
While I don't agree with all the closes, we entrust the crats to gauge consensus, and, it's been in there (the bot policy) for a while that they close controversial BAG noms. Kingturtle came forward and did that, and, I appreciate the effort. SQLQuery me! 08:53, 16 May 2008 (UTC)
(ec)As I said somewhere above, I think that's rather irresponsible of them. We appointed them because we trusted them to evaluate consensus: they are still answerable to the community (all the more so, in fact). I naturally recongnise that the crats are volunteers just like the rest of us; but as Kingturtle has shown, it's really not that difficult, and doubly so when the consensus is clear. It only took Kingturtle one minute to close your RfBAG. I actually think it's very unreasonable of them to claim that half an hour of their time every couple of months is too much to ask. If the bot-operating community, as a facet of the wider community which gave them their 'crat tools, asks them to do something with that status, then they should honour that request. Happy‑melon 09:01, 16 May 2008 (UTC)
Uhm, I'm not sure I follow? SQLQuery me! 09:02, 16 May 2008 (UTC)
I'm just saying that, if we establish a consensus that the crats should close all RfBAGs, then it's not really acceptable for the crats to turn around and say "tough, we're not doing it". Presupposing, of course, we can establish a consensus :D Happy‑melon 09:13, 16 May 2008 (UTC)
I think MBisanz has it about right. Any BAG member, or any community member in good standing for that matter, can close clear-cut requests when they're done, and the rest can be decided by the 'crats, who we've placed our trust in as a community to gauge consensus. krimpet✽ 09:16, 16 May 2008 (UTC)
That's the way it is now, yes. The question I'm trying to get a discussion on is, should that change? Happy‑melon 11:39, 16 May 2008 (UTC)
I for one don't. The current setup seems to work fine, and it seems rather feature creep-y and against the wiki spirit to have to call in a 'crat to do a job that anyone - a BAG member or non-member, admin or non-admin, could do as soon as an uncontroversial BAG request has passed the usual time limit. krimpet✽ 11:50, 16 May 2008 (UTC)
For my part, I don't think it's as easy as it looks, and I don't honestly think it's worked very well in the past. Just because a nomination has almost unanimous support, that does not mean it has consensus if it only has three or four supporters. All the RfBAGs that Kingturtle closed had way over the average number of contributors to the older RfBAGs (ie, the ones closed by BAG members). I don't know about you, but I just don't think that a nomination with five or six contributors represents consensus, no matter how long it's been open. There is no way I would have passed Snowolf's RfBAG, even though it's unanimous. Coren's RfBAG is a joke, as is Reedy's just below it (to be clear, I would have supported all of these if I'd been active here at the time). This area of bot policy seems just as confused as all the others; for every other discussion on wikipedia, the default ("no conseus") option is the status quo, which in this case means they should not be promoted. RfBAG, whatever BAG is supposed to do and however they're supposed to do it, surely must be about obtaining a consensus that the candidate is suitable?? Unanimity does not equal consensus, nor does silence in user-related discussions (and especially when we're trying desperately to get more eyes on the issue). Bureaucrats understand this instinctively - that's why they passed their RfBs. BAG members, in a fair few cases, don't appear to grasp the distinction. Of course, if the real issue here is that RfBAG is not about establishing a consensus for promotion, then that's something we need to be absolutely clear about. Happy‑melon 13:56, 16 May 2008 (UTC)
BAG requests with little participation should then be treated like relisted AfDs - if an unsatisfactory amount of people have contributed to the discussion to clearly gauge consensus, let it go on for a couple more days. :) krimpet✽ 14:02, 16 May 2008 (UTC)
My thoughts entirely - there is most definitely no deadline for things like RfBAG. My point is that, while I would trust a bureaucrat to know from long experience when a request needs to be left to run, BAG members have shown on more than on occasion a slightly worrying inability to come to the same conclusion :D. Happy‑melon 14:52, 16 May 2008 (UTC)

I can't stand 'no consensus equals consensus for the status quo'. It fills Wikipedia with incredible inertia. — Werdna talk 00:05, 17 May 2008 (UTC)

List of current bots and the tasks they are performing?

On Wikipedia:Bot requests, in the opening paragraph, it says:"You might also check Wikipedia:Bot policy to see if the bot you are looking for already exists." Maybe I'm overlooking something really obvious, but I can't find any such list on this page. Has it been moved somewhere else? And if there is no such list, how would I go about creating one?--Aervanath's signature is boring 17:29, 25 May 2008 (UTC)

There is Wikipedia:Bots/Status, but its pretty outdated. Mr.Z-man 18:27, 25 May 2008 (UTC)
Thanks.--Aervanath's signature is boring 19:10, 25 May 2008 (UTC)

I would think it might be a good idea to change the bot policy page to make it the responsibility of the bot operator to keep is entry on this page current. Dbiel (Talk) 21:16, 25 May 2008 (UTC)

Good idea, but it leaves out the part where the botop doesn't care anymore. This may be recursive, but perhaps a bot to check bot contribs and update the bot status? Another idea would be to make maintenance of the bot approval contingent on an entry in a "Current bots" page where the operator has to list current status and tasks, and expected run frequency, and if this is not kept up-to-date, approval is revoked after some arbitrary time period based on the frequency. In other words, if you say you expect to run the bot once a month and don't do anything for two years, it's no longer an approved bot. Franamax (talk) 22:18, 25 May 2008 (UTC)
Hmm, prior discussions have been not to revoke Bot status unless its abused or the owner agrees, so I don't think we could do that. I do intend over the next month to get that page up to shape, so maybe hold off until I forget :)? MBisanz talk 00:24, 26 May 2008 (UTC)
The problem is enforcement. I think you would have a fairly difficult time getting consensus to block a helpful bot just because the operator hasn't filled out all the proper forms yet. Mr.Z-man 08:36, 26 May 2008 (UTC)
Rick Bot updates WP:LA based on admin activity. It wouldn't be difficult to do the same for Wikipedia:Bots/Status. Addshore is restructuring the page - after it's stabilized I could add a task to update it (daily seems plenty often enough). -- Rick Block (talk) 16:02, 26 May 2008 (UTC)

How about a bot that goes through transclusions of {{Infobox Bot}} and uses that information to update the list? -- maelgwn - talk 05:47, 27 May 2008 (UTC)

Quick question

Which bot handles deletion tagging of orphaned fair use images? I've run across a few non-free images that seem to have been orphaned a while without being tagged - from what I recall, those used to be mopped up daily. Kelly hi! 12:39, 1 June 2008 (UTC)

Ah, never mind. Looks like BJBot is doing a run. Kelly hi! 15:34, 2 June 2008 (UTC)

Perl packages

Hi, sorry about spamming - if anyone can help here, i'll be very thankful: Wikipedia talk:Creating a bot#Perl_packages. Thanks. --Amir E. Aharoni (talk) 08:06, 2 June 2008 (UTC)

Eliminate BAG

The Bot Approvals Group has been by far one of the most argued over formal groups on Wikipedia. It's existences seems to come out of the will of 4 users who thought more process was needed in Wikipedia (link 1, link 2) In short, this process was never "ratified"/approved by the community.

While the idea of having some way to check the validity of a code is worthwhile, the community, not a group of users who approve new members themselves, should make the final decision in approving a bot. A formal group like BAG does NOT need to exist to check the validity of a code and the worth of a bot.

The process for joining BAG is shady and switches whenever their control is threatened. Previously discussion took place on the talk page of BAG. However, once a previously MfD was created about BAG and the community complained over the "cabal" of the process, BAG sought to fix this by allowing anyone to join. This was short lived, and BAG went back to the old way of adding users. Soon enough the community cried out again over the cabal nature, and BAG added itself to the RfA main page. Again, when one of the current BAG members would of failed joining BAG, (see link for ST47) the group switched back to the old way, which takes place on a unwatched talk page, of approving members.

This unwarranted and unapproved process needs to be stopped. No more reforms, no more process wonk. The community should decide the fate and usefulness of a bot, not a selective group of users. If, indeed, later down the road the community would like to have a group oversee the validity of a code, then a whole new community approved process can start.

Eliminate BAG, but continue to add bots to be approved on the WP:RFBA page. The community can then go there and, with consensus of the community as a whole, decide the usefulness of a bot. Those who are in BAG can simply comment on the validity of the code/script. If you want to complain this doesn't get enough traffic, then add it to the RfA main page. Monobi (talk) 00:07, 9 May 2008 (UTC)

As they seem unwilling to accept any new method of nomination (thus putting an end to the cabal concerns), I would agree with eliminating the BAG entirely and instead moving to an RFA style system for bot approvals (not to be held on the RFA page, but still to be decided by a bureaucrat (who would also of course set the bot flag upon community approval)). —Locke Cole • t • c 00:13, 9 May 2008 (UTC)
I proposed a cross transcluded discussion that will work. I just havent had time to work the details out. βcommand 2 00:48, 9 May 2008 (UTC)
Translusion won't work though; a discussion posted at WP:AN garnered zero responses, and AFAIK that's the only other place you could transclude them besides WP:RFA. (I suppose you could try WP:VPT, but I doubt it'll come close to the kind of feedback we were seeing with RfBAG, and that's the mark you need to overcome to convince me that what you're proposing is better). —Locke Cole • t • c 05:13, 9 May 2008 (UTC)
If you can get the community involved, more power to you. So far, community consensus on the bot approvals process has been an overwhelming "I don't care". --Carnildo (talk) 00:14, 9 May 2008 (UTC)
The arguments presented here don't seem to be very strong ones. The BAG membership process is shady? It's discussed at WP:BAG and WP:BOT, exactly where you'd expect to find it. Also, as others have said elsewhere, decisions are made by those who show up. In 2006, users discussed and then implemented the bot approvals group. If only a few people participated, oh well. Further issues can (and have) been brought up on the talk page and the policy is always changing (this is a wiki after all). There's no formal ratification or approval process (though Jimbo has suggested using ArbCom in the past (/me shudders)). Also, I think Carnildo's point is dead-on – nobody cares. --MZMcBride (talk) 01:21, 9 May 2008 (UTC)
Wikipedia:Adminship_poll/G is running 34-5 against having RfBAG on RfA. Although perhaps not the best announced (On signpost and 'crat noticeboard, I think, but didn't quite make the site notice), it's so highly lopsided that it suggests, just maybe, the community simply doesn't want to select BAG that way. Gimmetrow 01:34, 9 May 2008 (UTC)
If nobody cares, then I fully intend to run my bots without the approval of some poorly constructed group of users. Monobi (talk) 03:08, 9 May 2008 (UTC)
Then you will be blocked and further action will be taken against you. you have a history of poorly designed and coded bots that have had numerious complaints spreading the four main accounts that you have used. βcommand 03:16, 9 May 2008 (UTC)
Monobi, running your bots unapproved doesn't have the community consensus you've been pleading for in posts such as the one above. dihydrogen monoxide (H2O) 03:31, 9 May 2008 (UTC)
Monobi, WP did not originally have a bot policy. It developed because having completely unregulated bots created some problems. Of course nobody complains when a bot does good things and doesn't screw up, but it was the other cases that led the community to say that *someone* must approve bots. Anyone can comment on a bot request, but if nobody is responsible for the approval and has to answer for screwups, it's a form of tragedy of the commons. Thus BAG. BAG is pretty low in importance on WP, but it does serve a purpose. Gimmetrow 03:34, 9 May 2008 (UTC)
It's worth noting that most ways of determining community approval on Wikipedia don't involve closure by a person from a specific authorized group. Usually, if you want community approval for something, you either:
  • a) just do it, and see if anyone reverts you or tells you to stop, or
  • b) propose it somewhere, let people discuss it for a while and then, if looks like consensus has emerged in favor of it, do it.
Most such discussions will either continue until a consensus is reached naturally (or isn't, and the discussion just dies out), or they have a time limit but no particular designated group of closers. The exceptions, like AfD and RfA, tend to involve a simple yes/no poll to approve a single action (deletion, sysoping) that can only be carried out by a member of a particular group (admins, bureaucrats) and are thus naturally closed by the person actually performing (or declining to perform) said action when it is carried out (or not). BRfA is something of an anomaly here: even though it has a formal group of closers, it usually takes the form of a threaded discussion rather than a simple poll, it has no set time limit, and the people authorized to close it are not the ones with the actual authority to implement the decisions. In many ways, it would be much more natural if it was carried out more like, say, RfC. —Ilmari Karonen (talk) 05:46, 9 May 2008 (UTC)
BRFA is an RfC. Anyone can comment. It probably wouldn't make a difference if we said "BAG closers = admins + select approved non-admins", because probably only technically interested admins would get involved. Gimmetrow 00:02, 10 May 2008 (UTC)
That poll is nearly worthless. Besides, it's outweighed by the nearly fifty (or more?) contributors to a single RfBAG nomination. —Locke Cole • t • c 05:13, 9 May 2008 (UTC)
Just because I supported a RfBAG doesn't mean I support the process --Chris 09:09, 9 May 2008 (UTC)
Heck, just because I started a RfBAG for myself doesn't mean I necessarily support the process. (FWIW, I'm rather ambivalent on the subject, and not particularly convinced by the turnout of my own RfBAG: 24 !votes including several SPAs, as there would've been had it been closed on time, do not a very convincing consensus make.) I'm not sure what would be better, though; the best suggestion I've seen so far might be the one made by Betacommand above, involving multi-transcluded discussions (but please keep it out of ANI, thank you!), but I'm not really convinced even that would work all that much better. —Ilmari Karonen (talk) 09:53, 9 May 2008 (UTC)
I commented as well when it appeared the system was going to be pushed through despite the opposiiton, even though I am strongly against the system. As the only criteria for success was "people use it" it was set up so that after implementation it couldn't possibly fail unless people completely ignored it (obviously not going to happen on one of the highest watched pages) or the bureaucrats forced it to stop. The attitude of "other peoples' opinions are worthless because people used the process" was what caused me to abandon this discussion before. Mr.Z-man 17:57, 10 May 2008 (UTC)
I bet if put RfBAG nominations in the site notice, we would get lots of contributors. Should we put them in the site notice? Gimmetrow 06:00, 9 May 2008 (UTC)
Sure that sounds great, or ....

