Wiki143:Requests for adminship/TawkerbotTorA

The following discussion is preserved as an archive of a request for adminship that did not succeed. Please do not modify it.

TawkerbotTorA

Final (131/69/10) Ended Thu, 12 Oct 2006 13:03:54 UTC

Just for public record - there is an method in the code for sysops to stop this bot from operating. You edit a checkpage just like you do w/ AWB —The preceding unsigned comment was added by Tawker (talk • contribs) .

Script error: No such module "user". - the Script that likes to be called a bot (not fully automated - manually run) – Ok, this is not a human RFA. Jimbo was talking to Werdna about an automatic tor blocking bot that blocks the tor proxies listed on http://tor.noreply.org/tor/. It's a pain in the ass for human editors to do it... hence a bot is a lot better. It's a pretty much idiot proof bot, the code's at Wikipedia:Requests for adminship/TawkerbotTorA/code. However, as you might assume, it requires +sysop to do it's deed hence this RfA is being made - Jimbo doesn't want to overide community consensus. Any questions post away. One important note... this is NOT Tawkerbot2 - the two bots are totally unique (Tawkerbot2 is python, TawkerbotTorA in c#) Tawker 00:01, 5 October 2006 (UTC)

The bot is developed by me, and operated by Tawker (who was mean enough to name it after himself). We will be in full consultation, and ideally will both have shutoff rights. Just to note, this blocks the Tor users anonymous-only, with account creation disabled. This combination was suggested by Jimmy. — Werdna talk criticism 00:04, 5 October 2006 (UTC)

Addendum — as I'm not a sysop, I will not have the password to the bot. — Werdna talk criticism 00:06, 5 October 2006 (UTC)

See now, Tawker should name the bot after himself. That is the convention to name a bot after the person operating it, not after the author of said bot. -- RM 12:22, 6 October 2006 (UTC)

General comments

It may also be prudent to read this and this comment from Essjay regarding an earlier similar (but not the same) proposal for one of Tawker's bots to have sysop capabilities. – Ch acor 13:23, 5 October 2006 (UTC)
It's also prudent to note that it is a 100% unique task. I'd have to say that blocking a la Curps style is a fair bit more contoversial than doing a very simple janatorial task -- Tawker 17:18, 5 October 2006 (UTC)
Bot's request for approval page at BRFA --WinHunter ^(talk) 13:58, 5 October 2006 (UTC)
Just implemented an emergency stop facility. — Werdna talk criticism 16:42, 5 October 2006 (UTC)
- Please change the emergency stop to be deterministic. With that code there's over a 1 in 4 chance that the bot will still be running after 12 blocks; try just checking every fifth block or something like that. With a sysop-bot, we need to be very certain that everything's as it should be. In fact, this may be better with a slow throttle (say 30 seconds), and a check every block... --ais523 16:52, 5 October 2006 (UTC)
  - Why not make it check that stop page after every block? True, it would add time for each block, but that time should be insignificant and minimal, and if the block's malfunctioning, I would rather have the block stop at once rather than relying on a random number generator hitting the right number. (Or if checking every block really isn't optimal, then perhaps, instead of using a random number generator, use a counter variable that increments each time a block is made, and then resetting the counter back to 1 once it reaches the limit (10, in this case)? This would ensure that the stop isn't up to chance, and that the bot would stop in no more than a certain number of blocks.) However, I would still strongly prefer a check after every block. Thanks! Flcelloguy (A note?) 17:02, 5 October 2006 (UTC)
    - Hmmm... sounds like a server-smashing thing. How about every three unblock/block combinations, deterministically (counter). — Werdna talk criticism 17:08, 5 October 2006 (UTC)
      - Sounds good, and I understand the concern for resource hogging. Flcelloguy (A note?) 17:14, 5 October 2006 (UTC)
        Since writes are the most intensive on the database server, and retrievals are not so intensive, I don't see why you couldn't do a check every time instead of every 3 times. These type of issues have been brought up before, and generally if they are reads of a page this is fine. Especially given the controversial nature of this bot, it *should* check the emergency stop EVERY TIME, as the "extra load" is worth it. -- RM (member of the BAG) 12:26, 6 October 2006 (UTC)
Note Please note that though I've vocally opposed this bot, it has nothing to do with not trusting the operator, Tawker does a great job IMHO. As a WP:BAG member, I think this needs more work before implementing. — xaosflux ^Talk 01:40, 6 October 2006 (UTC)

Discussion (for expressing views without numbering)

How fast will it run and how often?Geni 00:19, 5 October 2006 (UTC)

Manually run, I was thinking of once a week. -- Tawker 00:30, 5 October 2006 (UTC)

back when User:Fvw did something like this it had a tendacy to seriouly flood recent changes.Geni 11:24, 5 October 2006 (UTC)

Couple of questions:

Do any other bots have admin status?
User:Curps did / does. It wasn't a "bot" per say, Curps just ran it without any consensus / approval on a "human" account. -- Tawker 00:30, 5 October 2006 (UTC)

I once did the same thing with User:Ram-Man, running an unapproved bot with admin priveledges. Not that I'm defending such action, but for historical note I thought I'd mention it. I suppose the point is that any admin can run an unapproved bot. -- RM 12:28, 6 October 2006 (UTC)
Is it to be run from somewhere secure (since it appears that the password for an account with admin rights will be held as plain text, I would think your firewall is going to be fairly busy)? Yomangani ^talk 00:21, 5 October 2006 (UTC)

Well, many admin passwords are held in cookies of web browsers. As for my local security my PC is a dynamic IP, is behind two layers of firewalling and NAT. In short, it's a pretty secure box. -- Tawker 00:30, 5 October 2006 (UTC)

How will addresses be unblocked when they are no longer running tor? Demi ^T/_C 00:26, 5 October 2006 (UTC)

The blocks will be for a fixed amount of time, I was thinking of a month. After that if they are no longer an tor exit node the block automatically expires, if it's open another block for a month and so on and so forth. Of course, an bot that keeps a list and then dymaically updates it would work too, that's just not written yet -- Tawker 00:33, 5 October 2006 (UTC)

We have had problems with users and connection hijacking, local machine compromises, etc. Can this be run somewhere supported by Foundation ops staff? Demi ^T/_C 00:26, 5 October 2006 (UTC)
Tawker is already a sysop. — Werdna talk criticism 00:35, 5 October 2006 (UTC)
Not what I asked. Demi ^T/_C 00:52, 5 October 2006 (UTC)
Well, it's not going to change the odds of an account hijack if the bot was run externally or if it's run by ops staff. It's not exactly something that runs on the cluster.. It would go through the 'net at some point anyways. (This comment added by Tawker · Demi ^T/_C 00:52, 5 October 2006 (UTC))
No, I mean from the tools server or somewhere similar. Demi ^T/_C 00:52, 5 October 2006 (UTC)
I don't really see the necessity for this. Plenty of sysops edit through unsecure channels. Unless you're going to restrict all sysop editing to https (which isn't such a bad idea, IMO), there really isn't a need to enforce this. — Werdna talk criticism 01:25, 5 October 2006 (UTC)
Can you confirm that this will only block exit nodes, and whether there any plans to change this? Jim182 00:29, 5 October 2006 (UTC)

From the code:

if ( (AccessPolicy.Contains("accept *:*") || AccessPolicy.Contains("accept *:80") ) &&
                    !AccessPolicy.Contains("reject *:80"))

So in short, yes, it blocks only exit nodes. — Werdna talk criticism 00:35, 5 October 2006 (UTC)

Will this bot ever get any other/new functions, and if yes, will these functions use the admin-tools as well? --Conti|✉ 01:01, 5 October 2006 (UTC)
Any new functions, unless they are enhancements to the Tor-blocking functionality, will be run on separate accounts and approved separately. — Werdna talk criticism 01:05, 5 October 2006 (UTC)
Why would this bot account ever be UNblocking? — xaosflux ^Talk 12:28, 5 October 2006 (UTC)
It's currently coded to unblock each node before adding a block. This ensures that the blocks don't prematurely expire on us, and makes sure they're soft. — Werdna talk criticism 12:35, 5 October 2006 (UTC)
The permature expiry behaviour doesn't happen and hasn't in a while, the block request will be rejected if the account is already blocked. --pgk 14:22, 5 October 2006 (UTC)
Is this going to block anons & account creation only, or allow registered users to edit through Tor nodes (eg. to circumvent evil firewalls)? [ælfəks] 01:10, 5 October 2006 (UTC)
To quote above, "Just to note, this blocks the Tor users anonymous-only, with account creation disabled. This combination was suggested by Jimmy." — Werdna talk criticism 01:19, 5 October 2006 (UTC)

From what I've read, it sounds like a good idea. But I would like some clarification for the less technically-proficient of us (not saying that I'm one! :-) ): what exactly will this bot do, in layman's terms? What exactly are Tor proxies? Where is the list at [1] originated from? What do all those strange-looking character combinations mean? Thanks! Flcelloguy (A note?) 02:42, 5 October 2006 (UTC)
Tor is a network of computers that will use onion routing to obfuscate (hide) your location. They forward traffic for you to other places, to help mask your location. All Tor proxies are currently blocked indefinitely on sight, but our coverage is very haphazard, and we basically just block whatever we find. As I explained to Jimbo, the Tor problem, as [he put it], is that I can't edit through Tor without switching proxies a few times every ten minutes, and yet vandals can make their edits with the unblocked proxies on a drive-by basis. The purpose of this bot is to standardise the soft-blocking of these proxies by blocking each proxy for a month or so, once per week, with the anonymous-only and no account creation bits set. This will allow users with accounts to edit through Tor to circumvent fascist firewalls, or simply out of habit, however it will prohibit vandals from editing and creating accounts through Tor to circumvent blocks. I don't really want to go into the specific technical details of the Tor directory — those who are interested can find more information on the Tor website. — Werdna talk criticism 03:00, 5 October 2006 (UTC)

