Wikipedia talk:Requests for adminship/TawkerbotTorA

Meta-comments

This is not a game. This is not some IRC funny-room. In that regard, I'd be particularly critical of the editors numbered 11, 13, 24, 28, 29, 30, 38, 39, 43 and 45. -Splash - tk 10:07, 5 October 2006 (UTC)[reply]

... nor is it a mudslinging match. — Werdna talk criticism 10:21, 5 October 2006 (UTC)[reply]

...it is, however, a forum in which we are positively encouraged to comment on each others comments. I'm sure you'd be among the first to order "discussion not !vote". -Splash - tk 10:59, 5 October 2006 (UTC)[reply]

Yes, discussion is better than voting. — Werdna talk criticism 11:30, 5 October 2006 (UTC)[reply]

Relevent supportinf numbers

Crossposted from Wikipedia:Requests_for_adminship/TawkerbotTorA#Discussion_.28for_expressing_views_without_numbering.29. Please reply there.

I've asked on the main page, but it's very busy, so I'm crossposting. I'd like to know how many blocks per month this bot will be expected to do, but accounts range from 250 to 1500. I'd like to know what these numbers are based upon. - brenneman ^{L} 04:20, 6 October 2006 (UTC)[reply]

End crossposted.

Is this a bot in search of a function?

I know that Tawker and Werdna have the best intentions, and they are much more knowledgable than I about bots and proxies. However...

Two points: 1. This RFA seems premature. The scope of the project is constantly changing. First there are 1000 proxies to block, then 600, then 250. Is this an automated bot or a script? How will it be controlled? How can it be shut down? Who will have the passwords? All of these issues are in constant flux in response to questions and comments. Why was the Bot approval process only started today after the RFA was filed. Werdna wasn't aware of the proxy list? It seems like the operating specs should be nailed down before the RFA, not during; same with the bot approval.

2. I generated a list of current tor proxies in about 10 minutes, formatted and ready for blocking. See User:Thatcher131/Torlist. I used the list at http://proxy.org/tor.shtml, which is not the list Wernda proposes to use, and maybe there is a reason the list I used is not appropriate, but it was better formatted for what I wanted to do. I simply copied it from the browser page, and manipulated it in Word and Excel to get it in the correct {{IPvandal}} layout. It would have been faster if I still had a copy of BBEdit lying around, but it still only took a few minutes. (Someone could probably write a simple script that would parse the IPs out of the list Werdna wants to use, if that list is significantly different.) I also trimmed the list to include only exit nodes. There are 359 on the current list.

If I can generate this list manually in 10 minutes, there doesn't seem to be too much need for a bot. A new list could be run once a week and a simple comparison script run to find new exit nodes not on previous lists, or even better, a script could take each weekly list and prune it of any already blocked IPs (assuming such is possible).

In short, I don't see this bot will be doing anything that can't be done with a couple of scripts and a little human help, so I see no reason to set aside the longstanding policy of not sysopping bots, even if the goal is worthy. Thatcher131 04:42, 6 October 2006 (UTC)[reply]

Adminbots in general

As this attempt to win community support for the running of bots with admin rights seems to fail again, I find it interesting to look at how little opposition admins who run bots from their own accounts get. The famous Curps vandalblocking and anti-WoW bot was apparently blocked only twice during its carreer (and after Curps left, plans to have an official bot with this function were shut down). I have also seen very little opposition to Cyde using a bot to delete empty categories as part of his CFD work. I am sure that other bots or scripts have been run from admin accounts from time to time, but all of these are without official approval. Whether the actual clicking of the block button on a bot-generated Tor proxy list is done by a human or a script running from a human's admin account or an adminbot running from a specific account seems to make no big difference to me in the result, perhaps just in where we place the blame if something goes wrong or in who we hold accountable. An admin using a script from his own account is certainly responsible for all the script does, but isn't a bot owner responsible in just the same way? At least the robotic actions don't clutter up the log if they are done from a separate account.

If we deny the running of openly declared bots with admin functions, doesn't this mean we invite admins who have repetitive tasks to do to just IAR and run the bot from their own account and hope nobody cares? Kusma (討論) 08:07, 7 October 2006 (UTC)[reply]

Well, take a look at Wikipedia:Requests for arbitration/Marudubshinki, where Maru is about to be de-sysopped for running an admin bot. In that case he originally ran an unapproved bot from a second account, and it received many complaints for making mistakes. Rather then get bot approval, he moved it to his main account and equipped it to do deletions. I suppose that looks like selective enforcement, although there is also an issue of self-unblocking in that case, and Maru basically ignored multiple complaints about the bot making errors.