Whack!

You've been whacked with a wet trout.

Don't take this too seriously. Someone just wants to let you know that you did something silly.

— xaosflux Talk 11:57, 9 May 2008 (UTC)

You're really, really not helping the conversation at all. —Locke Cole • t • c 15:24, 9 May 2008 (UTC)
actually that was a very effective message. βcommand 15:25, 9 May 2008 (UTC)
You would think so. —Locke Cole • t • c 15:27, 9 May 2008 (UTC)
Pot, meet kettle. --uǝʌǝsʎʇɹoɟʇs(st47) 18:12, 9 May 2008 (UTC)
  • I wish people would not move comments around. My reply above about the "site notice" was to LCs comment about "fifty contributors", which comes across as: "we got more comments this way, therefore it's better". I was not actually suggesting putting it in the site notice; I was illustrating the problem with the argument by placing it in a different context. Gimmetrow 21:08, 9 May 2008 (UTC)
    • This is a pet peeve of mine, for exactly the same sorts of reason, so make this a month-late "+1". Alai (talk) 15:39, 6 June 2008 (UTC)

Not Eliminate reviewing bots

Even if the community feels that having a special group of bot approvers is not the best method, I'm not getting the feeling that the community thinks that bots should just roam free either. If it is decided to abandon BAG, the approvals process will need to be replaced by something...perhaps xfd style where any admin can close out the discussion? — xaosflux Talk 04:17, 9 May 2008 (UTC)

Why not just a noticeboard format, where people can comment as they please? The crats can decide if they think there are outstanding objections. I don't think the "support/oppose" format matches well with bot requests; it's usually more like: "Here's an issue" / "OK, fixed" / "Here's a question" / "Yes, that's fixed"/ etc. — Carl (CBM · talk) 15:30, 9 May 2008 (UTC)
We tried something like that before the invention of BAG. The silence was deafening. --Carnildo (talk) 19:31, 9 May 2008 (UTC)
Right, pre March 2006. This was before local bureaucrats even made bots... I think times have changed somewhat. Al Tally (talk) 19:46, 9 May 2008 (UTC)
Have you seen how highly active WP:BON is, even after I added it to the noticeboard header? MBisanz talk 19:48, 9 May 2008 (UTC)
What does that have to do with anything? Al Tally (talk) 19:51, 9 May 2008 (UTC)
That is in theory a noticeboard where people can comment as they please about bots. MBisanz talk 19:53, 9 May 2008 (UTC)
"This is not the place for requests for bot approvals". We are talking about the place where bots are approved. Al Tally (talk) 19:56, 9 May 2008 (UTC)
So you'd like this discussion sitting on a noticeboard page for a month? Along with the 30 or so other open Bot reqs? Do we now hate people without gigabit ethernet connections? MBisanz talk 20:04, 9 May 2008 (UTC)
I don't want a noticeboard. I want a "request for bot status" page, similar to XfD. I never mentioned a noticeboard. Oh and people comment on RFAs all the time, which are a lot bigger than that. Al Tally (talk) 20:57, 9 May 2008 (UTC)
Yes, set up an XfA system for bots. It works for the other processes where it is implemented, and it will work for bots as well. Every other project I have encountered uses an XfA system for bots, why does Wikipedia have to be different? Monobi (talk) 22:01, 9 May 2008 (UTC)
BRFA is an XfDiscussion system. Gimmetrow 22:28, 9 May 2008 (UTC)
But the bots are approved by a little Cabal. The community should be doing that, not a self appointed club. Al Tally (talk) 22:49, 9 May 2008 (UTC)
The community left the job of approving bots with BAG, and the community expresses its opinion on BAG members whenever anyone wants to join BAG. Gimmetrow 23:55, 9 May 2008 (UTC)
The community gave BAG that job quite a while ago, not expecting that BAG would grow into the shape of a giant middle finger pointing at the community. And it's not sufficient to be able to approve members (in a horribly obscure forum, if the BAG gets its way) before they join BAG -- what about disapproving existing members who act as if they own the bot policy and go on profanity-laden rants against anyone who disagrees? rspeer / ɹəədsɹ 08:03, 12 May 2008 (UTC)
There is precedent at Wikipedia_talk:Bots/Approvals_group/Archive_3#Betacommand for precisely that, and of course User conduct WP:RFCs in general for bad behavior. MBisanz talk 08:10, 12 May 2008 (UTC)
Then MFD WP:BAG, if its deleted, then the community shows it doesn't want BAG, if its kept, then its shown the community wants BAG. MBisanz talk 01:09, 10 May 2008 (UTC)
Been there, done that. The result was "speedy keep", and we have to discuss it here. Apparently BAG is exempt from being deleted. Al Tally (talk) 06:29, 10 May 2008 (UTC)
(outdent) If left to 'crats to close, is the suggestion that crats approve all bots, or just the 'bot flag', if just the flagging, having any bot do anything (without a flag) is still likely to be excessive. — xaosflux Talk 03:33, 10 May 2008 (UTC)
There are really two issues here — flagging of bot accounts and community review of bot tasks in general — and it might be best if they were actually kept separate, so that we'd have "Requests for bot status", which would be an RfA-style poll closed by those authorized to grant the bot flag (currently 'crats, though I'd personally support giving this ability to a broader group), and a separate "Bot review", in the style of RfC or even Editor review, for discussing proposed and ongoing bot tasks to see if they enjoy community support. Of course, the processes would still be connected to some extent, since nobody's likely to support flagging a bot account unless there's consensus for it to perform at least one task, but it might still simplify things (or not; the idea is to actually make the new processes simpler and less formal, not just take the existing bureaucracy and double it). —Ilmari Karonen (talk) 14:03, 10 May 2008 (UTC)
The solution is simple, really. Tag WP:BAG with {{historical}}, then add the BRFA page to show up on the RFA page. The intended task will be listed on the page, at which point, if consensus exists for the function, the operator will run the bot. After everything has been figured out and the majority of those who have expressed their opinions are please, a 'crat will the then flag the bot if need be, or approve it if it is to run without a flag. The guiding principle for the 'crats will be WP:BOT, instead of the arbitrary rulings of BAG. Monobi (talk) 03:59, 11 May 2008 (UTC)
Is such a complex procedure really needed for yet another interwiki or newsletter-delivery bot? --Carnildo (talk) 17:50, 11 May 2008 (UTC)
It's no less complex than BAG approving it; except this way, the community will be instead. Al Tally (talk) 17:54, 11 May 2008 (UTC)
If there is not a stable consistant group that over sees bots, we will run into some major problems. there are bots that come up, often and are denied as often, (IE WelcomeBot). you might have consensus for the bot from a group of editors, but the bot should still be denied. given the nature of bots you need a stable group to manage them. bot approvals SHOULD NEVER be a vote. bot approvals function as a discussion and safety check. removing the bot oversight is asking for trouble. βcommand 2 19:38, 11 May 2008 (UTC)
Why should it be denied? Because you don't like it? That is precisely what is wrong with all this. The bots that get approved are ones that you like, and not necessarily the community. Other wikis manage without an arbitary little group, and so can we. I have never even mentioned having a vote, instead I said a discussion. In other words, it would be just like it is now, but without BAG. I understand you and Carnildo would hate the idea of your little club being disbanded, and the community rightly deciding on these issues, but BAG is simply not needed anymore. And if we don't get consensus here, we'll have to MfD it, and you'll have to provide some sort of an argument to keep it, instead of shouting "Speedy keep!" Al Tally (talk) 19:44, 11 May 2008 (UTC)
My little club? Perhaps you should look at how active (or rather, inactive) I've been at approving bots, and then cease with the personal attacks. --Carnildo (talk) 00:09, 12 May 2008 (UTC)
I haven't made any personal attacks. It strikes me as odd you'd show up here to defend this group if you say you aren't even active here. You've still not said how having a little group approve instead of the community is better. Al Tally (talk) 00:18, 12 May 2008 (UTC)

I don't think that BAG "rulings" can be described as arbitrary by any measure. We're faster than a one-week RfA. We've rarely approved any bot that didn't have community consensus. We've rarely denied any bot that wouldn't have community consensus — very few bots are declined at all. — Werdna talk 07:29, 11 May 2008 (UTC)

Eliminate bot approval entirely

Given issues with BAG lately (notably their inability to comply with the will of the community) I support eliminating the BAG entirely. But further, I think we should stop approving bot tasks for the simple reason that we don't require editors to get approval for their edits before they make them, so we shouldn't be requiring bot operators to get approval for tasks prior to initiating them. All edits can be undone/reverted, so any potential damage is temporary at worst. I think that really just leaves the matter of how one applies for and receives a bot flag. I believe this is something we can do a number of ways, but the two most obvious ones to me are:

  1. Implement an RFA-style system for obtaining the bot flag (as with RfBAG, this would give us the widest exposure to the community, providing more input than obscure talk page discussions)
  2. Implement a system where the candidate contacts a bureaucrat directly (or via the noticeboard for 'crats), is given a trial run/grace period of a week (during which edits will be performed sans bot flag, and observed for potential issues), and assuming no problems are found, is given the bot flag directly

The first option provides more input from the community, and could actually include portions of the second option (for example; during the RfBotFlag, a trial run may be performed to demonstrate that the bot is working correctly). The latter option is less formal and would still allow community input (nothing stops concerned editors from watchlisting 'crat talk pages and the noticeboard).

As for bots who perform disputed operations, we should handle this like we handle any editor making disputed edits: discuss directly with the operator, get comments from the community via RFC, go to AN/I and, if the matter isn't solvable with the bot operator, then the matter can be taken back to the 'crats to discuss removing the bot flag (and thus that editors privilege to perform unattended automated edits). Thoughts? —Locke Cole • t • c 00:20, 12 May 2008 (UTC)