Some more questions, this time from myself... when you say "manually run", about how many blocks/minute will be placed? When the bot is done with the list at the URL, is that all it will do? Or will that page be updated with more IP addresses? (And have you tested this bot in another wiki before? Can you link to that test?) Thanks! Flcelloguy (A note?) 02:42, 5 October 2006 (UTC)
In my own tests, about a dozen block/unblock actions per minute took place. This was connecting to a slowish server, so you can expect more than this on Wikimedia servers. I would recommend in the strongest terms that this bot is also ~~sysop-flagged~~ botflagged, to prevent the blocks from disrupting Recent Changes. — Werdna talk criticism 03:00, 5 October 2006 (UTC)
Thanks for the answers! I think you might have missed two of my questions, though: once the bot is through blocking the addresses listed at the Tor URL, is the bot's job complete, or does the URL somehow "change" or "refresh"/"update" itself with more IP addresses? In other words, is this bot working on a one-time job - blocking all of the IP addresses listed - or on a continous job, blocking all of the IP addresses listed that are changing continually? The second question relates to the tests: has the bot been tested on a wiki similar to here? (i.e. MediaWiki) If so, would you mind providing a link to the test? Thanks again, and apologies for being nitpicky on a wonderful idea! Flcelloguy (A note?) 03:07, 5 October 2006 (UTC)
Sorry, I could have sworn I answered those. I think I /thought/ the answers. Anyway, yes, the URL is updated regularly with the hosts of current Tor servers. It's a continual job. It has been tested on my own wiki [2]. Please note that the last few hundred blocks were me removing the blocks it'd placed because I didn't need to block Tor at that time. — Werdna talk criticism 03:56, 5 October 2006 (UTC)

P.S. "Sysop-flagged"? Uhh... aren't all admin accounts sysop-flagged automatically? Or did you mean bot-flagged (which I presume would hide even log actions from recent changes?), or did I miss something? :-) Thanks! Flcelloguy (A note?) 03:10, 5 October 2006 (UTC)
Yes, botflagged. — Werdna talk criticism 03:56, 5 October 2006 (UTC)

This bot would totally lay off all other admin tools, yes? If it doesn't (which is purely speculative at this point), what should we do (besides obviously desysopping)? – Ch acor 03:55, 5 October 2006 (UTC)
Yep, personally all we need is the ability to block and unblock but thats not a permission we can set per account at the moment w/o modifying the codebase. Any new features if any need comes up would need consensus again. -- Tawker 04:03, 5 October 2006 (UTC)
Totally. I would suppose its approval would be taken back and it'd be desysopped straight away if something like that happened. I can't imagine under what circumstances it would, though. — Werdna talk criticism 04:06, 5 October 2006 (UTC)

Where will the log that the code shows at User:Werdna/Tor List be? I would feel more comfortable if a message in the talk page of the blocked user explains that the IP is blocked due Wikipedia policy. Maybe a template similar to Template:Tl. We can't expect everyone using Tor to know that. -- ReyBrujo 04:49, 5 October 2006 (UTC)

Yeah, we'll tweak the block messages to something like Template:Tl — Werdna talk criticism 06:44, 5 October 2006 (UTC)

What sort of precedent will this set? And why will it be blocking account creation as well? --Gwern (contribs) 03:00, 5 October 2006 (UTC)

The precedent set is that sysopped bots must be approved by a community-wide discussion, whether that be a few days on RfA, a full-term RfA, or some discussion in another venue. I personally recommend a few days on RfA, and if there's a snowball's chance in hell that there'll be no consensus, we run it for the full term. I think a few days is sufficient if the code is provided, because there's not the usual RfA caveat that new evidence can come to light — it's all there, right in front of you, in black and white if the code is provided. Sysopped bots should additionally be cleared by a few members of the Bot Approval Group. As for the blocking of account creation (this should really have been in the comments section), I believe that blocking account creation is important because vandals could otherwise create abusive accounts — although this is flexible. — Werdna talk criticism 03:04, 5 October 2006 (UTC)

Blocking an admin doesn't stop that admin from blocking users. Will this bot automatically shut itself off it blocked, or will some other (secure!) alternative exist to turn it off in an emergency, such as a protected page it automatically checks? - A Man In Bl♟ck (conspire | past ops) 12:20, 5 October 2006 (UTC)
Yes, it does. When a sysop is blocked, all of th asysop abilities are also blocked, except for self-unblocking. Naconkantari
My recent test appears to dispute this. — xaosflux ^Talk 12:41, 5 October 2006 (UTC)

Beg to differ. I just tested it. - A Man In Bl♟ck (conspire | past ops) 12:43, 5 October 2006 (UTC)

incorrect. I guess the obvious thing is to program in a protected page check every time a block is about to be issued. — Werdna talk criticism 12:45, 5 October 2006 (UTC)

Thanks for showing me that, I must have been mistaken. Naconkantari 14:41, 5 October 2006 (UTC)

Also, are you absolutely sure this won't cause robots to rise up and exterminate humanity? - A Man In Bl♟ck (conspire | past ops) 12:21, 5 October 2006 (UTC)
(ecx3)~~No, I don't know what the bot did before I disabled its electronic sentience and uplinks to MilNET.~~ Yes, I promise. Hopefully this will allay some of the concerns of the opposers. — Werdna talk criticism 12:31, 5 October 2006 (UTC)
There are many tor nodes on dynamic IPs what do you intend to do about those, reblocking them on each new ip on a frequent basis has reasonable potential for collateral damage with the trail of obsolete IPs left blocked. (I kept a database mapping tor node names to IP over a couple of months, this is a real issue). --pgk 12:24, 5 October 2006 (UTC)
(ecx3)As (I think) stated, blocks will be time-limited to two weeks or so and a job will be run weekly. Any issues can be resolved by an Template:Tl request. Of course, it can run more often and block for less if necessary — this is all parameters. — Werdna talk criticism 12:31, 5 October 2006 (UTC)
Well I'm not sure that's going to do it then. I'll check later but there are some nodes which change IP from day to day and possibly more frequently. Template:Tl is not the most friendly way of doing things (a) it's confusing to some users (b) the whole tor thing is confusing to some admins so they are likely to feel uncomfortable with unblocking. --pgk 12:38, 5 October 2006 (UTC)
The run period is, as I said, variable. Running weekly and blocking for a month was plucked out of thin air — it's not like it's the only way the bot will ever run. — Werdna talk criticism 12:45, 5 October 2006 (UTC)
And, of course, any short-term blocks for Tor are preferable to the status quo, where all Tor proxies are blocked indef on sight. — Werdna talk criticism 13:00, 5 October 2006 (UTC)
I'll have to disagree that this is the status quo, last time I checked there were more unblocked than blocked by a wide margin. (Though many are blocked through the Mediawiki proxy blocker, so it's not as bad as it first appears...) --pgk 14:22, 5 October 2006 (UTC)
As The Tor List is externally managed, will these entries be verified, or if the off-project site updates, will it affect use? — xaosflux ^Talk 12:28, 5 October 2006 (UTC)
That is one of the main tor directory servers (hardcoded in some versions of the tor client), there are others available also. --pgk 12:35, 5 October 2006 (UTC)

Has this been presented to the tech team, I've got to imagine they could parse a list of ip's in to the table more efficiently then a remote bot? — xaosflux ^Talk 12:28, 5 October 2006 (UTC)
Rather than developing this as bot code wouldn't it be better to develop such functionality into Mediawiki? This is a broad observation that we have many capable people who seem to want to solve every problem with a bot, when in fact the functionality might better be placed in the core software (or extension) and would certainly be less controverial (and less open to abuse) --pgk 12:33, 5 October 2006 (UTC)
(edit conflict, to both of you) Yes, I spoke to a number of developers. Simetrical is one. The developers see it as something that should be a stop-gap while we wait for software changes. I'm considering working on an extension that'll do just this. — Werdna talk criticism 12:34, 5 October 2006 (UTC)
Ack, I'm not a developer. I'm a MediaWiki dilettante who answers questions a bunch. Anyway, the reason we have bots altogether is that we have a lot more people with bot-writing capabilities (read: anyone who knows how to program) than people with commit access (read: people Brion/Tim/et al. trust to really for sure not break anything or slow stuff down or add messy and hard-to-maintain stuff to the codebase). Ideally all current bot functions would be done via direct operations on the database, but that's not happening, so we have to take what we can get. —Simetrical (talk • contribs) 21:24, 6 October 2006 (UTC)
Talking to Brion he has no problems w/ running this bot, and we discussed the MW extension - that will come in time -- Tawker 21:59, 6 October 2006 (UTC)

Has anyone checked if a steward can set the blocking permission on this account without setting the other admin flags? — xaosflux ^Talk 12:44, 5 October 2006 (UTC)
Possible, yes. I don't believe the developers are interested in doing so, though. — Werdna talk criticism 12:58, 5 October 2006 (UTC)
Technically possible, requires some MW code modifications. Angela said it's not currently possible w/ the current interface - ideally that's better than a full sysop flag as it's all the bot would need but right now it's an all or nthing approach. If blocking can be assigned pretty much turn this into a request for block permission -- Tawker 13:18, 5 October 2006 (UTC)

I will support only if this is implemented. A Steward could easily assign the flag.Voice-of-All 14:23, 5 October 2006 (UTC)

This flag could be created by anyone with shell access by adding a line to LocalSettings.php. It could then be assigned by stewards, but not bureaucrats. The latter difficulty is probably what Angela was referring to. —Simetrical (talk • contribs) 21:24, 6 October 2006 (UTC)

As this bot is a stop-gap, and must be replaced by a proper extension (that I will write), the developers refuse to create such a flag. — Werdna talk criticism 14:30, 5 October 2006 (UTC)

Thanks for the replies above! (I'm placing this down here for ease of reading/answering.) Tawker states on the bot request page that it'll be manually run, meaning that he'll be able to stop it in case problems arises; however, I'm sure we would all like an additional safeguard, since blocking will not work. You state above that it's possible to implement a functionality that would automatically shut-off the bot if the talk page is edited or if the page is protected, or something similar. Has this been implemented yet? If not, I would love to see such a preventative measure be coded (I imagine it would just be copying already-written code from other bots, like TawkerBot2, with this functionality already) prior to the bot gaining administrator rights. Also, a second question: I know this has been asked several times already, with you giving your assurances that it will not perform other tasks, but I wanted to make this explicitly clear, with replies from both you [Werdna] and Tawker. If the bot is coded for some other functionality not described explictly and clearly at the beginning of this RfA and does not gain express approval through another RfA, will both of you endorse the blocking, shutting-down, and immediate de-sysopping of the bot? Thanks. Flcelloguy (A note?) 14:24, 5 October 2006 (UTC)