It seems to me that running a bot that would parse the tor list, check to see if the IP was already blocked, tag the user talk page with {{tor}}, and call up the blocking screen, but require a human to actually click the block button, would be fine. More like running a super-powered script. You could probably do 5-10 blocks per minute that way. Is this how Cyde's bot works or does he run it unattended? De facto double standards are not generally not good, so a broader examination of admin bots is probably needed if there are more flying under the radar. Thatcher131 11:38, 7 October 2006 (UTC)[reply]

I don't know what his bot does; when he was asked about it he only said it was "standard CFD work". Kusma (討論) 08:42, 8 October 2006 (UTC)[reply]

If adminbots are outlawed only outlaws will run adminbots? Haukur 14:47, 7 October 2006 (UTC)[reply]

I warn against admining the bot right away

Tawker's account may have been hijacked. It started deleting and salting notable articles out of process and while that may look just like someone being really abusive, the account then said it was gone on a wikibreak with some vague excuse about thanksgiving (which is over a month away). So unless Tawker can be verified as still controlling the account, the bot may have been compromised, too. Anomo 12:40, 7 October 2006 (UTC)[reply]

You fail to realise Tawker lives in Canada. Thanksgiving in Canada is on Monday. WP:AGF. – Ch acor 14:49, 7 October 2006 (UTC)[reply]

Outside any serious evidence, I tend to DISAGREE that that account has been compromised. If you really think it has, please build the case for it at WP:AN/I, not on various discussion pages. — xaosflux ^Talk 16:02, 7 October 2006 (UTC)[reply]

Consensus for TawkerbotTorA

This discussion from Wikipedia_talk:Requests_for_adminship was relocated here. So, the bot currently has around 60% support. Given recent events, would anyone care to guess whether this is "consensus" or not? -- ALoan (Talk) 10:30, 7 October 2006 (UTC)[reply]

Technically, in determining consensus, the arguments themselves carry more weight than "votes". - Stephanie Daugherty (Triona) - Talk - Comment - 11:14, 7 October 2006 (UTC)[reply]

Goodness, and there we all were forgetting that. ALoan's question is sensible, however: Werdna indicates that Jimbo (or "Jimmy") has said this bot might be ok (no idea what he actually said, of course) so this would be a similar excuse to the Carnildo debacle for varying the normal operating of RfA at the last moment. Given that such variation changes the rationale of the participants it'd be nice to know well in advance if this is going to happen. Without wishing to repeat all the arguments from the Carnildo case, is is worth pointing out that the ArbCom has found (or is soon going to) that that RfA lacked consensus, so some clarity in advance in this case is warranted. -Splash - tk 14:05, 7 October 2006 (UTC)[reply]

Whether or not the current 75%-80% procedure for promoting is a sensible one (I myself like it), I think anybody would be having a hard case arguing that 4 out of 10 people opposing your nomination means that there is community consensus. That even when considering as suggested above that "arguments carry more than weight than votes" (this last statment is by itself problematic). Oleg Alexandrov (talk) 19:17, 7 October 2006 (UTC)[reply]

I personally think that the arguments put forward by (half of) the opposers are intelligent and significant enough to prevent the bot from operating right away. However, if we're given the sysop flag, we can hammer out the other details with the rest of the BAG. The rest of the arguments are, frankly, paranoia about bots with the sysop bit. None of them actually include any reason why bots should not have the bit. — Werdna talk criticism 14:12, 8 October 2006 (UTC)[reply]

This is backwards. You should hammer out the details with the BAG and then seek sysopping. There are many reasons given for why the bot should not have the bit; although it's easy to consider that they are not reasons when one finds them to be counter to what one would prefer. -Splash - tk 16:45, 8 October 2006 (UTC)[reply]

Some of those oppose votes really are mindbogglingly stupid. Is a bureaucrat expected to count not-a-votes from people who clearly haven't a goddamn clue what they're talking about or, in a few cases, literally don't make any sense? - David Gerard 15:52, 8 October 2006 (UTC)[reply]

David Gerard, I say yes and no to your question. No, obviously they should not consider poor reasoning. Yes, they need to address the underlying misgivings that give rise to the misguided concerns. I've come to realize that RFA, one of the few set rituals on Wikipedia, functions as a mechanism to reinforce established community norms. When the misgivings are not addressed splits in the community thinking occurs. Smart discussion during RFA that address the poorly thought out reasons is the best way to go, I think. FloNight 16:14, 8 October 2006 (UTC)[reply]

A good number of the support !votes are also based on reasoning that has, time and again, been pointed as flawed (principally that the bot can a)be stopped, b)'recalled', c)prevented from acquiring new functionality without a return to RfA etc ). -Splash - tk 16:45, 8 October 2006 (UTC)[reply]

And a good number of support votes have no explanation whatsoever or are cliches. There is no reason to expect that those votes are more informed than the oppose votes. I happen to believe that the bot would be a good idea, and any bureaucrat can remove its sysop bit at any time if need be (without any hurt feelings on bot's side, obviously :) However, picking and choosing which votes are "valid" is extremely biased and should not be allowed. For good or bad, the community has spoken and that decision needs to be followed. Oleg Alexandrov (talk) 16:58, 8 October 2006 (UTC)[reply]

Actually, a Steward would be required, however I've written an extension to MediaWiki that would allow bureaucrats to do so. — Werdna talk criticism 00:55, 9 October 2006 (UTC)[reply]

So a bureaucrat can give, but they can't take away? - jc37 00:59, 9 October 2006 (UTC)[reply]

Exactly. It's both a policy and a technical issue, on smaller wikis, particularlly non-english ones, a rogue 'crat could remove everyone's sysop bit. On larger ones, the process of desysopping is harder, and is quasi-judicial rather than political, and rarely performed, so the need for a 'crat to desysop anyone is rare. - Stephanie Daugherty (Triona) - Talk - Comment - 01:15, 9 October 2006 (UTC)[reply]

On one hand, I would think that a "rogue" would be undone by any other bureaucrat, but since removal is rare, I can see why it could be seen as not useful. - jc37 01:23, 9 October 2006 (UTC)[reply]