you need a reality check. stop attempting to push your POV. its getting disruptive. I would suggest that you avoid bot related topics because you seem to fail to understand what bots are, why we have a bot policy and why we have an approval system. RfX is broken like a square wheel. Im done attempting to talk with people who are attempting to re-write bot policy when they know absolutely jack shit about bots. if you ever figure out what the term "bot" means please come back, but until then stop spewing ideas that my dog even knows wont work. βcommand 00:49, 12 May 2008 (UTC)
Betacommand, please read this carefully before you reply again. The only thing that is "shit" around here is BAG and the members viciously defending it. Al Tally (talk) 01:28, 12 May 2008 (UTC)
yeah right, read this and grow a brain. your comments here show your lack of higher brain functions, you might want to see a doctor about that. Like I said you dont know what your talking about so shut up. βcommand 01:31, 12 May 2008 (UTC)
I'm not seeing how further discussion here, on eliminating BAG or BRFA will be productive. We're obviously deadlocked with some people wanting to keep things the way they are and others wanting to eliminate BAG and/or BRFA. Since we're producing a No Consensus which defaults to a Keep, might I suggest that those proposing the changes write them up as a full proposal and drop it on Template:CENT, RFC-Policy, etc and get more people involved? Or if that won't do, write up a detailed MFD summarizing the various points? MBisanz talk 01:55, 12 May 2008 (UTC)
Not true MBisanz. The only people who want BAG kept are in fact BAG members, so this isn't a "no consensus" case. Monobi (talk) 02:21, 12 May 2008 (UTC)
Since when did BAGers not get a say in whether or not to keep BAG or BRFA. Running through this thread, I did a back of the envelop count that 3 people think we need to eliminate one or both parts, 7 object to eliminating one or both parts, and 4 didn't give a thumbs up or down. That seems more like a Keep than even a No Consensus. MBisanz talk 02:28, 12 May 2008 (UTC)
It would be nice to see the BAG members themselves produce a coherent proposal for reform, since they are indeed best positioned to evaluate most of the issues. Ensuing calm discussion from all the BAG members would be even more appreciated. They, after all, are the most trusted and most expert as regards botops themselves. Let's hear a unified and dignified response. Franamax (talk) 02:42, 12 May 2008 (UTC)
MBisanz, BAG has no interest in enacting community will, otherwise they wouldn't exist. Beside, what group would willingly get rid of their power? None. I'd recommend we place this on a RfC page, since this hidden conversation isn't going anywhere. Monobi (talk) 03:12, 12 May 2008 (UTC)
I object to the implication that a discussion to modify policy taking place on the talk page of the policy page in question is "hidden". This discussion was announced elsewhere. The initial changes to the bot policy that prompted this whole thing were announced multiple places, multiple times. No one's shown up because no one cares, not because anyone is hiding anything.--Dycedarg ж 03:19, 12 May 2008 (UTC)
Well I can sort of see Monobi's point, most participants are existing BAG members who seem oblivious to the issues inherent with the BAG. Getting wider community input on this would be better, though I don't think an RFC will help (I've tried that before and it hasn't worked out so well). —Locke Cole • t • c 04:22, 12 May 2008 (UTC)
I disagree that we're deadlocked, I think most here agree we need to change something about BAG and/or bot approval. What we need are reasonable solutions, and that's exactly what I've presented. Please comment on that, as more input is always better than less. —Locke Cole • t • c 04:22, 12 May 2008 (UTC)
Again, going by the numbers, I see 3 people saying we need major changes, 7 saying we don't, and 4 having no clear position, doesn't feel like a "most". MBisanz talk 05:52, 12 May 2008 (UTC)
It'd be nice if you weren't hung up on numbers. Your numbers are directly above though, in a poll about how BAG members are chosen. If you'd like, we can conduct yet another poll on this specific issue, since you seem to think BAG has community approval... —Locke Cole • t • c 06:00, 12 May 2008 (UTC)
Where did everyone here get the idea that polls are the way to resolve everything? Polls don't create consensus, they destroy it. The BAG should be considering what's going on and finding a way to reform themselves before things get worse and they have to get reformed from outside. rspeer / ɹəədsɹ 08:10, 12 May 2008 (UTC)
Can someone please tell me why we need to get rid of the whole approval system? I understand that we have problems with bag being self elected but what is with this? An Rfa for a bot flag? You can't be serious --Chris 08:37, 12 May 2008 (UTC)
You don't get it. The approval system will stay the same, with regards to making a page with a bot request, and people comment on it. The only difference is that BAG will not be approving, instead anyone can, and once consensus is reached a bureaucrat flags. Al Tally (talk) 08:56, 12 May 2008 (UTC)
To be fair, the title of this section is "getting rid of bot approval entirely" :D Happy‑melon 10:22, 12 May 2008 (UTC)
Oh I totally agree with approving bots. Just not the idea that a little cabal gets to do it instead of the community. Al Tally (talk) 10:33, 12 May 2008 (UTC)
Other than that flagging is done by bureaucrats rather than stewards, how does your proposed system differ from what we did prior to BAG? --Carnildo (talk) 09:10, 12 May 2008 (UTC)
It doesn't. Al Tally (talk) 09:45, 12 May 2008 (UTC)
Which, mind you, isn't necessarily an argument against it. The original reason for creating the BAG was that stewards didn't want to spend the time and effort to decide whether or not local consensus existed in ambiguous cases or for proposals with few if any comments. (In hindsight, that problem might've been better solved by just relisting and more widely advertising any proposal without some quorum of commenters and a clear consensus one way or the other.) Now that local 'crats can flag bots, the situation is somewhat different, and it would be even more so if we were to give the ability to flag bots to an even wider group (like, say, all admins). —Ilmari Karonen (talk) 22:10, 12 May 2008 (UTC)

I don't know where you're getting the idea that the bot approvals group has a "complete inability to comply with the will of the community". Sauce? — Werdna talk 12:45, 12 May 2008 (UTC)

I haven't said that. I personally am not interested in how well or badly BAG do the job. The whole point of my argument is they shouldn't be doing this. The community should. BAG should not exist. It has no purpose, since the community can easily do what it does. Al Tally (talk) 14:00, 12 May 2008 (UTC)
The community does approve bots, by way of the group whose members the community approves. This is HOW the community has chosen to approve bots. And it's pretty easy to join BAG; hardly anyone who actually wants to do this job is excluded. Gimmetrow 14:13, 12 May 2008 (UTC)
Gimmetrow, this is NOT how the community chose to have a group approve bots. This "policy" was added by a user without discussion on any other forum or noticeboard, it was not created under community consensus. Also, just because people "vote" for BAG members doesn't mean they approve of the process. Monobi (talk) 20:10, 12 May 2008 (UTC)
The community also has a say in the approve process of bots simply by their participation in the discussion that is part of each bot approval. The simple fact is (which is true for almost everything in life) most people just do not want to get involved. I believe that bots need to be approved by a committee such as BAG. What might help is a formal appeal process to deal with special cases where BAG may not be reflecting the general consensus of the community as a whole. Dbiel (Talk) 20:08, 12 May 2008 (UTC)
So basically, absolutely nothing will change. The WP:RFBOT format will still be used, BAG will just be replaced by an unofficial group of people who end up closing 95% of the bot requests (after the novelty wears off after the first few weeks) because they are the only people who care enough to do it. So remind me, what is this supposed to solve/fix/improve? —Preceding unsigned comment added by Mr.Z-man (talk • contribs)
Actually it appears people are suggesting we don't have any bot approval, and we deal with bots only if they are problematic. Monobi (talk) 22:07, 12 May 2008 (UTC)
I removes the idea of an unofficial group that made themselves official without any consensus. They didn't need to exist in the first place, there was no consensus for them, so removing is going back to the status quo. Al Tally (talk) 21:45, 12 May 2008 (UTC)
Do you have any actual evidence there was not consensus for BAG, or are you just speculating? Gimmetrow 21:52, 12 May 2008 (UTC)
The status quo of mid-2006? Mr.Z-man 21:53, 12 May 2008 (UTC)
See this. That doesn't look like consensus of any kind.
If it was never legitimate in the first place, it is not legitimate now, especially as it isn't even needed. Al Tally (talk) 22:06, 12 May 2008 (UTC)
I don't see any lack of consensus there. Any other evidence? Gimmetrow 22:15, 12 May 2008 (UTC)
Well there was that first MFD when the community decided to Keep the BAG as the bot approving group. MBisanz talk 22:48, 12 May 2008 (UTC)
I didn't say a lack. 3 or 4 people saying "Yeah whatever" isn't a consensus. Al Tally (talk) 22:50, 12 May 2008 (UTC)
People were clearly uninformed that BAG was introduced without consensus, and without need. Al Tally (talk) 22:50, 12 May 2008 (UTC)
Wait, so now the argument is that the community wasn't properly informed before making a decision? Who are we to judge that the MFD voters didn't look in the history, understand it, and still vote to keep? Couldn't we blame 50 other things, good and bad, on the lack of the community understanding something anyways? MBisanz talk 22:57, 12 May 2008 (UTC)
The first MFD was Keep (reform), and as far as I know, it wasn't really reformed in any significant way. —Locke Cole • t • c 02:52, 13 May 2008 (UTC)

(undent) Ignoring for now the past (I really don't care if BAG started as a heavy metal garage band). Is the argument that what BAG does doesn't work or just that there isn't enough community input. If the former, there's been no evidence presented to that effect (in fact the evidence I've seen only speaks to the founding of BAG and the opinions of a couple current members). If it's the latter, then I ask Who is going to make the project suddenly care about bots? The only time anyone even notices a bot is when it does something they don't like, and even then the average user seems to have no idea what bot are and aren't capable of. Even if we "eliminate BAG," People will just go on ignoring bots, yelling whenever anything goes wrong and arguing about the past instead of debating the merits of the discussion. It will still be a couple of dedicated Bot Operators conducting the requests and still that same group nagging 'crats to close the discussions. Removing BAG will have no appreciable effect except to give someone fodder for the arguments that there "isn't a legitimate process" and that WP:IAR therefor allows them to run whatever bot they want to. BAG is doing a fine job approving bot and (ignoring the need for more help) deals with requests incredibly efficiently. I see no need to rock the boat just because the group developed out of need rather than explicit approval and consensus. Adam McCormick (talk) 00:36, 13 May 2008 (UTC)