Well, Tawkerbot2 get shutdown from irc or blocking, there's no talk page notify code in there (it would kind of defeat the purpose of the bot). However, it's easily possible to make a checkpage that the bot would check every min or so while it's running (to check before every block is a tiny bit of a resource hog... actually, if we do it to check for talk page messages and keep the talk page protected it would work. In response to your second question, if the bot runs any functionality not approved and disclosed in this rfa, yes, it should be de-sysopped. I know that that won't be necessary but yes, in the stuck by lightning chance I went nuts, yes, that would be the best course of action -- Tawker 16:59, 5 October 2006 (UTC)

I'll second that. If we've gone that nuts, block and desysop it pronto. Sounds like a hefty block for one or both of us, too. We will do the right thing and seek approval for any additional admin actions that we want the bot to take. — Werdna talk criticism 17:08, 5 October 2006 (UTC)

Thank you both. Flcelloguy (A note?) 17:14, 5 October 2006 (UTC)

Has the foundation taken out comprehensive cover with Old Glory Insurance yet? -- Netsnipe ► 17:23, 5 October 2006 (UTC)
Yes, I believe Danny has been in negotiations for some time, but we pulled through okay. He had some trouble posting, as the firewall at his work blocked Wikipedia, and he couldn't edit through Tor. — Werdna talk criticism 17:33, 5 October 2006 (UTC)

Questions: Forgive me, but I am unclear on what is going to be done by this bot. I'm gathering from the above that there is a list of Tor IPs that need to be banned, and this bot will parse this list for those IPs, yes? Will these blocks be permanent or temporary? If temporary, how long? Is the list of IPs static or is it ever-changing? How many IPs are there? Before I can consider supporting or opposing, I need answers to these questions. --Durin 21:02, 5 October 2006 (UTC)

Temporary blocks, of a duration to be determined on the BRFA for this bot. As I've stated numerous times, I believe that the RfA is the place to discuss the policy and idea of the bot, and the BRFA to figure out technical details. — Werdna talk criticism 00:07, 6 October 2006 (UTC)

Some of my questions above remain unanswered. I do feel that approving the bot for admin status and then working out the details at WP:BRFA is backwards. I want to know exactly what this bot will do before voting in favor of it. I'm strongly inclined towards opposing right now, and may do so later if for no other reason than it's generated such a large amount of discussion for people asking clarification. --Durin 22:45, 6 October 2006 (UTC)

Question: I apologise if this has already been answered, but the swarm of questions above confuses me - about how many tor IPs are there that need to be blocked? An average number may give more reasoning to show that this is a job that is much too tedious for human sysops to manage (we all have lives, of course). Cowman109^Talk 21:35, 5 October 2006 (UTC)

My most recent data indicates that this is approximately 250 blocks. — Werdna talk criticism 00:07, 6 October 2006 (UTC)

Question: IIRC in the past a list was passed to Brion and added to the existing Mediawiki Proxy Block list. Has this been explored as an option? Generating the list and passing to the Devs on a regular basis? It would seem to provide the same functionality without the objections. --pgk 22:12, 5 October 2006 (UTC)

I wasn't aware of this proxy block list. Still, it'd be nice to fix it with a bot that does the job, and replace that with a MediaWiki extension (I'm already in the planning phase for it) — Werdna talk criticism 00:07, 6 October 2006 (UTC)

This seems to already be covered in part by meta:Proxy_blocking, and works for all projects, not just en: couldn't the existing process just be extended to cover this if needed? — xaosflux ^Talk 01:20, 6 October 2006 (UTC)

"Fix it with a bot?", in what way do you think it is broken? The point here is that it's a manually controlled list of nodes, if you can generate that list it can be updated by the devs, no need for bots doing lots of blocking/unblocking, no need for bot with admin etc. i.e. Without an admin bot it can offer the same functionality today (You could generate the list and hand it to the devs today). This would certainly work as a stop gap until a dynamic mechanism for TOR can be built into mediawiki. --pgk 06:20, 6 October 2006 (UTC)

Moved from User talk:Aaron Brenneman#Actually

We're talking about anywhere from 500-1500+ blocks a month depending on tor activity - honestly, you'd just click block links in a row, it seems a better use of one's time to edit and not drone on w/ bot work -- Tawker 03:46, 6 October 2006 (UTC)
End moved section
That statement is correct, as far as it goes. There is a cost associated with any approach. However, let's first turn ourselves to the question of how big a job this is. The stated estimates range from Werdna's 250 to your 1500. That's quite a wide margin for error, can I ask what these figures are based on? - brenneman ^{L} 04:06, 6 October 2006 (UTC)

I made a previous estimate of 500-1500 and gave this to Tawker. This was based on my own general knowledge. My revised estimate is 250 — when I ran it a couple months ago, it made 236 blocks. — Werdna talk criticism 05:33, 6 October 2006 (UTC)
The number also changes wildly, one week Tor nodes are cool to run and suddenly there's 1000, the next they're boring and there's fewer. It goes up and down kind of like a roller coaster -- Tawker 06:09, 6 October 2006 (UTC)
If that's the case this bot will be useless, run if 1 day the next there are 1000s more unblocked and we are stuffed. To put some hard figures on this though 16-Aug 294 nodes, 23-Aug 304 nodes, 30-Sep 311 nodes, 6-Sep 302 nodes, 13-Sep 275 nodes. i.e. it is dynamic but not that dynamic. --pgk 06:20, 6 October 2006 (UTC)

Can we have the link to where this was discussed with JW, please? Sorry if it's in here somewhere already. - brenneman ^{L} 06:22, 6 October 2006 (UTC)
The discussion was on IRC, in private message. I will not disclose logs of it, as the discussion had other topics, and contained opinions not for public release. — Werdna talk criticism 07:12, 6 October 2006 (UTC)

I believe Joshbuddy has a question at the end of this page, down at the very last neutral, that you might have missed. Would you mind responding to that? Thanks! Flcelloguy (A note?) 14:55, 6 October 2006 (UTC)

I have some new questions regarding issues brought up since I last commented. The first is related to Joshbuddy's question (see above): if this is indeed to be "manually run" and watched the whole time, as argued above, what makes a bot necessary? What distinguishes this from a normal script that could easily be run on your or another admin's account? The second question pertains to the number of blocks: I had previously, judging from your estimates and my quick glance at the testwiki log you linked above, believed that the number of blocks required to be much higher than now estimated (~250). If the number is this low and has little variation, as Pgk demonstrated above, and we already have multiple admins volunteering to block all of them on a monthly basis given an organized page (which has been created already), what is the use for a bot? Finally, new concerns have also been raised that this is a "temporary" measure until MediaWiki implements or revises its list to deal with this. If this is indeed the case, and if the changes will be coming soon (or relatively soon), what purpose will the bot serve, given that we already have an alternate solution (manual blocking)? Also, once the technical fix is implemented, the bot will be useless, not being approved for any other functions. Finally, in the bot request for appoval page, you (Tawker) write that you are "going to login to this one [the bot account] at least once a week to ensure that it's running ok (and to clear the new messages thingie.)" This implies that the bot will be running all the time, even when you (Tawker) are not available or logged-in, yet you stated explicitly elsewhere that this was going to be manually-run. This seems to be conflicting information; would you mind clarifying this? Many thanks. Flcelloguy (A note?) 02:32, 7 October 2006 (UTC)
- I patiently await your replies. Thanks! Flcelloguy (A note?) 01:00, 10 October 2006 (UTC)
  - Found it finally. It is a script... I just like the Tawkerbot name :o - In short, I don't want to run this under my normal user account because a) I said no automated apps as "Tawker" and I rather stick to my word and b)IMHO bot edits and automated stuff should be flagged as such on a unique account. It makes it easier for me to find my own blocks :). As for the number of admins, I'm sure they'll try but it's such donkey work that people will forget and it will turn into another WP:OP or copyvio list... backlog galore. As for the clear messages, that's a stop measure and the clearing if it is well just that, making sure I haven't missed anything. The script is manually run... I double click on an application when I need to run it. Sorry for the delay, this has been so long that I'm getting lost in it -- Tawker 01:14, 12 October 2006 (UTC)
  As this RfA seems doomed to fail, I invite all sysops to help out with either compiling the script with a delay and running it manually, or getting going on this list. Please, please, please, block indefinitely, anonymous only, with account creation disabled. I would really like to be able to edit through Tor on this account. — Werdna talk criticism 04:11, 10 October 2006 (UTC)
  - This request would likely be handled better at WT:OP, have you personally verified that these are OP's? — xaosflux ^Talk 05:00, 10 October 2006 (UTC)
- I would still like some answers to my questions here, prior to my "vote", Werdna and Tawker, even if you don't feel that the RfA is going to pass. This would be greatly appreciated. Thanks! Flcelloguy (A note?) 00:50, 12 October 2006 (UTC)
To add some current numbers (assuming I didn't screw up the scripts), there are at time of writing 299 tor nodes which port 80 available, 148 are blocked indefinitely, 3 are blocked but not indefinitely. The rest aren't blocked, though they could be trapped by the mediawiki proxy blocker (some tor nodes are don't know which ones) --pgk 17:02, 7 October 2006 (UTC)