Adam makes some very good points. BAG does not prohibit community involvement. Eliminating bot approval entirely is far more likely to create many more problems that it could possibly resolve. The only thing I see missing is a formal appeal process to deal with specific issues where BAG may be approving or not approving something contrary to community consensus. These would mostly be a very small percentage of the requests presented. Even without a formal appeal process in place, there are other means currently available to address any problem that might arise. Dbiel (Talk) 01:19, 13 May 2008 (UTC)
I've said it before and I'll say it again. BAG needs a component of members from the wider community, elected by the wider community. Two-approval, single-veto - one technical, one non-tech. Either side can deny approval or halt existing ops, both sides need at least one assenting member. We can't throw away the technical knowledge contributed from the existing BAGers - but we need a formal component specifically sensitive to community concerns, as opposed to just looking at whether the code works or not. Franamax (talk) 01:59, 13 May 2008 (UTC)
Having both technical and non-technical members on BAG would be a good thing, but setting up a single vote as a veto gives BAG members way too much power and is moving about as far away from consensus as it is possible to get. Remember that any admin can block a bot account that is acting outside of community consensus. Also BAG members already have an obligation to take all comments, both technical and non-technical, into consideration when making their decisions. Discussions are currently open to all. To get more community envolvement, the simpliest way would be to post more notices that specific discussions are underway, rather than changing where or how they are currently being handled. Dbiel (Talk) 02:28, 13 May 2008 (UTC)
I find it hard to believe that all of our technically oriented users are so completely out of touch with the community that this would be necessary. Mr.Z-man 02:35, 13 May 2008 (UTC)
Not all of them are out of touch, but many of them are (I'd provide a list, but I suspect it would be viewed as a personal attack). —Locke Cole • t • c 02:52, 13 May 2008 (UTC)
My argument is that BAG is simply unnecessary. We don't have Wikipedia:Editor Approval Group or Wikipedia:Editors/Requests for approval, nor should we have something like this for bots. Bots are basically just editors performing edits but in a rapid/automated fashion. As with normal editors, when we run in to problems with them, we can follow the presently existing methods of dispute resolution. The only thing we need to decide is how we give out bot flags, and I don't think we need BAG for that either. I guess Adam I just don't see the need (your words) for BAG, or that there ever was a need for it (the only historical note is further above where, apparently, Stewards were the only ones capable of giving out bot flags, but with bureaucrats able to do so, even that's no longer an issue). —Locke Cole • t • c 02:52, 13 May 2008 (UTC)
I encourage everyone to ignore all rules and run their bots without this arbitrary ruling. Monobi (talk) 02:57, 13 May 2008 (UTC)
I don't think anybody is willing to be blocked for doing this. I could be wrong though. - Rjd0060 (talk) 03:10, 13 May 2008 (UTC)
How does blocking an automated bot help the project? Or more specifically, how does an automated bot running without approval hurt the project? —Locke Cole • t • c 03:48, 13 May 2008 (UTC)
Following Locke Cole's Monobi's recommendation most users will find that their account would be blocked very quickly. Dbiel (Talk) 03:18, 13 May 2008 (UTC)
For doing what, might I ask? Not receiving some approval by a process created out of the will of a single person? Monobi (talk) 03:28, 13 May 2008 (UTC)
That isn't how I would describe it, but to keep things simple: Yes. - Rjd0060 (talk) 03:35, 13 May 2008 (UTC)
It's not my recommendation (though I personally see no problem if someone were to run an automated task without approval), please be careful in placing attribution next time. —Locke Cole • t • c 03:48, 13 May 2008 (UTC)
Thanks for pointing out my error, it looks like I got sloppy using copy and paste. My appoligies. (original post edited to reflect correct attribution) Dbiel (Talk) 03:56, 13 May 2008 (UTC)
The problem with running an automated task without approval relates mostly to the sheer number of edits it can make in a short period of time and the potential problem of reverting them (if shown to be necessary), dependant on the type of edit and the edit frequency of the articles involved. Remember that semi-automated tasks do not require approval. One of the most important functions of approving a bot is to determine if the edit itself falls within community consensus. Dbiel (Talk) 04:11, 13 May 2008 (UTC)
A couple of things:
  1. Again, we don't require normal editors to request permission to edit, so it follows we shouldn't be requiring bots (or rather, their operators) to request permission to edit either
  2. Bots need not necessarily have the bot flag, in which case there are limits in place to how fast edits can be performed (the bot flag relaxes these limits, but in theory without a BAG, many bots would likely be tried without a bot flag, limiting any potential "damage")
  3. As to potential damage/issues, as with any edit, bot edits can be reverted/fixed.
  4. Dispute resolution should handle just about any other problems. We have other policies (the blocking policy, etc) to handle bots which act against consensus.
Most of all though, and I'm reiterating this, I just don't buy in to the idea that a bot (especially an unflagged bot) will cause so much damage that it can't possibly be reverted/fixed. And these are all issues even with the approval process in place right now (many bots are closed source, so while a "test run" may not reveal any problems, it doesn't mean a full run of the bots operations won't). —Locke Cole • t • c 05:18, 13 May 2008 (UTC)
It's not a question of whether it's possible to revert the damage done. The point is that the amount of damage done will increase and it will take someone's time to fix it. I know you don't see the threat, But there is one and anyone who's ever had to clean up after a bot gone rogue will tell you that it's a pain which is mostly avoided with an established approvals process. Bots are not editors and they do need to be treated more harshly. It does not follow from the lack of editor approval that there shouldn't be bot approval. Even if that were the case, a bot is much more prolific than any editor (take BCBot for example) and reverting all it's work is a pain that is completely unnecessary. The cost of cleaning up after a profusion of unscrutinized bots is much higher than the cost of (even if we take it as you state it) your perceived notion that BAG is causing some imagined harm by just existing. Adam McCormick (talk) 05:36, 13 May 2008 (UTC)
(ec)One problem with that however, in the course of 'ordianry non-bot editing', a user can not cause the type and severity of damage that a broken / poorly designed (or thought out) bot can, regarding point one. If one of mine broke, just right (and, it can happen, easier than you would think), for instance (especially using the new version of distributed editing framework I've been working on), I could cause WEEKS of damage repair work, and very real damage to the project. And that's just me. Imagine having a few of those happening in short order. Human editors simply cannot cause problems on the same level, in the same way. Heck, we (whom are supposed to be fairly knowledgeable in this stuff) make mistakes too, see User:Pageview bot for instance. Imagine if things like that occurred more often, as a result of very non-technical people approving things that should not be. That being said, as I've said before, I would much prefer to return to a system of 'anyone can join', that we had for nearly a year. If folks are approving inappropriate bots, they can be removed, kinda like a clerking system like is in use at a few other places. I think, this would be a good balance. SQLQuery me! 05:38, 13 May 2008 (UTC)
I see your argument, but I don't think you see the need so let me attempt to illustrate. Let's take my own bot for example. Recently, the interlinking of categories led him to make a few hundred spurious edits before he was stopped. Would this ever happen to a a well-meaning editor? No, it can only really happen to an automated process following rules. If no bot ever needed approval events similar to this would be the norm of bot behavior. I've seen requests from people who only vaguely know how to program who want to edit every page on wikipedia. Bots that could wreak excessive havoc without any way of really constraining them. Without a specific group, without explicit approval what stops a particularly malicious vandal from making tens of thousands of edits before they can be stopped?
But even ignoring this extreme example, let's look at your "editor for approval process." On some levels I agree that bots are a reflection of their operators, but I can only really justify this for semi-automated bots. A fully automated bot operated unattended for days, or weeks in some cases, making thousands of edits without direct oversight. Can the same be said of an editor? If you post something on an editors page there is at least some expectation that they will read it, and react to it. The average member of the community can't do anything about a bot, can't really stop it from annihilating their work, have almost no recourse (Short of ANI). The only way to prevent this is direct oversight, direct supervision of editors who know bots. Not just programming them but operating them. I couldn't care less if BAG can program, but as i said those who care try to join, and those who care would become the de facto BAG even if the group was dissolved. Adam McCormick (talk) 05:11, 13 May 2008 (UTC)
I'm not advocating anarchy, of course, and this Bot policy page should codify requirements for those choosing to operate bots (specifically that secondary accounts be used, they identify themselves as bots on their user pages, etc). To take your example though, your bot made these "few hundred spurious edits" despite their being a BAG. Do you see my point there? Vandals who choose to operate a bot would be treated like any vandal we deal with: blocked and banned (plus I doubt any vandal would actually try to gain approval, but I think I get the idea).
Regular editors do have a recourse though, contacting the bot operator, using dispute resolution, and so on. Bot operators are just like editors, and like editors, there will be disputes about the kinds of edits they make. The significant difference is the speed at which they make edits, and that's really it. Everything can be undone if the community disagrees with their edits, and normal processes can be used to deal with difficult bots/operators (AN/I, temporary blocking, etc).
BTW, a dissolved BAG does not preclude there being a WikiProject or some such for bot operators to join and to self-police other bots as they deem necessary. But they certainly don't need the power of policy to note a misbehaving bot on AN/I or to contact a 'crat if a bot operator has lost the support of the community in operating their bot. —Locke Cole • t • c 05:27, 13 May 2008 (UTC)
I think you rather overstate the "power" BAG has. Anyone can note an allegedly misbehaving bot at AN/I. In almost the opposite of what you suggest, if a bot is brought up at AN/I, a BAG-approval somewhat reduces the chance it gets an immediate block. Gimmetrow 05:37, 13 May 2008 (UTC)

Civility

Since things have settled to some extent, no names need be mentioned, but any further issues with civility (including personal attacks, baiting or trolling) will very likely result in a block. Gimmetrow 02:07, 15 May 2008 (UTC)

I'm not entirely sure why people grow so hostile on this page, but it's more than a little annoying when some people randomly decide to blow up without actually contributing anything to the conversation (and in fact, seem intent on derailing discussion because they don't seem to like where it's going). FWIW, my opinion remains unchanged: BAG needs to go. —Locke Cole • t • c 08:18, 15 May 2008 (UTC)
LC, just so you're not surprised, when I said "no named need be mentioned", I was also referring to you. The way you present your arguments tends, unintentionally or not, to generate a defensive mindset. How do you think saying "their inability to comply with the will of the community" is likely to be received? Gimmetrow 08:37, 15 May 2008 (UTC)
About as bluntly as I intended it to be, I imagine. The simple fact is (see poll above if you have any doubts) there is a big disconnect between what the community believes and what BAG believes. And so far most BAGgers seem content to try and wish it all away (or ignore it, or resort to childish outbursts with much bluster but little in the way of logic). I'm not saying there hasn't been some reasonable discourse here, but despite the arguments presented being sound (and the support being there in some cases), members of BAG seem unwilling to budge or acknowledge they're in the minority.
FWIW, I'd tried to be gentler earlier on but my attempts to discuss this were usually ignored until I made an edit to the policy page directly. So hopefully you'll understand if I've slowly ran out of patience with this process. Maybe at some point there'll be some give from BAGs side? Compromise isn't out of the question from my end, but I haven't seen any that doesn't involve maintaining the status quo. —Locke Cole • t • c 09:00, 15 May 2008 (UTC)
It's rather difficult to find a compromise on "eliminate bot approval entirely", don't you think? What do you really want, and why? Gimmetrow 19:21, 15 May 2008 (UTC)
I want a transparent method of selecting BAG members that gets the most input from the community. We have a method, but for reasons that will forever elude me seemingly, members of BAG seem unwilling to accept it (I'm talking about the RFA-style system). Is it so unreasonable to get wide community input on something like this? —Locke Cole • t • c 00:53, 16 May 2008 (UTC)
What's not transparent about a discussion here? It would help if you would say clearly the problem you see, and how this solution would fix it. What specifically did the other system produce that you don't like? I know there are some systemic problems with the existing bot approval system, but most can be addressed by changing the bot policy. I don't really see how changing the method of BAG membership addresses these. Gimmetrow 01:47, 16 May 2008 (UTC)
I do not know I we are looking at the same thing or not. But any poll indicated above hardly reflects what the community thinks. There are just not enough members of the community responding to draw any conclusion, one way or the other, as to what the community actually wants; unless it might be to say that they just do not care; and even that is a bit of a far reach as the vast majority of the community does not even realize that this discussion is going on. Dbiel (Talk) 19:13, 15 May 2008 (UTC)
Last time I checked, when there's no consensus for a new system, we use whatever is in place. We don't just say "there's no consensus for anything" and then have nothing. If we did that everywhere, we would never have any stable policies and given some of the responses on the Adminship poll, we probably wouldn't have RFA either. I should also note that most of the people who opposed BAG membership on RFA were not BAG members. The RFA system was also spearheaded by Coren and supported (on the poll) by likely future BAG member User:Dihydrogen Monoxide - its quite clearly not as simple as BAG members vs. everyone else and the us vs. them mentality isn't helpful. Mr.Z-man 01:03, 16 May 2008 (UTC)
I'm sure this has been hashed and rehashed a million times already but I really have to disagree with this section. I think BAG has said many many many times that they wish for more community input on bot approval. Does the community ever provide any? No. So how does dissolving the only people who *do* provide input get the rest of the community to provide input? Q T C 04:19, 5 June 2008 (UTC)
Because we need to start over with something that works. BAG members ignore community consensus and revert war to keep their status quo going, ergo their use to the community is quickly approaching nothingness. There's various incidents, recent incidents, such as an early BRFA closure (with a resulting revert and protected page) that need to be considered here as well. How does prematurely closing a discussion and protecting the page benefit the community? How does a member of BAG abusing an undisclosed sockpuppet and using it in an edit war over BOT policy benefit the community? BAG is not good for Wikipedia, and IMHO should cease to operate. —Locke Cole • t • c 04:48, 5 June 2008 (UTC)
Distancing myself somewhat from LC's rhetoric, but not necessarily from his points, and with a heavy heart:
I must point out a recent page containing the quote "He is also a recently-appointed member of the Bots Approval Group, and has in many respects been offering a 'non-bot-operator' opinion that was so craved by the community", the same page containing a statement by the "He": "I stand by my BRFA approval, which was made from a technical perspective only. In that sense, I think it's been agreed that there isn't an issue. I should have asked for more community discussion prior to the approval".
I don't want to pick at any wounds, but I do need to point out the disconnect here. BAG must create it's own solution, and it should be satisfactory going forward. Please, guys, figure something out. You shot down RfBAG, you've dismissed my explicit two-component tech/non-tech solution, you haven't introduced answerability, there's additionally "admin bot whose owner refuses to post at ANI" (my quote and maybe bad grammar). Things are not going well here, any ideas? Franamax (talk) 05:02, 5 June 2008 (UTC)
Well, I am of the opinion that BAG is not merely a technical review, but also gauges community consensus. It's part of WP:BOT#Bot requirements, even. I'm making no statement about Fritzbot, but sometimes the "its only a technical review" is a sort of defensive response. And admins running scripts on their admin account are routinely bits of drama, but this case seems to have been handled reasonably well. It's not like any of these issues are not known, but it's a little difficult to address them sensibly while under constant attack. Gimmetrow 05:18, 5 June 2008 (UTC)
True enough, but where is the "constant attack" coming from? Is it those people who just totally hate computers, or is it from the (seemingly unending, maybe you'd agree) series of concerns being raised by reasonable members of the community about the way that bots and some members and non-members of BAG are going about their business? You can, of course, decide to put me on one side or the other. Whichever you decide, if you want to characterize it as constant attacks, it's not going to go away, right? That's why I'm asking you guys to start blue-skying some solutions. Franamax (talk) 05:30, 5 June 2008 (UTC)
I believe that the "gauging consensus" aspect is key, and is ultimately if anything the more important aspect. (That's not to say the 'technical' aspects are to be neglected, or that an explicit hammering-out-consensus on-page is going to be required in every case.) If there's no agreement on that -- and in the past there's been both resistance to it from within the BAG (from members that wish to scope the scope in purely technical terms), and from elsewhere in the community (from people who don't want the BAG to be able to 'nix' tasks) -- I believe we'll have to look again at some other mechanism for ensuring that consensus is somewhere in the loop. But for the sake of having as 'lightweight' a process as possible, I think it would be preferable to have both dealt with at once, and hence that the BAG consists of people who are trusted to make sensible decisions on technical grounds as and when those arise, and to judge consensus. Doesn't matter so much whether those are in separate heads, or combined in one, so long as there's a general recognition of what's going on. For my money, the "gauging consensus" and "being in touch with the community" aspects are sufficiently important that an RFA-like BAG-joining process is the least-worst option, though I won't claim it's in any way a perfect solution, or indeed the only acceptable one. Alai (talk) 10:26, 5 June 2008 (UTC)

When is a script a bot?