Support

1st Support Seems like a great idea. --CFIF ☎ ⋐ 00:06, 5 October 2006 (UTC)
(Damn, I got edit conflicted!) Hmm...an unusual RfA, to say the least. But, I assume it's gone through all the necessary bot approval steps, and I trust its owner enough to be pretty sure that it won't go crazy. ;) --Mr. Lefty (talk) 00:07, 5 October 2006 (UTC)
Support as co-nominator — Werdna talk criticism 00:08, 5 October 2006 (UTC)
Absolutely This is a much-needed bot that will automate and clear weeks worth of backlogs. Naconkantari 00:10, 5 October 2006 (UTC)
Full support, the code seems fine and I trust Werdna and Tawker. Draicone ^(talk) 00:12, 5 October 2006 (UTC)
Eh support sure, I trust Tawker and this sounds like a Good Thing. However this seems more like a request for a bot that can block than a request for adminship, I mean it wouldn't really be creating a new admin or anything, and we already trust Tawker with admin tools. I don't really think the RfA is necessary personally. --W.marsh 00:15, 5 October 2006 (UTC)
Of course. I don't really think the RFA is necessary, but I guess it makes sure people are aware of its existence. Cowman109^Talk 00:21, 5 October 2006 (UTC)
Sure why not, if it operates well and does only what it has to do. —this is messedrocker (talk) 00:22, 5 October 2006 (UTC)
Sounds good. --Interiot 00:30, 5 October 2006 (UTC)
Support - obviously a good idea. Yomangani ^talk 00:37, 5 October 2006 (UTC)
ABSOFRIGGINLOUTLY Clearing up the massive backlog will be a great thing. Even with an infrequent glitch or screw up (that would only be expected), the benefits will far out weigh the continued back log IMHO --Mystar 00:50, 5 October 2006 (UTC)
Support Great idea, it seems like it would work well. Hello32020 00:55, 5 October 2006 (UTC)
Your recent support has been reverted by an automated bot that attempts to recognize and repair support votes to Wikipedia RfA articles. From what I've seen from TB2 and TB4 so far, the chances of the bot totally screwing up are minimal. -→Buchanan-Hermit ™/?! 00:58, 5 October 2006 (UTC)
Hey, AVB is outraged.... how can you forget AVB :) -- Tawker 02:12, 5 October 2006 (UTC)
Support per nom. Michael 01:07, 5 October 2006 (UTC)
Support No worries and a great idea. FloNight 01:11, 5 October 2006 (UTC)
Support, useful, satisfying answer to my question. --Conti|✉ 01:13, 5 October 2006 (UTC)
Support of course.--Hús ö nd 01:29, 5 October 2006 (UTC)
Support Interesting and useful idea, and I'd like for enWP to give it a try. Thus, I, for one, welcome our new bot admin overlords.-- danntm ^T _C 01:34, 5 October 2006 (UTC)
Oh please yes. Mackensen (talk) 01:40, 5 October 2006 (UTC)
Support. A good idea in good hands. RyanG e rbil10(Kick 'em in the dishpan!) 01:50, 5 October 2006 (UTC)
We should have software support for this at some point, but until then, this is certainly a good idea to enforce m:No open proxies. I think some people are slightly too paranoid about admin-bots generally, but this one has an extremely simple task to do and the chance of it going amok is probably going to be roughly zero. But even if it does, it's not like there's anything whatsoever it can do that can't be reversed in a couple of hours. —Simetrical (talk • contribs) 01:52, 5 October 2006 (UTC)
~~Oppose. I don't believe the candidate wrote the answers to the questions itself.~~ Support per all of above. I will also add to Werdna that I am glad to see you still contributing strongly despite the outcome of your recent RfA, and hope you will be a sysop your personal self someday soon with this bot work counting strongly in your favor. Newyorkbrad 02:03, 5 October 2006 (UTC)
Support - sounds good to me. —Khoi khoi 02:15, 5 October 2006 (UTC)
~~Oppose~~ per TawkerbotTorA never accepted the nomination. Just kidding Support ofcourse. Stubbleboy 02:34, 5 October 2006 (UTC)
Support. I would probably prefer this sort of thing being dealt with by WP:BRFA in the future (ie. a policy is needed dealing with bots getting a sysop flag). This bot sounds like it will do a good job. DarthVad e r 02:38, 5 October 2006 (UTC)
I trust the users - CrazyRussian talk/email 02:42, 5 October 2006 (UTC)
Support seems like a great idea. Wikipediarul e s 2221 02:47, 5 October 2006 (UTC)
Support. Ral315 (talk) 03:38, 5 October 2006 (UTC)
Support. Daniel .Bryant 03:39, 5 October 2006 (UTC)
Support Weird but cool.UberCryxic 04:03, 5 October 2006 (UTC)
~~Provisional support, leaning neutral. I'm still not sure about letting a bot having admin tools, but am willing to give it a go. – Ch acor 04:06, 5 October 2006 (UTC)~~
Weak support. I would be happier if there had been a short trial period, and then a de-sysopping, so at the long-term RfA (this, in other words), we'd have some performance data, but I guess that'll just have to happen on the fly... --Gwern (contribs) 04:58, 5 October 2006 (UTC)
Strong support however, note that the bot approval proceedure also applies - Stephanie Daugherty (Triona) - Talk - Comment - 05:17, 5 October 2006 (UTC)
Strong Support we need to bolster technical tools for defense of content. Excellent idea. Rama's arrow 05:28, 5 October 2006 (UTC)
Support, seems perfectly reasonable, and I trust Tawker & Jimbo. Themindset 05:31, 5 October 2006 (UTC)
I'm going to cry now. — Werdna talk criticism 06:49, 5 October 2006 (UTC)
Support With the bot-approval process in place and trusted editors in control then this should be tightly controlled and efficiently run. (aeropagitica) 05:45, 5 October 2006 (UTC)
Great idea! Jorcoga 06:14, 5 October 2006 (UTC)
Support our new Tor blocking overlords. Let's give 'em something to Tawkerbout. Can't sleep, clown will eat me 06:29, 5 October 2006 (UTC)
Could we have a pun-reverting bot next, please? - Richardcavell 23:26, 9 October 2006 (UTC)
Support' This sounds good (& safe) to me and definitely save much of human time for such tedious task. --WinHunter ^(talk) 06:35, 5 October 2006 (UTC)
Support Given the experience of the users involved, and the restrictions on the account, it seems like a totally acceptable proposal. Alphachimp 06:58, 5 October 2006 (UTC)
P-p-p-powerbook support. Simple, predictable, repetetive tasks? Sounds like an excellent job for a bot. Saves the rest of us time to worry about more important things. I was initially concerned that more people might need access to a stop button of some sort, but Werdna explained to me on IRC that the bot would only be running once a week, so it seems less important with that in mind. Keep a close eye on it, but let it save us some effort. I trust these two hooligans. And all else aside, remember that this is only a soft block on Tor -- anybody is free to start an account from other IPs, and then edit via Tor. Could solve a lot of problems. Luna Santin 07:02, 5 October 2006 (UTC)
SSupport MER-C 07:04, 5 October 2006 (UTC)
Support under the parameters and limits described above. I would encourage the development of a server-side solution (like the easy incorporation of external blacklists) but in its absence this is a way to maintain the proper blocks. Demi ^T/_C 07:10, 5 October 2006 (UTC)
And I, for one, welcome our new robot overlords. For those who oppose all sysop bots by principle, then why on Earth do we bother running an Autoblocker to begin with? This is not a bot that can be gamed by users, so I strongly support its use. -- Netsnipe ► 07:44, 5 October 2006 (UTC)
Support per Netsnipe. Tawker's already proved the worth of his bots. No reason not to trust these users. Firsfron of Ronchester 10:06, 5 October 2006 (UTC)
Support I found a problem recently with a user who loved to sockpuppet using all IP's which were found on a TOR proxy list, and reporting them was basically a hassle and just contributed to a backlog at WP:OP. It is pretty clear that the IP's are not generally used for any good faith edits as anonymous IP's. Arguing that this will disadvantage users if it is anon only and account creation disabled is null for me as they can visit the page without a proxy once to create the account if they really want to. Any controversial blocks can go through the normal unblock process at WP:OP, or the bot can have a "big red button" for emergencies... Ansell 10:19, 5 October 2006 (UTC)
Support, but the bot must never use admin capabilities for any other purpose without a new rfa. ×Meegs 10:29, 5 October 2006 (UTC)
That simply cannot be enforced; see me comment slightly further down. -Splash - tk 10:57, 5 October 2006 (UTC)
Are you suggesting that Werdna and Tawker would deliberately break a promise that went that much against community consensus? Remember who we are dealing with here. JoshuaZ 11:56, 5 October 2006 (UTC)
We are dealing with the person who, in their own RfA said (non-verbatim) "I will not sysop my bot, because the community wouldn't like that" and then went to RfBot and said "please could you sysop my bot". -Splash - tk 12:38, 5 October 2006 (UTC)
Ok, let me try it this way, Tawkerbot2 is an apple, TawkerbotTorA is an orange. I think it was pretty clear that that comment Splash quoted was with respect to Tawkerbot2 (the apple) and not TawkerbotTorA (the orange.) - The TB2 sysop proposal was after I had request from multiple well respected users to formulate a proposal (and the location was also suggested by several very well respected members of this community) and to be honest, throwing an idea out for discussion isn’t evil. I think the important thing is the fact that this is NOT running on the Tawker account and has been put up for consensus. If there are any changes (say have it automatically probe other open proxies and confirm that way) it would need to be proposed and discussed in a public forum. In short, it has to be run like a crat gathers consensus on promotion, will full transparency. -- Tawker 17:07, 5 October 2006 (UTC)

Splash, you are correct, and I do share your concern, but I am accepting Werdna and Tawker's word. ×Meegs 12:31, 5 October 2006 (UTC)
Unless I'm missing something Tawker's behavior was consistent with what he promised. He did not just run a bot through his own sysoped account, he went back to get permission from the community. JoshuaZ 13:05, 5 October 2006 (UTC)
Support of course. the wub "?!" 10:43, 5 October 2006 (UTC)
Support on the understanding that this bot will do "what it says on the tin" - Wedna has stated any changes of scope in this bot's behaviour will be run under a different account name and will require a separate RfA. Fair enough - who will envigilate this? --Mcginnly | Natter 10:51, 5 October 2006 (UTC) Changed to oppose - see below.

This is not a condition that can be imposed. A sysopping grants total access to the sysop tools to an unlimited degree. -Splash - tk 10:57, 5 October 2006 (UTC)
Personally, I trust Werdna and Tawker but perhaps if Splash is more wary, we can compromise and get them to committ this bot to CAT:AOR - that way we don't need ArbCom delays to hobble the bot if needed? --Mcginnly | Natter 13:09, 5 October 2006 (UTC)
Support – we could do with more of these 'adminbots' – Gurch 12:02, 5 October 2006 (UTC)
Support sounds fine. SMC 12:22, 5 October 2006 (UTC)
Support. I for one welcome our robot masters. I can understand the concerns of the bioforms listed under oppose below, but I feel the TOR anonymizers is a problem that we haven't fully uncovered and while a bio could do the work it would be horrible drudgery. This silicon based admin will in any event be under significant scrutiny. Perhaps we need to make another bot admin that verifies that the first is indeed blocking TOR nodes and isn't sweeping them under the carpet for nefarious purposes that end up with us all living down near the core of the earth and having huge sweaty sex-dances. Syrthiss 12:58, 5 October 2006 (UTC)
Support. It's regretable we have to block Tor users at all, but with the vandalism problem we do. This is a perfect candidate for automation, a simple and repetitive job. The code involved isn't overly complex and there doesn't seem much scope for huge problems. The operators, I believe, are trustworthy. Let's try it. --kingboyk 13:10, 5 October 2006 (UTC)
Support the notion. However, I think that at least for the first few runs, and preferably thereafter, there should be a steward standing by to desysop the bot in case of emergency. --ais523 13:17, 5 October 2006 (U T C) Changed to oppose. --ais523 15:52, 6 October 2006 (U T C)
Support, good idea. Also a good idea to apply through RfA and not just run the bot on a normal admin account, which is the way most adminbots seem to be run (how many are there?). Much nicer if the bot is talked about and approved openly. A MediaWiki-based solution or a move to a toolserver-based bot might be a good idea in the long term, though. Kusma (討論) 13:28, 5 October 2006 (UTC)
Support. As a WP:BAG member the bot seems OK. I am assuming that a)if it is blocked, that it is programmed to NOT do anything and b)the flag issue will be worked on to that it won't need +sysop eventually.Voice-of-All 14:42, 5 October 2006 (UTC)
Support Yes please. --InShaneee 14:46, 5 October 2006 (UTC)
Support Serves a needed purpose. Benefits outweigh risks. Not subject to human frailty and will be supervised byntrusted users.Cheers, :) Dlohcierekim 15:10, 5 October 2006 (UTC)
Support. Does something useful, will be run by a trusted user who will watch it carefully, supported by Jimbo. These things seem to outweigh the quasi-religious "bot must not have sysop bit" arguments made by some users below. -- SCZenz 15:15, 5 October 2006 (UTC)
Support, on the condition that the bot only does what is laid out in this rfa without new approval being sought (so I guess I'll just have to AGF that this will always be the case). Petros471 16:17, 5 October 2006 (UTC)
Support all bot admin actions will be logged and I assume that Tawker will take full responsibility for his pet. Good luck! ˉˉ^anetode╦╩ 16:49, 5 October 2006 (UTC)
Support, opposition presents no compelling concerns, purpose seems to be a useful preventative measure. Christopher Parham (talk) 17:07, 5 October 2006 (UTC)
My concerns have been addressed. Flcelloguy (A note?) 17:14, 5 October 2006 (UTC)
~~I am removing this while I continue to discuss the issue; see my comment with new questions and concerns above. Flcelloguy (A note?) 02:36, 7 October 2006 (UTC)~~
Readded support after further consideration and the questions being responded to. Flcelloguy (A note?) 02:10, 12 October 2006 (UTC)
Support - I think that the opposition has little technical merit and is basing its accusations on emotion and personal opinion rather than fact, and until I see someone broach reasonable proof without also ad hominem attacking the creators, I will keep this support. —K e akealani 17:33, 5 October 2006 (UTC)
Support—this is precisely the sort of mindless task that we ought to have a bot to perform. We need a better process for giving sysop bits to bots in the future, though. TenOfAllTrades(talk) 18:19, 5 October 2006 (UTC)
Support automated (monitored) method of enforcing m:No open proxies. --MPerel ^{( talk | contrib)} 19:15, 5 October 2006 (UTC)
Support until proven as a bad idea Let's see how it goes, I'm interested in seeing what happens. If it doesn't, a crat can always de-sysop it. The ikiroid (talk·desk·Advise me) 19:18, 5 October 2006 (UTC)
Support. Human admins make mistakes (I know I do). A robot admin probably will make less mistakes (or maybe equal to), and cover a wide variety of Wikipedia backlogs, blockings, etc. --Nishkid64 20:29, 5 October 2006 (UTC)
Support. G.H e 21:23, 5 October 2006 (UTC)
Support I don't understand on word of the techno-bable, but I trust those involved. The oppose arguments don't stack up. They seem either technophobic, or untrusting of either of the assurances of the ops, or the ability of the community to call a halt before we're too far down the slippery slop. Could bots do more int he future? Well, IF the case can be made that the specific proposal helps the encyclopedia, why not? Let's trust ourselves to make those decisions (or reject them) when they come.--Doc 22:04, 5 October 2006 (UTC)
What makes you think we would even be offered the opportunity thus to decide? Why not just ask Jimbo (I see this is a "Jimmy" issue, though) on IRC if it seems like a good idea? Or just go ahead and have a user that went though RfA do whatever it pleased with its sysop bit. The rest of us can, after all. -Splash - tk 23:21, 5 October 2006 (UTC)
Support, clearly a pressing need. Tawker and Werdna have my trust, and the oppose votes make no sense to me. Daveydw ee b (^chat/_patch) 22:41, 5 October 2006 (UTC)
Support -- These open proxies need to be blocked, and the script appears to be the best way to do this. John254 23:39, 5 October 2006 (UTC)
Support Anything that'll help the wiki. -Royalguard11(Talk·Desk) 23:42, 5 October 2006 (UTC)
Support this is what bots are for. Most of the opposes sound like they've been reading too much sci-fi. Opabinia regalis 23:58, 5 October 2006 (UTC)
Support - very useful and safe bot. I assume:
Any new functionality of the bot would recure reconfirmation through the RfA
Simple blocking of the bot will stop it. Thus any of 1000+ admins could esily stop it if anything goes wrong abakharev 01:09, 6 October 2006 (UTC)
See the discussion area above, blocking this bot will NOT stop it, it will require a bot-based check. — xaosflux ^Talk 01:32, 6 October 2006 (UTC)
However, any steward or developer could stop the bot by desysopping it, in the highly unlikely event that it malfunctioned. John254 01:56, 6 October 2006 (UTC)