In WP:AN/I#Massive redirect deletions, an administrator apparently used a script in a very bot-like manner. Another administrator had to block his account for 1 minute to kill the script.

When does the use of an automated or semi-automated script cross the line and become subject to WP:BOT?

My recommendation:

  • Any fully-automated script is officially a bot.
  • Any user whose use of semi-automated scripts causes problems can be told by an any administrator for the next [insert short time period] ["your use of this script" or "your use of scripts to accomplish this task"] [plus optionally: "on this set of articles"] shall be subject to the requirements of a bot, except that a separate account is not needed.

Think of the latter as a partial block: The user may continue editing Wikipedia but his privileges are curtailed for a period of time.

What do you all think? davidwr/(talk)/(contribs)/(e-mail) 04:14, 1 June 2008 (UTC)

Not clear what you're getting at. If an admin could tell any user that script use is subject to bot requirements, but the user could keep editing, then effectively any admin could temporarily approve a bot. As it is, any fully-automated script is a bot and needs approval, and semi-automated scripts which edit "fast" or "a lot" are likely to need approval like a bot. Finally, a 1 minute block may or may not interrupt a script; the script may just keep retrying the edit. But if it works it can avoid some drama. As it turns out, the user had stopped running the script on his own. Gimmetrow 04:53, 1 June 2008 (UTC)
A script that edits quickly is just a bot by another name. What kind of meaningful distinction can you make between one computer program for editing Wikipedia and another? Calling your bot a "script" shouldn't make it exempt from approval. So I agree with Gimmetrow, and I disagree with davidwr's exemption for semi-automated "scripts" that handles them differently from semi-automated bots. rspeer / ɹəədsɹ 10:10, 1 June 2008 (UTC)
(I come here after reading the ANI thread) a script used to make 20 edits per minute for a long lenght of time to make thousands of edits should most certainly be treated as a bot because it has the potential to cause the exact same issues that bots cause, independently if whether there is a user appoving each edit or not.
Also, at high edit rates doing a repetitive task, there is no chance that the editor can actually evaluate every edit, not to mention the issues with keeping his attention level high and not start making mistakes, and the user is just imitating a bot's behaviour by hitting "edit" mechanically, so the edits are being as indiscriminate as those of a bot, and should not be treated as normal edits. --Enric Naval (talk) 15:47, 1 June 2008 (UTC)
Is that the distinction between a script and a bot -- whether there's manual approval or not? But then all semi-automated bots would be "scripts" as well. Also, there's no way to actually tell whether there's manual approval (consider the recent case where Betacommand claimed he manually approved every broken DEFAULTSORT edit he made, or MZMcBride's rapid deletions of redirects).
I think the only thing that could legitimately be called a "script" and not subject to the bot rules is a low-edit-rate tool that integrates with the user's editing to help them make edits (for example, TWINKLE). But if the "script" suggests the edit in full, and the user's only input is to click an "approve" button, it's a bot. rspeer / ɹəədsɹ 19:21, 1 June 2008 (UTC)
I'm not so sure I agree with you. I use reflinks.py on Pywikipedya as if it were a script, although I could use it in an "insert name, blindly click save" fashion if I wanted to. In this case, the distinction between bot-like use and script-like use is in the chair not the computer. As a matter of course, I typically preview the changes and, when it doesn't do the job right or completely, tweak the output before saving. Other editors may use the same tool in a "look at me, I'm a monkey, I can fix bare links fast" click-and-save way, which can get them into trouble since this particular script occasionally gives bad output. davidwr/(talk)/(contribs)/(e-mail) 22:03, 1 June 2008 (UTC)
My point is that we can't tell what's going on between the keyboard and the chair, and we shouldn't make rules that pretend we can tell. We should judge whether something is a script or a bot by (a) the edits it makes to Wikipedia, and (b) the way the code is designed to interact with the user (which in most cases we should know). What's going on in the person's brain shouldn't make the difference. rspeer / ɹəədsɹ 23:10, 1 June 2008 (UTC)
If we use that definition as a bot (no user input except setup and to approve each edit), AWB use would require bot approval, as would many user scripts, including some modules of Twinkle. Mr.Z-man 23:28, 1 June 2008 (UTC)
The definition of a bot has traditionally been full automation, specifically the lack of user input to approve each edit. AWB users who check and approve each edit are not running bots. AWB tasks which do a lot of edits (and so impact a lot of articles) may also warrant review. Gimmetrow 23:47, 1 June 2008 (UTC)
Aren't many AWB tasks run through the BAG? Aren't many BAG requests for semi-automated bots? By Gimmetrow's definition, there could be no such thing as a semi-automated bot, yet we see them frequently. rspeer / ɹəədsɹ 01:46, 2 June 2008 (UTC)
Semi-automated tasks are often run through the BAG for technical review, yes, often because they want to edit from a separate account ending in *bot. However, if they're not editing "fast" or not doing "a lot" of edits, they don't strictly-speaking need bot approval. I don't think this is a new interpretion of bot policy/practice. Before the maxlag feature was developed, someone could run a script for small jobs on their user account at two edits per minute, approving each edit individually, without strictly needing bot approval. Gimmetrow 01:56, 2 June 2008 (UTC)
(undent) OK, so exactly what is "fast" and what is "a lot" of edits? Will we always revolve around "I could do that many with a tabbed browser" as the standard response? Or the alternative "you're a f-'in liar I never used no bot prove it"? What's the answer here: when is a script a bot? Franamax (talk) 07:00, 3 June 2008 (UTC)
Well, that is the problem, isn't it. If I recall correctly, the policy used to be that anything more than 2 edits per minute for a sustained period was bot editing. Since maxlag was introduced, we don't have a specific edit rate for a target, but more than 5 or 6 edits per minute is too fast to really review anything that isn't trivial. "A lot" is also grey. The intention here is to let someone use AWB to do random odd tasks without needing further bot approval, but prevent someone making non-consensus changes on 10000 articles. Gimmetrow 07:13, 3 June 2008 (UTC)
Looking at the edit patterns of some editors, not all of whose names start with Beta or MZM, there are 10-minute pauses followed by a burst of high-rate edits. This is indicative of preparing a "job" within a script tool then setting it loose. I'm not convinced that qualifies as individual consideration of edits, as opposed to building an action list and setting it loose. The danger there is that the operator concentrates on the tasks, rather than the results; that is where the "bot" definition should provide some numerical clarity. The maxlag parameter is a bit of a red herring, it doesn't limit any page rates, you can go as fast as you want until the server tells you it's busy, then you back off a little bit.
Again though, there's a lot of back-and-forth over what's a bot, script, tabbed browser, human - can we quantify a per-minute and per-hour definition? Franamax (talk) 08:14, 3 June 2008 (UTC)
Well if you look at my contribs to images you'll see such a pattern. That pattern would be me opening 30 tabs in firefox, running the FURME script to get to an edit screen on each, then running back through the tabs and saving. MBisanz talk 08:18, 3 June 2008 (UTC)
I'm well aware of the ambiguity. I imagine it would take some work to settle down on specific definitions. I would venture these numbers: assisted editing over 6 edits per minute for a long period (long enough not to be tabbed browsing), or doing any one task involving 1000+ articles (even at a slow pace), should get bot approval. But that's just a view, it's not policy. Gimmetrow 08:24, 3 June 2008 (UTC)

The 1 minute block did nothing but add an entry to my block log. The block was made after the deletions had stopped. --MZMcBride (talk) 06:23, 3 June 2008 (UTC)

As I stated above. Gimmetrow 07:13, 3 June 2008 (UTC)

Essentially what we have at the moment is a giant duck test: if it quacks like a fully automated edit, it's subject to the policy. That's pretty much the best we can do, short of looking over someone's shoulder as they're making them (or as they have a pint while their script is doing so), though I'm certainly not opposed to tweaking wording to make current understanding and practice clearer. Defining it in terms of numbers of edits or edit rate per unit time isn't desirable, since any given figure will inconvenience those making legitimate "semis", act as "cover" for those pushing the envelope of unapproved fully-automated edits, or some combination of the two. But if someone were, say, making over three edits a minute solidly for an hour, I'd ask them if they were running a bot. If they said flat out that they were not, and offered some remotely-convincing explanation of what they were up to, I'd take it on faith that was the case, assuming there were no other issues. If someone's edits start to break all credibility of human stamina and attention span, or start to make "bot-like errors", I'd be somewhat more insistent on the matter. If that's the sort of thing that people "get" from reading whatever statement we make on semi-automated editing, I think that's good enough. Alai (talk) 16:42, 3 June 2008 (UTC)

Proposed bot change

The following discussion is archived. Please do not modify it. Subsequent comments should be made in a new section.
  • I'm going to close this poll now, as there is not presently consensus to implement this change at this time.

While it appears that this could be a useful suggestion to implement, there is no perceived benefit for the effort. (I.e. Bot names need the word bot in the name at present, and, non-bot names cannot have it, at present.) in doing all the work to tie a new namespace to a usergroup, redundant to a phrase already in nearly all these names. FWIW, there are presently 7217 account names containing the word 'Bot', and, 438 flagged bots. SQLQuery me! 18:04, 8 June 2008 (UTC)

Hello I would like to propose a chage. Instead of bots haveing a page starting in User: have there pages Start with Bot: Or Auto:.

If you want I can make a demo page for a bot. (If the operator of the Bot agrees.) S.T.H. ( P/T/C ) 20:32, 14 May 2008 (UTC)
Actually, in new Mediawiki versions it should be possible to achieve this quite easily. Similar to the Image: namespace becoming the File: namespace. MBisanz talk 23:18, 14 May 2008 (UTC)
Thankyou S.T.H. ( P/T/C ) 23:21, 14 May 2008 (UTC)
This is not as simple as renaming a namespace, the software needs to change all logs of edits made by bot-flagged accounts to point to Bot:Botname instead of User:Botname (without changing human editor edits). If this was like image to file, all users would appear to be bots. FunPika 00:02, 15 May 2008 (UTC)

Support

  1. S.T.H. ( P/T/C ) 20:32, 14 May 2008 (UTC)

Oppose

  1. Technically impossible without User: redirects I think. FunPika 22:47, 14 May 2008 (UTC)
    1. Its Very much possible. S.T.H. ( P/T/C ) 22:54, 14 May 2008 (UTC)
  2. Not worth the effort. Mr.Z-man 23:15, 14 May 2008 (UTC)
  3. Possible or not (wouldn't each bot still have a userpage?) it seems like unnecessary clutter. Adam McCormick (talk) 23:29, 14 May 2008 (UTC)
    The Benifet would leave more usernames open. S.T.H. ( P/T/C ) 00:03, 15 May 2008 (UTC)
    • The only way more user names would be available is if the character count limit of user names was being exceeded by add "BOT" to the name, which does not seem to be the case Dbiel (Talk) 00:37, 15 May 2008 (UTC)
    With current policy, all new bots are expected to have "bot" somewhere in the name; most have it at the end. Non-bot accounts should not contain "bot". This already distinguishes bot and non-bot accounts. Gimmetrow 00:14, 15 May 2008 (UTC)
    (e/c)No it wouldn't, not without doing this in pretty much the hardest way possible. The easy-ish way to do this would be to create Bot:/Bot talk: namespaces and simply move all the bot userpages into the new namespace. Redirects in userspace would still be needed because everything else in the software still would use "user:." The way you seem to be suggesting would require a new type of user, that isn't actually a user, and would have "botpages" instead of userpages. Besides all the new code that would have to be written to support this (it would affect contribs, deleted contribs, page history, and logs) a script would have to be run on the database to re-attribute all the edits and log events. And a new system would have to be set up to create "bots" separate from the normal user signup. Even then, we still probably wouldn't allow the re-use of the usernames as we don't currently allow non-bot usernames to contain "bot" and it would probably be pretty confusing to have a bot and a user with the same name. Mr.Z-man 00:15, 15 May 2008 (UTC)
  4. Not worth the effort to reclass the millions of bot edits that have already been made. I have not seen anyone claim that the current naming convention of bot user names is confusing. There is already a category Category:Wikipedia bots that groups all bots together. It just does not seem to serve any valid purpose Dbiel (Talk) 00:37, 15 May 2008 (UTC)
  5. Oppose, a lot of effort would be required, without signifigant benefit. — xaosflux Talk 02:17, 15 May 2008 (UTC)
  6. Total make-work. Happy‑melon 08:14, 16 May 2008 (UTC)
  7. To much work LegoKontribsTalkM 00:15, 23 May 2008 (UTC)
  8. Appreciate the thought of proposing this, as it makes sense in theory, but in practice: 1/ a lot of effort; 2/ very little practical gain; 3/ we've a lot more pressing matters to be going on with both in- and out-of bot related areas. Oppose at this time. Anthøny 19:22, 1 June 2008 (UTC)

Neutral

  1. Could you please explain to me the problem for which this is the solution? You mention leaving more usernames open, but as the vast majority of bots these days have "Bot" in the name, I don't see how this limits the average editor's choices. Are you saying that usernames ending in "Bot" are highly desired and (after this change) should be allowed? Or am I missing the actual problem? - AWeenieMan (talk) 00:07, 15 May 2008 (UTC)

2. I'm not sure. True, it may be better if they started with "Bot:", but the word "bot" is already in the name. It would be confusing if there were two of the same word, unless that were to be removed...? Chimchar monferno (talk) 16:39, 8 June 2008 (UTC)

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

Post-close observation

  • Actually, I make that a pretty clear consensus (to reject). Alai (talk) 19:36, 8 June 2008 (UTC)
    • LOL, same thing as no consensus to implement. I was trying to word it nicely. SQLQuery me! 21:36, 9 June 2008 (UTC)

Appropriate wording

From the bot requirements section:

  • uses informative messages, appropriately worded, in any edit summaries or messages left for users

Suppose that I want to create a bot which posts messages to talk pages (in this case, WikiProject talk pages, not actual user talk pages). Would the actual wording of the messages posted have to be approved? Or would it be sufficient to file an approval request for "posting messages to WikiProject talk pages, notifying them about XYZ", without specifying the actual wording? --B. Wolterding (talk) 12:07, 3 June 2008 (UTC)

Most likely the wording would be checked during the trial process yes. The wording could be improved later though if it was necessary. -- maelgwn - talk 12:32, 3 June 2008 (UTC)

No consensus

Per the above poll the method of choosing BAG members clearly lacks consensus. Despite my best efforts to educate the editors involved in reverting, they insist on pushing their preferred version on to this policy page. To that end, and since it's clear there's no consensus for the policy as it stands, I've placed {{disputedtag}} on the page. While I'd be more than happy to sort this out and would welcome some sort of agreement that would meet my single criteria (consensus), I'm not holding my breath. —Locke Cole • t • c 03:46, 5 June 2008 (UTC)

And yet, what is the process currently in use? Personally, I don't much care one way or the other how a group of minor administrative functionaries are selected on some website, and I doubt many others really do either. But this "proposal" was handled completely the wrong way, and that's why I'm skeptical. A lot of that wasn't your doing, but it's something you need to deal with if you want to continue. Supposedly, the change is to get more people involved or something like that. That begs a few questions: 1) Is it really beneficial to have more people involved selecting BAG members (as opposed to, say, commenting on BRFA), 2) why this particular "proposal" among at least a half-dozen that might achieve the same goal, 3) does it actually achieve the goal, and 4) are the problems associated with the change sufficiently offset by the benefits proposed or achieved? (Any change of rules has the inherent disadvantage of the adjustment period, so it has to have enough benefit to offset that or it's simply not worth it.) If you want the change to go through, you need to come up with something about each of those questions. Gimmetrow 04:59, 5 June 2008 (UTC)
To answer your questions, again, this is strictly my opinion: 1) Yes, I believe it is beneficial to have more input on BAG member selection, more discussion should always be encouraged, and additional input provides added potential for discussion. 2) Why not this proposal? In the test runs that were made it clearly demonstrated that it gave 3 to 4 times more input than the old method. To be clear, I'm not married to this particular method, but any alternative would need to at least double the amount of input from the community. 3) In test runs it tripled or quadrupled the amount of input compared to the old method, so I believe that's a "yes" in so far as achieving the goal. 4) I think the problems are minimal given the expanded input it provides and the additional opportunity for members for the community to be aware of BAG member selection. If I've not covered something, please let me know so I can try and address it. —Locke Cole • t • c 05:16, 5 June 2008 (UTC)
The problem is that input does not necessarily correlate to helpful comments. I really don't think number of participants should be sole the criteria we use to determine success. DHMO's BAG membership request was fairly nondescript, but his recent RFA was a massive dramafest. Would putting his BAG request on the RFA page have led to a similar result as his RFA? Its impossible to know, but its certainly a possibility. Is such a result desirable? I certainly hope not. Mr.Z-man 05:25, 5 June 2008 (UTC)
You say it was a dramafest, and I'm not judging (I haven't looked at the RFA), but were the concerns expressed by the opposing editors valid? If so, then the RFA served its purpose. As the bar is lower for BAG, I'd expect he wouldn't have had any problems using the proposed method (afterall, being included in BAG doesn't give one the same abilities as being given +sysop). I know more input isn't always better, but as a wiki we encourage discussions, yes? So more input should provide more opportunity for discussion (and certainly more opportunity for potential issues to be addressed). —Locke Cole • t • c 05:35, 5 June 2008 (UTC)
I'm glad you recognize more input isn't "always better". We have WP:Redirects for deletion, and they don't usually get a lot of comments, but we still have Redirects for deletion. Those interested comment. It's really the same here. I've been mostly satisfied with the group of approvers; it's BRFA that needs more input, in my opinion. Gimmetrow 05:54, 5 June 2008 (UTC)
The two are somewhat interlinked, however. If the group of "approvers" is small, there will often be less input/discussion. That would be a possible reason to expand the BAG, and perhaps specifically to expand it in a direction intended to take it wider than the "clique of bot-coders" it's been accused to being. Or else, it's an argument for making it a more "welcoming" place for non-BAG members to contribute, without the imminent prospect of them being shouted down as speaking out of turn process-wise, not being technically qualified to speak here, or just plain having splenetic venting directed at them. (I speak as having all of the above myself, and given that I'm a Wikipedian of several years standing, an admin, a bot-op with a "top 20 edits" bot, have done a little bit of bot-coding myself, and have a higher degree in CompSci, I think I can say that the "threshold" is not so much "high" as "totally whimsical".) Alai (talk) 11:32, 5 June 2008 (UTC)
(e/c) My point is that RFA is pretty well known for being one of the most politicized areas of the project. AGF seems to be suspended on a regular basis and the civility rules seem to be relaxed. I think I've asked something like this before and never gotten a straight answer: Why on earth would we want to expand that process? I just don't get it. RFA is one of the most controversial processes we have, and its the best solution we have for BAG membership? Mr.Z-man 06:01, 5 June 2008 (UTC)
Because it's the least-worst solution to that problem, and it seems likely to be least-worst solution to this one. Bear in mind that certain, erm, "somewhat controversial" Wikipedians have suggested a more "star chamber" model for RFA in the past (including running RFB on such a 'ticket'), and it's been shot down in flames). People object less to a "star chamber" model for BAG because fewer people care about it at all, and most of the remainder care about it less than they would RFA or RFB, probably due to a combination of it being seen as a niche activity, because there's there no actual technical permissions involved. Most of the time, RfBAG (or its present and past equivalent) exists in a quiet little corner where fewer people pay it any attention, so the BAG is able to tell itself that it's "less broken" than RFA. I believe that's drawing the wrong conclusion from the available evidence. Any time there's a Bot WikiDrama on AN/I (or elsewhere), that's implicitly a failure of the current system, including (but not limited to) the BAG selection process. Alai (talk) 11:40, 5 June 2008 (UTC)

I thought I was looking on the wrong page. Is this a WT:BOT thing or a WT:BAG thing? I think it would be useful to separate the technical aspects of bot policy from the BAG membership issues and the BAG membership selection issues. Which is why I tried to start a discussion at WT:BAG. Could someone centralise things and put notices in relevant places? Carcharoth (talk) 11:02, 5 June 2008 (UTC)

For now, BAG membership issues do fall under the bot policy (see here), and I don't think WP:BAG is policy etc. (at least, last time I checked it wasn't). So yeah, this should be discussed here. And I agree that we shouldn't be putting anyone through anything like my RfA, and thus think that the status quo—ask on WT:BAG if you can join the BAG—should be upheld. giggy (:O) 11:16, 5 June 2008 (UTC)
But reasoning from what it says "for now" is problematic, since what it says currently is precisely what's being edit-warred over! I do agree that it should be discussed here, until such time as the version here stabilises, but I think it's very bad practice to say "it's policy due to inertia, regardless of lack of current consensus", and it might be better in the medium-term to refer the membership question over to the BAG page, for the very reason that it's not tagged as policy. Alai (talk) 11:45, 5 June 2008 (UTC)
Oh not again with the inertia, I think that was the term that was used previously to discount all the opposition to the new system. Please don't go there. This isn't some new concept. When we don't have consensus for something, we use the status quo. We don't just abandon everything until we come up with something better or figure out a new term to discount the opposition. Then we're stuck using the first thing that isn't broken, regardless of how good it really is. Mr.Z-man 15:07, 5 June 2008 (UTC)
Can you suggest a better term than "inertia" for what you've just described? It being asserted (especially by revert war) that the policy remains the policy, when there's manifestly no consensus for it, seems to me to be precisely that. Perhaps we could call it "present policy being primarily conditioned by past policy, and at most secondarily conditioned by consensus", if I have my Buddhist terminology right. And it's especially unedifying when the main reverter-to-the-status-quo is precisely the person whose actions have made the status quo untenable. Alai (talk) 15:43, 5 June 2008 (UTC)
Well, I'm sorry I brought it up. Now, can we actually discuss a solution? The discussions on WT:BAG may not have consensus, but if you don't simply discount all the opposition, neither does the RFA system. So, what should we do? Mr.Z-man 15:56, 5 June 2008 (UTC)
I thought I had already sketched one. Refactor and/or mark the current process as provisional/proposed/disputed. Move detailed "status quo", proposals and counter-proposals to the BAG page, where instability is less disruptive, since it doesn't run interference on (the rest of) bot policy, and is all already marked as proposed in any case. Continue to operate whichever version there's the least drama over in the meantime. I agree that there's not (yet?) consensus for the RFA version either, so I would equally opposed to emplacing that as putative policy (if not more so, despite my preference for it). Alai (talk) 16:25, 5 June 2008 (UTC)

BAG request: Bjweeks (BJ)

My request to join the BAG is here. BJTalk 06:57, 6 June 2008 (UTC)

Approved or unapproved

How does one go about finding whether a bot is approved or not? I stumbled across an IP (91.198.174.201 (talk · contribs)) claiming bot status in edit summaries (with some slightly dodgy edits). --Rushmore cadet (talk) 12:24, 27 June 2008 (UTC)

A flagged bot is represented by a "b" alongside its edits in recentchanges/watchlists. IPs can't be bots, and this one is faking it for reasons unknown to me (since I didn't look at its contribs, just responding based on what you've said). If it's being disruptive you can request administrator attention at the relevant incidents noticeboard. giggy (:O) 12:52, 27 June 2008 (UTC)

That looks like it is a m:toolserver IP address and someone is logged out by accident. Not idea which account it 'should' be though. If you are going to block it TURN OFF AUTOBLOCK. -- maelgwn - talk 12:59, 27 June 2008 (UTC) It is already blocked by the look of it, see User_talk:Soxred93 and [[1]] -- maelgwn - talk 13:03, 27 June 2008 (UTC)

It seems to be all sorted now. Thanks. --Rushmore cadet (talk) 13:14, 27 June 2008 (UTC)

Category traversing for tagging

I suggest to add the following restriction to Wikipedia:Bot_policy#Restrictions_on_specific_tasks:

Bots should never operate on a list of categories and/or traverse the categorization tree to identify target articles for tagging and related tasks. Because Wikipedia categorization is not hierarchical, but instead a web-like structure, such an approach would be doomed to failure. Bots should only operate on sighted article lists for tagging and related tasks. Such lists can be generated from categories or from keyword searches. User:AlexNewArtBot is creating such lists of new articles based on keywords.