Well, if the bot would have a big red button like User:Tawkerbot2 it is OK with me. Also it is writing something to the talk pages and to the log, would banning it make it crash? Even if its talk page would have clear instruction what file to edit so to stop it, it is still OKI with me abakharev 04:15, 6 October 2006 (UTC)
Weak support. I'm going to throw myself into this one because I trust Tawker, and I do not think that the bot will develop its own intelligence and enslave us all until everyone conforms to NPOV using admin tools. A trial run wouldn't have hurt, but I can't imagine it would be hard to stop it, seeing how quickly we've desysopped actual humans in times of need. Grand master ka 03:17, 6 October 2006 (UTC)
Oh, I almost forgot... Is the bot going to thank us on our talk pages for participating in its RfA? Grand master ka 03:19, 6 October 2006 (UTC)
Support It's just doing one thing. This should not be a problem, and on the off chance there's some screw up, someone will notice it fairly quickly. Bastiq▼e ^demandez 03:41, 6 October 2006 (UTC)
Support Though machines will inevitably enslave us all, this bot is performing a single repetetive task and would be caught fairly quickly if it were misused. It has Jimbo's support, so I see no problem. – Someguy0830 (Talk | contribs) 05:52, 6 October 2006 (UTC)
Support I run a similar script on en.wiktionary.org, under my account wikt:User:Connel MacKenzie and have not encountered any major problems, to date. --Connel MacKenzie - wikt 06:03, 6 October 2006 (UTC)
I'd like to take a moment to explain my "support" vote, despite my reservations that TawkerbotTorA (as proposed) does not do nearly enough. As a side note, I will say as a sysop and checkuser on en.wiktionary.org, that I have never seen a legitimate edit on Wiktionary, originating from a tor node. Nor do I know of anyone who has. (I expect some "oppose-voting" twit now, to connect through tor and point to their one edit here. Go ahead.) I personally have reverted hundreds (probably thousands) of edits originating from the tor network. For en.wiktionary.org, because we deal with a smaller community, the number of people who even understand what tor is is rather small. None of the regular contributors care to delve deeply into (inherently lawbreaking) cracker tools such as tor; the contributors that do stay around on en.wiktionary.org care about lexicography. We get very frequent complaints about templates being too hard to use/too incomprehensible (e.g. "How did my entry get into wikt:Category:English nouns?") There is absolutely no foreseeable need to allow tor nodes access to the English Wiktionary. That is why the nodes are completely blocked there, not allowing registered users, not expiring after one month. Again, the only "contributors" that have edited en.wiktionary.org from tor nodes have been vandals (adolescents who think it is funny to post goatse in a dictionary.) Because those vandals have created thousands (~15,000 or ~70% of total) of sleeper accounts, it would be insanely unsafe for en.wiktionary.org to lessen those blocks, or to allow "registered users" to access there from tor nodes. I also do not think the Wikipedia concept of unblocking is coherent. If one check at one time shows that host as being down, you would unblock it (or let a block expire.) That is frightfully insane. Once a "bad" node has joined the tor network, it is unlikely they will overnight see the light, and become "good people." Even if they periodically disable their exit point, it is only to circumvent an automated check, such as TawkerbotTorA. As soon as the node is white-listed, you can rest assured that the node is active again. To the lunatics voting "oppose" here, please do not visit en.wiktionary.org; you'll find your attitude is not appreciated there. And if you push some "free the tor nodes" agenda, you will be blocked. Perhaps on Wikipedia, the concept is that you can always add more sysops if rampant vandalism is becoming too common. That concept is strangely backwards: why allow known bad contributors access in the first place? You don't have an infinite pool of sysops to revert the vandalism. It is not as if you have dedicated contributors using tor: you have only a theory that someone in China might someday learn enough hax0r 5ki11z to use tor, until their government catches them and shoots them. Meanwhile, the script kiddies are having a field day. For those of you (Wikipedia admins) that doubt what I say, please try this experiment: unblock all the tor nodes that are currently blocked. If you really think they should be unblocked, then, why, by all means, you should unblock them all, right? But please don't act surprised if you are de-sysopped. Despite all these concerns I have about the inadequacy of TawkerbotTorA, I vote support without reservation, as it is the first of many (far too-cautious) baby steps in the right direction. --Connel MacKenzie - wikt 16:41, 11 October 2006 (UTC)
Support Indeed, useful function, solid owners. - TheDaveRoss 07:18, 6 October 2006 (UTC)
Support. I'm happy with the guarantees given.-gadfium 07:38, 6 October 2006 (UTC)
Please stop vandalism support. This is obviously a great idea and I support it :) -- Grafikm ^(AutoGRAF) 08:13, 6 October 2006 (UTC)
The Curps experience shows that we are ready to be ruled by bots. What especially convinces me that this is so is that when Curps' blockbot went awry and blocked good-faith contributors people's reactions were not to block Curps but to blame the victim and only grudgingly unblock. After all, to err is human. Haukur 11:06, 6 October 2006 (UTC)
Support. Yes. Definitely. Tawker, I am in your. Something or other. :) --Woohookitty^(meow) 11:28, 6 October 2006 (UTC)
Support per above. —Jared Hunt October 6, 2006, 12:19 (UTC)
Support. Let it be said that I support this bot fully and trust its operator fully as well. Update: I don't know who will read this, and whether it will sway some votes, but as a member of the Bot Approvals Group, I will firmly stand against feature creep for this bot. If it is granted the admin status, it will not get any other functionality unless it goes through an RfA. I am quite sure the other members of the BAG would agree with me. This should satisfy many of the supporters and opposers who have this concern. The only way this bot could run with new features is if it did so unapproved, which is the same as if a regular admin used his regular admin account as a bot. Both are not allowed, but both are technically possible and logically identical. Since blocking is an important task and needs oversite, this bot should not run with a bot flag. I don't see why else we should not approve of this task. Tawker is a respected bot operator and Werdna is a respected bot writer and operator. Many of the opposes are on the same level as when the rambot was first opposed simply because it was a bot. I had done over 3,000 rambot-style edits manually without a bot, and no one complained. But when I did the same 30,000 type articles with a bot and told people about it, then suddenly there was a problem. This anti-bot bias has no place in Wikipedia. The actions performed by the bot are useful, have the implied approval of Jimbo, and are just plain a great idea. The bot code has been published, and many people have already commented on it, so it's great on that front too. -- RM 15:03, 7 October 2006 (UTC)
Weak Support subject to the assurances given here that this will not become an example of scope creep. Just-good-enough will do, but this can only be a temporary solution. Angus McLellan (Talk) 12:39, 6 October 2006 (UTC)
Support. I assume good faith. Let's give it a try. This is not asking for a license for nuclear weapon construction. If there is something wrong with this, Jimbo or the ArbCom can remove adminship as needed. --Ligulem 12:47, 6 October 2006 (UTC)
Support. It seems to me the concerns of those who oppose are unfounded. If I've understood correctly, it's more of a script run by Tawker than an AI roaming around 24/7. D e lta Tango | Talk 15:18, 6 October 2006 (UTC)
Support. I trust Tawker's ability to write a script that won't cause the servers to implode. This is apparantly a huge boring task no-one wants to do, so why not make it semi-automated? --tjstrf 16:17, 6 October 2006 (UTC)
Support. Hemmingsen 16:45, 6 October 2006 (UTC)
Support. There is a need for this, and Tawker has put in significant work making a bot for this tedious task, and I'd hate to see that go to waste. Adminship should be no big deal for approved bots when we trust the owner as much as we trust Tawker: I really don't think this can cause much in the way of problems. Mango juice^talk 17:49, 6 October 2006 (UTC)
Support. It's tor nodes... Sasquatch t|c 18:41, 6 October 2006 (UTC)
Meh, Support. I will be very, very, very hulk-style angry if I later learn this account or this bot undergoes any capability-creep without full community buy-in. —Bunchofgrapes (talk) 22:35, 6 October 2006 (UTC)
Support I trust werdna and tawker completly. -Ravedave ^{(help name my baby)} 01:27, 7 October 2006 (UTC)
Support. A good idea. Bucketsofg 01:36, 7 October 2006 (UTC)
Support. Its really no big deal. --Jay(Reply) 02:51, 7 October 2006 (UTC)
Support while I can see the objections to having a bot with admin tools, only Tawker will have the password for the account, and the code looks clean. Added to my complete trust of both Tawker and Werdna, I see nothing wrong with this. jam es^(talk) 03:03, 7 October 2006 (UTC)
Support. If it really sucks, you can request de-sysopping. pschemp | talk 04:05, 7 October 2006 (UTC)
Support—the battle of Cain and Abel played out before our very eyes. I favor technology; especially technology with a human at the kill switch. Let's do it. Williamborg (Bill) 05:48, 7 October 2006 (UTC)
Support I had my doubts until I read the question and answer section. As long as it's implemented as expressed here I'll stand behind it (unless of course it becomes too buggy to remain feasible). Durova 05:56, 7 October 2006 (UTC)
Support Nice idea! Good precedent! Congratulations, my friends! Forward to the Brave New World! --Rednblu 07:26, 7 October 2006 (UTC)
Support, of course under the expectation that its functions will be firmly restricted to exactly what the proposal says now. (Which, contrary to what Splash says, can indeed be enforced, not technically but by social contract) Fut.Perf. ☼ 09:26, 7 October 2006 (UTC)
Strong Support I trust both users running this bot. The task is important, the tools are in trusted hands, there is nothing to fear.--Konst.able 11:01, 7 October 2006 (UTC)
Support. As a non-human admin account, there is little doubt that the flag would be removed easily if this bot ever caused problems. Let's do it. NoSeptember 11:46, 7 October 2006 (UTC)
Strong Support Oppose !voters seem to have seen too many sci-fi movies :) Give it a try, why not? If anyone knows what they are doing, Werdna and Tawker do. And if it cocks up, there'll always be someone to say "I told you so" and ~~remove the flag~~desysop. — riana_dzasta _{wreak havoc-damage report} 12:06, 7 October 2006 (UTC)
Support - Sounds like a good idea, from the sounds of things, has Jimbo's support, why not?--Hamedog_Talk|^@ 14:31, 7 October 2006 (UTC)
Support more pros than cons, and it can always be put on "admins for recall" Agathoclea 15:28, 7 October 2006 (UTC)
Support This is a very interesting concept. If this goes through we might be able to create bots to assist in clearing the admin related backlogs. Tarret 01:04, 8 October 2006 (UTC)
Support - David Gerard 02:12, 8 October 2006 (UTC)
Tony Sidaway 02:19, 8 October 2006 (UTC) This bot can be stopped by removing its bot flag or blocking. Its primary functions can be disabled by desysopping. Its actions can all be reversed. Tawker is a respected and well behaved editor.
Some clarifications: blocking will not stop the bot, as an admin account can still perform admin actions, such as blocks, while the account itself is blocked. Removing a bot flag will also not stop the bot in progress, as bot flags do not affect the functioning of an account except whether or not the edits appear in RC. Yes, desysopping will stop the bot, though. Please see the above discussion for more on this. Thanks! Flcelloguy (A note?) 02:44, 8 October 2006 (UTC)
Conditional support - as long as the bot is completely supervised. --Ixfd64 05:02, 8 October 2006 (UTC)
Support I've read the discussion, many good questions and responses. Tawker does a fine job, I support the janitorial tasks of this bot, and on wiki almost anything can be fixed. There seems to be little chance of malicious behavior here relative to other admin requests, some of which I have a harder time with. Needs the e-stop and someone paying attention every so often, and it looks like it will have those... Good luck! Augustz 05:20, 8 October 2006 (UTC)
Weak and conditional support This should be monitored severely. Any overstepping of the proposed admin duties should result in swift desysopping, no questions asked. Pascal.Tesson 06:20, 8 October 2006 (UTC)
Support, the safeguards seem to be in place. Andrew Levine 08:46, 8 October 2006 (UTC)
Support. It's simply a bot helping the humans by doing boring work. Shanes 10:02, 8 October 2006 (UTC)
Support. ~~I, for one, welcome our new blocking overlords.~~ Update: I now feel that Tawker should simply run this under his account or have TawkerBot's approval updated appropriately to be allowed to perform this as a scheduled action. I remain a support because I have confidence in Tawker's ability to run this safely and RfA is not the place for the design of the bot's mission. - CHAIRBOY (☎) 15:09, 8 October 2006 (UTC)
Cautious support - I don't trust robots. I think the main problem with them is that they're incessant, and have no way of understanding "enough's enough". Remember The Sorcerer's Apprentice? Anyone building and running bots should watch that Fantasia segment at least once a year. I've seen complaints made on Village Pump about how annoying it is that we've got so many bots and semi-bots running around alphabetizing everything. Some things, it turns out, God never meant for us to alphabetize. That said, I think it's inevitable that we'll have bots doing things like blocking eventually, and this particular occasion, with responsible people running it, and with many eyes on it and hands on its plug, is probably about as safe a way to break the ice as we'll find. Besides.... when they take over, we'll make great pets. -GTBacchus^(talk) 00:43, 9 October 2006 (UTC)
Support Abuse of Tor by vandals and banned users is becoming a real problem. ≈ jossi ≈ t • @ 02:49, 9 October 2006 (UTC)
It's only a very rough metric, but the number of red-linked talk pages on User:Thatcher131/Torlist seems to suggest otherwise. Even for those IPs where there was a talk page (and hence a warning for vandalism) a random sample of ten had only one with more than five edits. Can I ask what the statement "becoming a real problem" is based on? Not suggesting that it's not true just wanting to be able to evaluate it for myself. - brenneman ^{L} 05:42, 9 October 2006 (UTC)
Conditional support. If the bot can be shut down by any admin, go for it, if this requires a steward, no go. -- Kim van der Linde ^{at venus} 07:55, 9 October 2006 (UTC)
Support There seems to be a lot of Luddite mentality around here. Granting this RfA would benefit the project more than many a human RfA.
Support, I think this is an excellent idea.--Isotope23 20:18, 9 October 2006 (UTC)
Support - this is a no-brainer. - Richardcavell 23:26, 9 October 2006 (UTC)
Support Definitely need to let the bots do the grunt work. -Gphoto 01:32, 10 October 2006 (UTC)
Support--ragesoss 02:04, 10 October 2006 (UTC)
Support Doctor Bruno 16:36, 10 October 2006 (UTC)
Conditional support: provided the bot is able to create reasonably good contents (?) and the community agrees that this bot is capable of assuming bureaucratic functions after proving its worth as a bot-administrator, a number of them (administrators) are like that only (?). I feel excited at the prospect as we would not require the volunteers numbering more than a million to build the Project! --Bhadani 17:41, 10 October 2006 (UTC)
Support. I think the comment above about Luddite mentality is bang-on... which seems quite ironic considering this is an electronic resource. Aside from that, I think the RfA is absurd-- if a bot can do... let the bot do its thing and we can all build a better resource. If there is one concern... it is that the bot is written in C#, a language created by a convicted monopolist that has bashed the GPL at every turn. Nephron T|C 19:47, 10 October 2006 (UTC)
While I'm somewhat sympathetic toward your view of Microsoft, the choice of programming language is dictated by the circumstances in which the code must be written and executed. I have no reason to think that C# was an incorrect choice. - Richardcavell 03:04, 11 October 2006 (UTC)
Luddite mentality - I marvel at the extent of imputing motives and not assuming good faith, and making fun of other person's comments. That is the precise reasons that more than one million wikipedians are unable to create a real encyclopedia despite sinking of vast resources by donors. In case, you require a bot to do certain specified functions, it is certainly a very bold style to accord admin capabilities on it. We may require such bots for many routine admin activities, but it may be done without wasting our time by bringing this to an RfA but by initiating changes in our policies in this respect. BTW, the conclusions derived from such disruptive comments would be something like that if a "bot" were to write the comment: 66% of wikipedians have Luddite mentality. --Bhadani 16:40, 11 October 2006 (UTC)
Green light here. Especially since it's little more than transfer of (Tawker's) power to a sub-account. M isza 13 20:14, 10 October 2006 (UTC)
Support - I know I am already one and this is just a way of showing automated edits vs not... the bot otherwise might already be one :) -- Tawker 15:06, 11 October 2006 (UTC)
Ummm, Aren't you in affect !voting in your own RFA here? — xaosflux ^Talk 17:04, 11 October 2006 (UTC)
Ummmm... doesn't the nom usually !vote support - my RFA can be found about 8 months ago. Ok, I got bugged about the fact that I didn't !vote here :) -- Tawker 17:39, 11 October 2006 (UTC)
Support. Tawker knows what he's doing, we need to crack down on open proxies, and it's got measures against malfunction. haz (talk) e 15:12, 11 October 2006 (UTC)
Unequivocal support - It's a transparent script which does what it says on the tin - unless there is an automatic function on wikipedia by which a sysop is unblocked when they try to edit as a blocked user- which there may be, I don't know - then I don't see how the bot can circumvent a block placed upon it - it's been written by Tawker, someone we can trust, and the important thing to remeber is that it won't re-write itself to unblock if it gets blocked! M a rtinp23 21:26, 11 October 2006 (UTC)
Also, looking at the code, there are a few lines which read:
if (ed.GetWikiText("User:Werdna/StopTBTA").Contains("Stop"))
return; //Shit hit the fan

For the many who don't understand this stuff, it means that if the page User:Werdna/StopTBTA (which I'm sure will be created) has been editted by an admin to contain the word "Stop", the program will die (ie stop editting!). Simple! M a rtinp23 21:32, 11 October 2006 (UTC)
Wholehearted support with (cow)bells. --Alf ^melmac 23:21, 11 October 2006 (UTC)
100% total absolute Support - frankly I dont get it. Every possible safety measure is built in, it only blocks known listed Tor IPs, it has an emerency feature, it has been fully tested without error, its code is completely transparent and it's developed by Tawker who has given personal assurance and runs bots we now can't live without?! This is crazy! Get it going and lets clear some backlogs! Glen 01:05, 12 October 2006 (UTC)
Support. Rebecca 10:07, 12 October 2006 (UTC)

Oppose

The counterpoints provided by Werdna and Talker are too close to The Terminator or The Matrix to my taste. Humans out of control are bad enough, but even the possibility of a robot out of control could be disastrous. No thank you. P eople Powered 01:30, 5 October 2006 (UTC)
Can you provide some form of logic here, or are you just idealistically opposed to bots? I think that, given the fact that we've published the code, I've tested it for three weeks on my own Wiki, and we've got two trustworthy users running it, there's very little chance that it'll get out of control. In fact, I suspect it's more likely that a himan would get out of control. — Werdna talk criticism 01:36, 5 October 2006 (UTC)
Who's to say that a human controlling these robots wouldn't get out of control? The rules of Wikipedia seem too vague and open for just such a device to cause serious havoc. It's not technology, it's sociology i'm worried about here. P eople Powered 02:43, 6 October 2006 (UTC)
Well, short of some sort of radioactivity frying my computer and messing up code, programs usually function as intended. Which counterpoint are you referring to anyways, I'm a little confused. Computers only do what they're told to do - there's a reason we test code in sandboxes. How is a computer more likely to screw up than a human? I'm just a tad confused -- Tawker 02:09, 5 October 2006 (UTC)

~~And if it were to get out of control, it would be blocked and its actions reversed.~~ —Centrx→talk • 06:02, 5 October 2006 (UTC)
I've been informed that blocked admins can still block users. Anyway, it is a manually operated bot—script would be a more accurate name; any problems would happen whilst someone is running it and can be stopped and reverted—as long as no artificial intelligence is added to the code. —Centrx→talk • 06:11, 5 October 2006 (UTC)
The Wikipedia:Bots/Requests for approval/TawkerbotTorA page specificially states this is NOT manually operated. — xaosflux ^Talk 01:32, 6 October 2006 (UTC)
Oppose per JoshuaZ and People Powered and the fact that I just don't like the idea of giving a bot admin status. - Mike (Trick or treat) 02:31, 5 October 2006 (UTC)
Uh, what? You're opposing because you don't like the idea. Please, I'd like some facts and logic here, not half-baked prejudices. — Werdna talk criticism 02:48, 5 October 2006 (UTC)
Oppose - as it doesn't allow account creation for users on tor proxies. There are many legitamate reasons to use tor, if account creation through tor proxies is allowed I would support (if I have misunderstood the situation drop me a note) --T-rex 03:08, 5 October 2006 (UTC)
Enabling account creation through open proxies opens the encyclopedia up to username creation vandals. Naconkantari 03:09, 5 October 2006 (UTC)

This would defeat a large part of the whole point of blocking these proxies. JoshuaZ 03:15, 5 October 2006 (UTC)

Additionally, tor isn't like an ISP. I would oppose blocking ISPs (like AOL) on this basis but tor is a bit of a special case. Our policy is to just block open proxies, period--the alternative to the blocking described would be to just flat-out prohibit tor, this actually opens up access. Besides, the blocking policy isn't exactly the point of this request, as I understand it. Demi ^T/_C 07:10, 5 October 2006 (UTC)

It's policy to block open proxies on sight. Editing from them is forbidden. Account creation is blocked to prevent the mass-generation of vandal accounts. There are legitimate reasons to use tor, yes, but Wikipedia doesn't allow it. Mackensen (talk) 12:00, 5 October 2006 (UTC)

As I understand it, not preventing account creation would make the exercise a waste of time. The application says that logged in users using Tor won't be blocked. If we allow an anon over Tor to create an account, a vandal would just create a new account over Tor (bypassing any autoblock on their own IP address) and then vandalise away without being blocked. The proposed solution is about as lenient as can be (the only workable alternative is banning editing over Tor by anyone, logged in or not). --kingboyk 12:16, 5 October 2006 (UTC)
Given the explainations sent to me it appears if I have read the situation correctlly, so I maintain my oppose of this rfa --T-rex 16:24, 5 October 2006 (UTC)
Apparently not, account creation will be blocked on the IPs regardless of the bot. The difference is it will take someone 3 hours to get through each batch without the bot. —Centrx→talk • 18:57, 5 October 2006 (UTC)

(reply to original comment) -- it's important to not allow account creation from proxies, or else ANY IP address that has an anon-only block can be circumvented by creating a new account through a proxy. Mango juice^talk 17:49, 6 October 2006 (UTC)
Absolutely not. They'll be back next week wanting to block accounts with dodgy usernames, the week after that with wanting to delete short pages, the week after that wanting to use it to close AfDs and automatically nuke possible copyvios (all have been suggested in the past). Only this time, they'll say "the bot has already been approved, so we don't have to go to RfA", we just go to RfBot, a page that will approve bots for just about anything without thinking about it. What the hell's the meaning of the username, anyway? If Jimbo endorses it then he should just sysop it by himself rather than seeking a figleaf RfA. I also object to Werdna's attitude here. Don't treat people like they're really bloody stupid just because they don't agree with you. Expansion because I know what people are going to say on closure: the other points raised by the opposers (principally lack of preparation, lack of necessity and lack of switchoffability) are also valid and for clarity I adopt those positions also. -Splash - tk 09:54, 5 October 2006 (UTC)
Objecting to my attitude? I really don't see how I'm treating anybody as "bloody stupid". I'm asking for some sort of logic, rather than feelings and prejudice. I can accept logic and facts as a reason for withholding the sysop bit for this bot, however I cannot accept personal prejudices. For the benefit of the closing bureaucrat, the concern here is that this requuest for adminship will open the door for using the same account for other purposes. As I've stated numerous times, there is no way I would possibly ever even consider doing something other than the task it's been approved for. I assume that the user has read the RfA page, and therefore does not trust myself and Tawker. — Werdna talk criticism 10:11, 5 October 2006 (UTC)
You understand correctly: I do not believe that the bot will remain restricted to its presently-proposed task. In the past, Tawker (whom I generally respect) has made similar promises and not adhered to them. -Splash - tk 10:15, 5 October 2006 (UTC)
Personal issues of trust are unresolvable by discussion. I will let the current tally speak for itself. — Werdna talk criticism 10:16, 5 October 2006 (UTC)
trust has nothing to do with it since I would regard such ations as those of a rouge bot and act acordingly.Geni 11:26, 5 October 2006 (UTC)

Splash, do you have specific examples where Tawker has done so? JoshuaZ 12:21, 5 October 2006 (UTC)
Yes. See my reply to you further up. -Splash - tk 12:38, 5 October 2006 (UTC)

Wow, just wow. There are ways to oppose a nomination without indicating your sovereign contempt for everyone here. There's a pressing need for this and you're treating it like the thin end of the wedge for some kind of wiki-bot-dictatorship. I'm sorry, Splash, but I don't see any real grounds for concern and I'm puzzled by your hostility. Mackensen (talk) 12:05, 5 October 2006 (UTC)
I didn't express contempt for anyone, save distaste at Werdna's approach. Although yes, I suppose I have to disagree with each of the supporters, but that's the dichotomy of !voting. I listed some examples of things that people have mooted having bots for in the past, right up to actually asking for a sysop bit in one case. Tawker on his RfA said "don't want a bit for my bot" then went to RfBot and said "please can have a bit for my bot". Having sysopped this bot through RfA, when someone queried whether it should now also do "X", anyone objecting would be dismissed with "it went through RfA, it can do what it wants", "you'd have to be really silly to not want this dull task done automatically" and "we all love tawker anyway". Hey, that's not entirely unlike this RfA. Thin end of the wedge, yes, absolutely that's what it is. But I can't see where I talk about a dictatorship. -Splash - tk 12:38, 5 October 2006 (UTC)