Please see User_talk:TinucherianBot, Wikipedia_talk:Bots/Requests_for_approval#TinucherianBot, Wikipedia_talk:WikiProject_Food_and_drink, and Wikipedia_talk:WikiProject_Computing#descendant_tagging for the relevant discussion on recent problematic bot runs. The operator and assumed a hierarchical categorization and by traversing the category tree a large number of completely unrelated articles got tagged for Wikiprojects. Cacycle (talk) 04:55, 5 July 2008 (UTC)

Uh, the consensus from what I had understood it was categories are fine but recursive categories were not. BJTalk 05:01, 5 July 2008 (UTC)
I would support this, the current state of Wikipedia's category system means that running such a bot on categories without checking all the pages in them is bound to be error-prone. In reply to Cwii, this isn't a vote, care to provide a reason? Mr.Z-man 05:10, 5 July 2008 (UTC)
This isn't a vote? Damn. Well my bots rely on categories, if I can't use categories I can't run bots since I have no toolserver account. There's no need for such a rule. CWii(Talk|Contribs) 05:14, 5 July 2008 (UTC)
This is only aimed at tagging and related tasks. As the current problems demonstrate, it is important to manually check the list of articles. Categories and sub-categories will inevitably have articles in them that are not in scope of a certain tagging task. CWii: How would you formulate it in order to prevent bots tagging amok? Cacycle (talk) 05:16, 5 July 2008 (UTC)
That's easy!
  1. Don't do anything stupid
  2. Tell the requester of tagging to double check the list
  3. Look through it yourself.
  4. Run the bot.
  5. If anything happens blame it on requester and mass rollback.
  6. rinse and repeat.
Hey, it's late. CWii(Talk|Contribs) 05:20, 5 July 2008 (UTC)
The problem is that simply reviewing the list of categories is not sufficient, as recent incidents show. Some categories may contain some pages that should be tagged and others that shouldn't be. In reply to your earlier comment, this doesn't apply to all bots, only wikiproject tagging and related bots. Forbidding all bots from using categories wouldn't make any sense. Mr.Z-man 05:29, 5 July 2008 (UTC)
Wikiproject bots? Eh, Why would I care. CWii(Talk|Contribs) 05:30, 5 July 2008 (UTC)
If you use AWB, why would you need a toolserver account to generate an article list from categories? Mr.Z-man 05:36, 5 July 2008 (UTC)
Nooo.... I use python now. But I would use it to run queries for more advanced bots. CWii(Talk|Contribs) 05:37, 5 July 2008 (UTC)
Oppose Catagories can be very helpfull when tagging articles, theres no need to come outright and say that no catagories can be used when tagging articles. All that is needed is care, and that was not present in the case that resulted in this discussion . I believe that a policy to ensure that care is taken may be appropriate but this sort of policy is not necessary. Printer222 (talk) 05:41, 5 July 2008 (UTC)
Nobody disputes that categories are helpful, but feeding bots categories instead of sighted ↓lists of articles for tagging does not work on Wikipedia for the obvious reasons stated above. This is merely meant to prevent bot operators new to Wikipedia to waste their (and our) time on a flawed approach. Cacycle (talk) 06:10, 5 July 2008 (UTC)

(outdent) That isn't what the policy is for, that should be under "bots frequently denied" or slap the BAG members so only more experienced operators get approved to run them. BJTalk 06:14, 5 July 2008 (UTC)

Feeding bots catagories that have been checked and are specific to the task can be helpfull. For example when i was tagging articles with the WikiProject Swimming tag, using the Category:Swimming did work. And yes i did check each sub cat. Food is so broad and then feeding the bot catagories because it is so board, doesn't work. You cant just generalise that feeding bot's catagories doesnt work, just because of one case where it didn't work. Printer222 (talk) 06:51, 5 July 2008 (UTC)

Aww, damnit, didn't see this discussion when I made this revert. Anyway, Cacycle is correct. Categories are useful, but they aren't perfect, and every list should be double checked before running a bot through it, not after. You check anyway, right? —Giggy 06:55, 5 July 2008 (UTC)

The whole point of an automated script is so that a human doesn't have to check every single article by hand. This is especially true when doing WikiProject tagging, as in the majority of cases the bot operator will not be familiar with the subject matter and so wouldn't be able to tell, short of opening and reading every article, which ones were within a project's scope. Recursive category parsing is a serious problem, and I think at WP:BOTREQ we've now reached the stage where no one will handle a recursive tagging request (if you want to make a point to the poster, just look for Category:World War II - it's a subcat of just about every top-level category I've looked at!), but expecting the bot operator to be responsible for eliminating false positive articles on a list of potentially tens of thousands of targets defeats the whole purpose of having a bot in the first place. Happy‑melon 10:03, 5 July 2008 (UTC)
Melon: Sorry for the confusion, I meant "sighted lists of articles", not "sighted articles" and have changed my comment above accordingly. Cacycle (talk) 17:06, 5 July 2008 (UTC)
if done properly the error rate with tagging is very low. Ive been running tagging bots for years, you just have to use your brain and be careful. NEVER automatically dig into subcats, but rather generate a list of sub cats and have the wikiproject go over the list. having more than one person related to the project that reviews the categories is a good thing. yeah there will be errors, but it would be the same errors that would happen with reviewing of individual pages in the list. βcommand 14:23, 5 July 2008 (UTC)
Of course it is OK to let a bot run on sighted categories. And there are actually some rare examples where you can use selected categories for blind tagging (such as the main category of a project), but we must make clear that this cannot be a general approach. Feel free to change the text above accordingly to make this clearer. Cacycle (talk) 16:33, 5 July 2008 (UTC)

Depending on the task, recursively going through all the sub categories sometimes makes perfect sense. Not always, but to outright ban the practice makes no sense --T-rex 15:08, 5 July 2008 (UTC)

Please could you describe an example where running a bot on subcategories would be adequate for article tagging. I have thought about this and could not come up with one. Thanks, Cacycle (talk) 16:33, 5 July 2008 (UTC)
In my experience recursive categories are very useful for generating a list that WikiProjects can filter. Once, and only once, atleast two users and myself have gone through the lists will I do a run. Yes it ends up being time consuming; but the current category system is fubar and I believe with the change to the bot policy regarding this matter something should be brought up at the village pump about the category system. §hep • ¡Talk to me! 18:06, 5 July 2008 (UTC)

Regarding the Wikiproject tagging: Articles are assigned to a Wikiproject because that project is supposed to work on them. If so many articles are added that the project can't even go through the list of titles due to lack of capacity, then something is terribly flawed I think. So: generating a list of potentially project-related articles (using categories), posting that list to the Wikiproject for review, have project members remove the false positives, and then parsing the reviewed list for automated tagging, seems a reasonable approach to me. By the experience mentioned above, the manual step should not be skipped. Whether this needs to be codified in policy is a different matter; but I don't object at that point. --B. Wolterding (talk) 19:39, 6 July 2008 (UTC)

Request for BAG membership

Per the bot policy, I am making this post to inform the community of my request for BAG membership. Please feel free to ask any questions/comment there. RichardΩ612 Ɣ ɸ 16:09, 5 July 2008 (UTC)

What is the threshold for manual edits being covered under bot policy?

If someone was creating articles at the average rate of 3 per minute, for periods of nearly an hour at a time, but they claimed to be doing it completely by hand, would their edits still be covered under the Bot policy? For example, 126 stub articles in 47 minutes. Kaldari (talk) 23:58, 22 July 2008 (UTC)

No. BJTalk 00:27, 23 July 2008 (UTC)
OK, if there is no threshold, I will no longer be requesting BOT approval for any edits that I perform, as I assure you I am going to do them all "manually". Feel free to delete my BOT account. Kaldari (talk) 00:49, 23 July 2008 (UTC)
… the threshold is automation. BJTalk 00:57, 23 July 2008 (UTC)
So anyone that wants to can circumvent this policy by simply stating that their edits are not automated, no matter how many edits per minute they are making? Kaldari (talk) 01:20, 23 July 2008 (UTC)
Sir Blofeld is an exceptional case - he really should have a bot flag on his account. But for whatever reason his editing activity is well known and he does not have a bot flag. I would classify his edits as informally approved.
In general, the people who notice "unapproved bots" are the recentchanges patrollers. If one of them notices you making a large number of seemingly automated edits, and you aren't on their mental list of exceptional users, they may leave you a note to find out what's going on. — Carl (CBM · talk) 01:32, 23 July 2008 (UTC)
Blofeld requested a bot account to create articles and it was denied when it was determined he would still be individually creating each article and it would not be automated in any way. MBisanz talk 01:40, 23 July 2008 (UTC)
If that is true, doesn't Wikipedia need some way to deal with regulating prolific manual editing as well? In case no one has noticed, Sir Blofeld has created 30,000 of the shortest stubs on Wikipedia, many without any sources or references, and some with questionable sources (like maplandia). I believe Wikipedia would be better off without these stubs until enough information is available on the internet to create decent articles, whether by bot or manually. But apparently the rest of Wikipedia has no problem letting him run rampant with no oversight whatsoever. I don't see how this is improving Wikipedia. Kaldari (talk) 03:17, 23 July 2008 (UTC)
If someone (such as myself) objects to the mass editing habits of someone who is "informally approved", what then is the appropriate channel for raising that objection? I already tried raising the objection on their talk page, but everyone there insists that there is no policy or guideline against what they are doing. Am I left then to manually review this user's edits as fast as they make them, nominating every article I can't verify for deletion? That would obviously be impossible. Do I have any alternative other than just completely ignoring the fact that Wikipedia is being flooded with thousands of unverifiable stubs (a large percentage of which seem to be based on single entries in GEOnames or maplandia)? Kaldari (talk) 03:23, 23 July 2008 (UTC)
WP:VPM is one place you could start a discussion if you disagree with the articles being created. If your concern is with sourcing, you could ask on WT:V. If your concern is with notability, WT:N. Have you tried raising your concerns directly with Blofeld? — Carl (CBM · talk) 03:48, 23 July 2008 (UTC)
Please see Wikipedia:Village pump (proposals)/FritzpollBot for some discussion about article creation. Making regular edits at over 5/minute for an extended time is likely to be considered bot-like, but I don't think we've ever discussed if semi-automated article creation at a somewhat lower rate would be enough to be bot-like. But you could always send the stubs to articles for deletion, perhaps in groups, or possibly to speedy deletion in some cases. Gimmetrow 03:57, 23 July 2008 (UTC)
Well, please don't mass-tag them. You can make a list and use a single AFD for the entire list. I don't believe there are any CSD criteria that apply to stubs of geographic locations. — Carl (CBM · talk) 04:02, 23 July 2008 (UTC)
I would hardly call working for 45 minutes at a time "bot-like." If he was working for several hours at a time, then maybe. But even semi-automatic scripts aren't even covered under the bot policy. As long as he's manually approving each edit, there's no real problem, at least as far as compliance with the bot policy goes. I fail to see how "no oversight whatsoever" is a huge problem - besides the usual recentchanges and newpages patrol, there's as much oversight as there is for any other user. Bots generally have even less oversight after approval, so I'm not sure how getting a bot flag would help with that. If the rest of Wikipedia has no problem with this, that's generally called a consensus Mr.Z-man 18:24, 26 July 2008 (UTC)

English Wikipedia opt into the global bot policy for interwiki bots

On Meta there is now a global bot policy for interwiki bots. The new policy will allow interwiki bots to have a global bot flag on wikis that allow it. I propose we opt into the system, while I don't think it was created with en.wp in mind, it would still be a great benefit to us. The current process for approving interwiki bots is woefully lacking. Now the bots are approved with only a short sanity trial without regards if the interwiki links made are even correct. Allowing the bots to be globally approved, with input by native speakers and accountability across many wikis would be a large improvement over the rubber stamping system we have now. BJTalk 18:03, 26 July 2008 (UTC)

  • I'd support the idea, but it is important to clarify that these bots are to be used solely for interwiki links, and that the flag is granted by stewards. . - Rjd0060 (talk) 18:08, 26 July 2008 (UTC)
  • I would definitely support this. Interwiki bots that already operate in multiple projects are pretty much guaranteed approval here. This would just make it easier for the operators. Mr.Z-man 18:33, 26 July 2008 (UTC)
  • I would support the idea. Right now it seems as if there is a major language barrier between wikis allowing interwiki bot operators to get away with a lot more than they'd be able to normally. --Nn123645 (talk) 19:14, 26 July 2008 (UTC)
  • My only concern is that we don't have control over what Meta is to include later in their "Global bots" section. So as long as we clearly define what global bots are allowed to do locally (i.e., only interlanguage work), it sounds fine to me. --MZMcBride (talk) 19:52, 26 July 2008 (UTC)
  • Sounds fine to me too, with the caveat mentioned by MZMcBride above. —Ilmari Karonen (talk) 19:56, 26 July 2008 (UTC)
  • I'd agree on the merits of opt-ing in for interwiki bots only. Global bots found to be doing other work without local approval should be blocked with judicious use of the banhammer. Q T C 05:18, 27 July 2008 (UTC)
  • The addition of EnWP to the standard bot policy list was reverted. I support our addition per BJ. —Giggy 05:35, 27 July 2008 (UTC)
  • I support adding the clear local definition here and at our global rights policy: WP:GRU. MBisanz talk 02:10, 29 July 2008 (UTC)
  • Comment:Now the bots are approved with only a short sanity trial without regards if the interwiki links made are even correct.? Are you saying that's the current system, or what it will be? Because that's not how interwiki bots are approved here. So, does the global bot policy have any consequences for us other than default approval for a few interwiki bots? Global-approved interwiki bots could presumably still be blocked locally if necessary. Also, even under the global system, bots would still need to get individual approval on a few wikis first, so we'll still review some. Gimmetrow 02:21, 29 July 2008 (UTC)
    That is how they currently are approved. The operators often speak poor English and the BAG (unless I'm missing something here) can't read the language of the interwiki links to see if they are correct. There is almost never any discussion about the bot, they are quickly trialled and approved. BJTalk 02:44, 29 July 2008 (UTC)
    Well I guess I'm the ogre then. While I assume the linking algorithm is "correct" in established frameworks, I do check what wikis the bot/operator has sought or received approval and spot-check a couple links in a language I know. Gimmetrow 02:52, 29 July 2008 (UTC)
  • Comment: So may enwiki be added now to the list of opt-inned wikis? — VasilievV 2 13:06, 2 August 2008 (UTC)
    I'd say so. —Giggy 13:38, 2 August 2008 (UTC)
  • I'd support, but let's not make the mistake we always make and assume that "support on some bot talk page" == "support from community" when community has not been informed. Any objections to refactoring this as an evil poll and spamming some boards with it? Happy‑melon 13:55, 2 August 2008 (UTC)
    Go for it, also I've worked what I hope is the spirit of this discussion into our WP:GRU policy. MBisanz talk 20:13, 2 August 2008 (UTC)

Above comments refactored from a threaded discussion. Happy‑melon 20:53, 2 August 2008 (UTC)

  • This is after all where we go to make decisions about bots. Chillum 21:46, 2 August 2008 (UTC)
  • Support Of course. Chillum 21:04, 2 August 2008 (UTC)
  • Support Prodego talk 21:44, 2 August 2008 (UTC)
  • Support -- Yes, it would speed up approvals. Meta is the best place to handle interwikibots. =Nichalp «Talk»= 06:08, 3 August 2008 (UTC)

I have made the changes to the relevant Meta pages: [2][3] —Giggy 06:20, 3 August 2008 (UTC)

It looks like you want to opt in to global bots and otherwise not use the standard bot policy. That is possible, but the requirements for global bots are defined by that policy rather than en-Wikipedia policy. (Of course, you can block global bots that operate here against the local policy.) —{admin} Pathoschild 22:35:26, 03 August 2008 (UTC)
That sounds about right, I think. —Giggy 01:29, 4 August 2008 (UTC)

BAG membership nomination

Per the bot policy, I am making this post to inform the community of a request for BAG membership. Please feel free to ask any questions/comment there. SQLQuery me! 03:14, 28 July 2008 (UTC)

Request for corporate memory on common bot screwups

I have blocked Lightbot for exceeding its mandate and subsequently screwing up. It was converting raw units to {{convert}} calls e.g. "25 ft" -> "{{convert|25|ft}}". I first noticed the problem when it did so in a direct quote,[4] and shortly afterwards someone pointed out that it had twice made such a conversion to a book title.[5][6]

The reason I am bringing this here is because we seem to be going over the same ground again and again. It has long been established that bots that make textual changes to articles don't work, because it is impossible to distinguish between material that is amenable to improvement, and material that must be preserved exactly as is, such as quotes, book titles, etc. But we don't seem to be learning from our mistakes. It frustrates me that we have to keep going over the same stuff, and I'm sure it must frustrate the BAG even more. Don't we have some kind of corporate memory in this area? e.g. required reading for bot operators? Somewhere where we can stick a bloody great blinking text box that says "Bots must not make textual changes!"

Hesperian 00:21, 1 August 2008 (UTC)

The project namespace is supposed to be our coporate memory (duh), but it has been hijacked, and is currently running one of the biggest and longest running nomics in the history of mankind. :-P --Kim Bruning (talk) 00:30, 1 August 2008 (UTC)
I have to say, as an outsider to this discussion, but as someone with plenty of experience writing bots and web crawlers in non-Wikipedia contexts, I'm not very clear what is meant by "bots must not make textual changes". What I assume you mean is that a bot mustn't treat an article as an opaque string and (effectively, if not actually) apply a regex to it. A bot must be able to parse the logical structure of the article, and act on that. Presumably implied in this is a requirement that bots should only alter parts of a article that are sufficiently structured that the changes of a false positive are practically zero.
So, for example, if I have a bot that replaces º, the Spanish masculine ordinal indicator, with °, the degree symbol. When would it be reasonable to make this replacement? If an article contains the sentence "Longyearbyen's latitude is 78ºN" would it be appropriate for a bot to apply a regex such as s/\<([0-9]\+)º([NSEW])\>/\1°\2/g that would replace them? I can't at present imagine a situation where an incorrect conversion might arise, but there may well be such situations. But I can't see what "textual change" might mean if this is not an example of one. However, I think we have quite a lot of bots that make precisely this sort of change.
— ras52 (talk) 12:44, 1 August 2008 (UTC)
Copy-paste the text of a page (as served, not as it appears in the edit box) into a raw text editor, such as Notepad or vi. The result is the text, with no links, no typeface, no layout, no images. That is what I mean by The Text.
Every time a bot tries to make wholesale changes to The Text, it screws up. We've had bots that capitalised florida in specific epithets. We've had bots that "corrected" the spelling of the French word mariage. We've had bots that fixed spelling in direct quotes, and now we have a bot that puts unit conversions in book titles.
You've put forward the safest example you can find, but even in this case it is easy to construct a example where the bot would fail. What would your bot do to the following sentence?
As an example of a textual error that a bot could safely fix, Ras52 offered the sentence "Longyearbyen's latitude is 78ºN".
Hesperian 13:19, 1 August 2008 (UTC)
The bot would change it. An editor then notices the mistake, reverts the edit and notifies the botop who blacklists the article or changes the coding so that anything inside quotemarks are ignored. The bot then goes back to making 499 good edits for every 1 obscure mistake. A question: do the edits ClueBot make fit your definition of a textual change? ~ AmeIiorate U T C @ 14:04, 1 August 2008 (UTC)
ClueBot is a special case and doesn't mark its edits +b. But Lightbot was approved as manually assisted, so the operator should be looking at context. BJTalk 14:09, 1 August 2008 (UTC)
Where did this discussion go wrong? This thread was not intended to start a competition to see who can think of the most foolproof example of a bot-fixable error, nor a competition to see who can trip me up and make my definitions/terminology look silly.
My point is that people are repeatedly letting loose bots that change text erroneously, because the bot-owners fail to realise that some text, such as quotes and names, should not be "corrected". The fact that we are eternally going over the same ground, and making the same mistakes, instead of learning from them, is a problem. What can we do about it?
Hesperian 14:21, 1 August 2008 (UTC)
I'm sorry if you thought I was trying to trip you up and make your definitions/terminology look silly. But I'm still genuinely struggling to understand what you mean. Do you think that it is reasonable for a bot to change 78ºN to 78°N irrespective of where it is? I'm assuming that you don't. But if so, can you give some examples of edits that you think a bot could legitimately make? — ras52 (talk) 14:56, 1 August 2008 (UTC)

I have also noticed incorrect bot "corrections" within quotations in article text. I would suggest that anyone reviewing a bot request that involves text changes, makes sure that the bot author has taken quotations into account. There's no reason for this issue to keep coming up. Kaldari (talk) 17:15, 11 August 2008 (UTC)

I've proposed to the bot's owner that I will unblock it, provided it is restored to its original functionality; i.e. not adding {{convert}} to articles. —  Tivedshambo  (t/c) 17:28, 13 August 2008 (UTC)
Fine with me. Hesperian 23:39, 13 August 2008 (UTC)

Proposal

After discussion at the adminbot RFC, I decided this was the best way to go. It's pretty much what Krimpet proposed there, and it seems to have gotten a lot of agreement that it was a good plan. I think this is the best way to go from here. --Chet B. LongTalk/ARK 00:20, 17 August 2008 (UTC)

So, these bots don't need to go through RfA anymore? Shouldn't we make a guideline that bot accounts should be separated from admin accounts or something? It's a bit unclear. Majorly talk 00:49, 17 August 2008 (UTC)
It is already policy, the admins running bots from their accounts are blockable in theory. 00:54, 17 August 2008 (UTC)
Right so if I go and block one of them, everyone is happy and pats me on the back? Or is a thread started on an admin noticeboard, threats of desysopping, blocking, drama etc started? Majorly talk 01:35, 17 August 2008 (UTC)
The latter most likely, for the time being. If the proposal passes admins should start splitting out their bots. After some grace period the policy should start being enforced. BJTalk 01:57, 17 August 2008 (UTC)
Endorse this policy, as long as adminbot operators are granted a period of time of at least 2 weeks after the policy is no longer proposed to apply for BRFA. --uǝʌǝsʎʇɹoɟʇs(st47) 13:28, 17 August 2008 (UTC)
Yeah to me that sounds like a good time to let every one currently running one to separate. --Chet B. LongTalk/ARK 14:23, 17 August 2008 (UTC)
I've added the two-weeks to the policy. Can someone volunteer to contact these admins when this gets marked as policy? I may not be actively editing (I shouldn't even be here now).--chaser (away) - talk 17:33, 17 August 2008 (UTC)

I've rewritten the text for style and presentation (diff) but I haven't (I hope) changed the meaning. I fully support its adoption. Happy‑melon 14:40, 17 August 2008 (UTC)

I think the "transition rule" should explicitly state that adminbots predating this new policy may continue to run on their operator's account during their BRFA. Any objections if I add this? A complex and controversial BRFA can easily take several weeks, and it would be insane to turn off all our vandal-blocking bots for that time just to satisfy a bureaucratic requirement. Of course, if the BRFA eventually concludes that the bot should not be approved, then things are different, but hopefully this won't happen, at least not in all cases. (Otherwise we'll just rename the site Grawp-o-pedia and be done with it...) —Ilmari Karonen (talk) 18:07, 17 August 2008 (UTC)

Proposal seems fine for me. By the way, Wikipedia:Requests for comment/Adminbots is still listed on Wikipedia:Centralized discussion, you may want to change the link to here at some point. --B. Wolterding (talk) 22:37, 17 August 2008 (UTC)

As I mentioned on WP:BN (diff), I propose that there be two stages: Technical and idealistic. The BAG can approve the bot on the technical side, and the community can approve the idea of the bot (but not its code) on the idealistic side, which would essentially turn de facto practice into de jure procedure. Please see the diff for more specific information. Thanks, —Animum (talk) 02:25, 18 August 2008 (UTC)

How does that idea compare with this current proposal? How would you propose to implement it in real terms? What advantages would the resulting process have over this proposal? I'm not trying to bite, just to demonstrate that we need concrete, detailed, on-the-ground proposals at this point in time. Happy‑melon 09:15, 18 August 2008 (UTC)
Quite the opposite, as it's a perfectly reasonable set of questions to ask. However, I'm a bit confused as to what you want to know that isn't outlined in the diff. Could you clarify? (Apologies for being thick. :-P) The one question I can answer without clarification is the last: This would simply mandate that the technical side and the idea receive approval. De facto to de jure, as I said above. —Animum (talk) 23:57, 18 August 2008 (UTC)
I admit that I'm somewhat confused as well, about Animum's proposal. Anyway, let me try to put it this way. In the usual bot approval process, three aspects are investigated as far as I can see.
  1. Is the bot technically sound, and the bot operator experienced enough to deal with technical issues?
  2. Is the bot's task in line with community consensus?
  3. Can the bot's task reasonably be done automatically, or does it require human judgment?
Number 1 is clearly for the BAG to decide, that's what the group is for. Number 2 does, I think, not pose a problem for adminbots; for all our administrative procedures are codified in policy to considerable detail, and if the bot's actions are not 100% in line with this policy, it should clearly be rejected - no extra forum is needed for that. (Unless this forum wants to discuss policy as such, which is outside the scope of bot approval.) Number 3 is, as far as I understand, what Animum wants to deal with in a separate process. However, I think that this a problem that benefits from the judgment of editors who are experienced with bots; and so it goes to the BAG again. That doesn't mean that discussion is limited to the BAG - on the contrary, every editor is invited to comment. But to me, setting up a separate process seems extra bureaucracy; or a means of creating extra drama. --B. Wolterding (talk) 00:24, 19 August 2008 (UTC)
"Number 2 does, I think, not pose a problem for adminbots; for all our administrative procedures are codified in policy to considerable detail, and if the bot's actions are not 100% in line with this policy, it should clearly be rejected - no extra forum is needed for that." Exactly, all I want is to make the community's will as important as the BAG's approval, but it admittedly takes the community's consent into account considerably, so this is pretty much unneeded except if we want to write it into policy more poignantly in order to reflect practice. Also, I'm not trying to get #3 discussed, for it can be discussed elsewhere. Does this pretty much clear everything up? —Animum (talk) 02:47, 19 August 2008 (UTC)
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:Bot_policy/Archive_22&oldid=1145852115"