If another bot task comes to mind another proposal goes to RfA - do you honestly think a bot named TorA is going to be a good name for a username blocking bot anyways - that'd suck!. The main reason this went to RfA is Jimbo is not a dictator - community consensus is a good thing. Now, would you care to bring to light which promises I have broken. I'm actually rather confused on that statement. This isn't TB2 by the way - no connection except me (hell, it doesn't even run on the same OS) -- Tawker 13:28, 5 October 2006 (UTC)
We can aargue what Jimbo is/not another time. I replied to your latter point above and below; I'll spare the kilobytes and refer you there. I understand that it is not TBx. -Splash - tk 13:55, 5 October 2006 (UTC)

Why do you think this would go unnoticed? Why not have it that the bot could be de-sysopped quickly, in the recognition that it is not a person and so normal ArbCom rigmarole would not apply? —Centrx→talk • 19:03, 5 October 2006 (UTC)
How would we go about having that? If some new community architecture is needed, then it should in place first, not ad-hoc as we go, sort of maybe perhaps afterwards. -Splash - tk 23:21, 5 October 2006 (UTC)

And why do you think Tawker or Werdna would prevent the emergency stop function? If either were to do that, the problem would not be that the bot has a sysop bit, but that either of them are allowed to remain users at all. —Centrx→talk • 19:08, 5 October 2006 (UTC)
This seems not to be related to my comment? -Splash - tk 23:21, 5 October 2006 (UTC)
Oppose. Unless there is a real emergency (i.e. Willy on Wheels being bot-blocked by Curps to prevent huge and laborious clean-up operations), I am very uncomfortable about handing a block button to a bot which does not possess human eyes. Tor-IPs and open proxies can be a pain, but as long as those forbidden IPs are not editing, they do not present such a threat to Wikipedia that they require blocking which is so urgent that we need a bot to do it. The danger of a false positive is present, and one such incident will be more serious than missing two unused tors. Sjakkalle (Check!) 11:55, 5 October 2006 (UTC)
Oppose This is an Isaac Asimov story, right?--Londoneye 12:21, 5 October 2006 (UTC)
Oppose Initial oppose, may review this after additional questions above are dealt with. As a new bot proposal this probally should have gained it's consensus at WP:RFBOT before coming to seek priv's. My major oppose reason is that this appear to be designed to block editors/potential editors based information not validated by an admin. — xaosflux ^Talk 12:32, 5 October 2006 (UTC)
Just FYI, pretty much every block on WP:OP is verified by a human running a script that basically does everything automatically - it loads a page via the open proxy to see if it is open or not and if it's open it gives a blocklink for the user to click on and press block. Any sysop could manually go through and block every IP on that list, are you concerned about the generation source or the fact that nobody confirms that those ip's are listed on the tor proxy page? -- Tawker 17:18, 5 October 2006 (UTC)
My oppose on that note would be that should an externally hosted list have errors, get corrupted, be hacked (server side, rouge wise, dns poisoned, etc) if these addresses are not validated as actually being effectual open proxies, this could lead to the blocking of legitimate anons. — xaosflux ^Talk 18:10, 5 October 2006 (UTC)
Oppose, for several reasons. Firstly, putting this through RfA is wrong, because the nominators are only seeking to give the bot sysop permissions for very specific uses, and adminship grants far wider permissions. Secondly, if the existence of this bot is sanctioned by Jimbo Wales, it doesn't need an RfA. Thirdly, this bot has a terrible name and needs a better (more explanatory) one, which will make some sense to those who find themselves blocked by it. Fourthly, the bot, if it is to be granted the sysop bit, should be operated by several different senior administrators, not just one, so that if it goes haywire it can be stopped regardless of whether Tawker is awake at the time. Finally, and most importantly, I oppose on the ground that this is a bot, and bots should not have sysop privileges in any event. - Mark 12:36, 5 October 2006 (UTC)
Forgive this long oppose, but it's going to come in two stages:
1. For the pragmatic reason that anyone who's worked in develop will feel in their bones: This won't end up temporary. It's a terrible kludge while being the best technical option we've got. And once it's running, the pressure to get a real solution will go to zero. In three years this bot will still be running, because it will be (barely) getting the job done and devloper time is one of the most precious resources we have.
2. For the philosophical reason that I don't want a bot with the bit. Again, once we've got one solution working, it's terrible tempting to apply that solution to everything. Cyde already appears to have a bot running via his main account, and once we validate its use, every step becomes easier. Why not have a bot close unanimous XfDs, for example? Unprotect pages after a certain time, then re-protect them if foo number of edit summaries with "rv" appear? And image deletion, for the love of Bob why doesn't a bot do that already?
While I'm sure this is a pain-in-the-pickle for someone to do, someone is actually managing to do it. And, as nicely as possible, this smells like a solution in search of a problem. It's not as though we're reading every week in the signpost how Tor Proxies are bringing us to our knees. Finally, I'd like to see some diffs provided by Splash, because he's made some strong statements that are relevent. - brenneman ^{L} 13:22, 5 October 2006 (UTC)
The link provided below by Chacor contains the stuff I'm referring to. That particular diff from Essjay is pretty accurate a summary (imo), but you will get the full story if you read that thread from the top, and also the (non-archivable) full RfBot discussion at (WoW gone as fast as normal vandalism) (I think that's where it ends). Tawker's RfA is here. You can see from there what he said. -Splash - tk 13:55, 5 October 2006 (UTC)
Thank you both. Since my apocryphal slippery slope has apparently already been started down, I feel more secure in that section of my oppose. It's not about not trusting these guys, or thinking that one prevaricated on the issue of suffixes, it's just that these things are easy and tehcnical solutions are seductive. But usually craptastic compared to a person's brain. - brenneman ^{L} 15:04, 5 October 2006 (UTC)
Christ on a cracker, we're talking about 250 blocks here? Once a month? There is no need for this bot, I will personally perform the required 250 blocks a month if the pasred list is proved to me. - brenneman ^{L} 02:48, 6 October 2006 (UTC)
Oppose, switched from neutral. I think Essjay got it right here regarding a similar earlier attempt to have one of Tawker's bots sysopped. – Ch acor 13:26, 5 October 2006 (UTC)
Oppose for a relatively narrow reason. Mike is absolutely right: the bot must have more than one human admin master, for situations when Tawker sleeps (carbon-based life is so frail.) Xoloz 14:33, 5 October 2006 (UTC)
Isn't the bot supposed to be manually started and not running continuously? In that case, it wouldn't need another master. Kusma (討論) 14:44, 5 October 2006 (UTC)
Call me crazy ("you're crazy, Xoloz"... "Why, thank you!"), but I worry that Tawker might have an off-day and lose track of things. A back-up person is never a bad idea. I hoping this concern of mine will be remedied before RfA close, really. It shouldn't be hard to give access to some other techno-savvy admin (Interiot?) Xoloz 15:58, 5 October 2006 (UTC)
Implemented an emergency stop function. Copies of the code with the username and password will be distributed to trusted sysops. Please note that technical parameters will be negotiated on the BRFA page once this RfA completes before the bot commences duties. — Werdna talk criticism 16:42, 5 October 2006 (UTC)
For the record, although my initial concerns were addressed, I am persuaded to reaffirm my oppose by the arguments presented below, especially Nandesuka's. Xoloz 15:37, 11 October 2006 (UTC)
~~Oppose.~~Strong oppose Sounds good on paper, but real life ??? too many unknowns. Checked the User's contributions and found that user TawkerbotTorA falls way short of my criteria for supporting. I encourage user TawkerbotTorA to not get discouraged- many Rfa's fail the first time around. Will support after several months of good editing and interaction with other users. Jcam 17:50, 5 October 2006 (UTC) Admin or not, bot operator has given me reason in the past week (through actions and words) to not assume good faith. Would not be surprised to see entirely different code and functionality after approval. Jcam 00:20, 12 October 2006 (UTC)
I think you subst'ed in the wrong template - this account isn't likely to have any interaction w/ users or edits :) -- Tawker 18:22, 5 October 2006 (UTC)
I think I also failed to mention that TawkerBotTorA has poor use of edit summaries. Of his zero edits, he/she/it has NEVER used edit summaries. Jcam 18:38, 5 October 2006 (UTC)
Well, that's a fair point. On the other hand, to run through some common reasons for opposing RfA's as to which Tawkerbot comes up clean, it's never been accused of being uncivil, never made a personal attack, never been blocked, never rouged an XfD, never had any unseemly userboxes on its userpage...... Newyorkbrad 22:34, 6 October 2006 (UTC)
Valid point. All we need is one of the Supports to say I thought he/she/it was one already ! to sway my vote Jcam 22:48, 6 October 2006 (UTC)
Oppose per Splash, Chacor, Aaron Brenneman and Thatcher131. Get a steward to give it the blocking bit specifically and distribute the password so it can be shut off even if the code's changed and goes haywire. ~~Having a non-admin (Werdna) administering an AdminBot gives me the willies;~~ this is much better handled by a human, and the prior examples of Tawker/Werdna trying to end-run their bots into Adminship makes my bloody skin crawl. Sorry guys, no, not as currently presented. If this is such a big deal, make a list of proxies and distribute the load among our human admin corps. I'm not seeing the pressing need. -- nae'blis 18:58, 5 October 2006 (UTC)
Oppose Bots should not be granted admin status due to the reasons given above. --S iva1979^{Talk to me} 20:43, 5 October 2006 (UTC)
Oppose Per the above listed reasons, and also from the standpoint of having any bot running as sysadmin status is a bad thing. Not saying this bot's creator will do this, but the rogue possibility always potentially exists of the program being altered either accidentally or deliberately with disasterous consequences. EDIT: Also, while opposed on a philophosical basis to banning of systems such as TOR (I believe ulimately more harm is done to humanity by such restrictions), if the issue really is such a pressing and critical matter it should be done by hand. Create WP:TOR and get people involved in finding/blocking them. · XP · 20:50, 5 October 2006 (UTC)
Oppose. I'm still not entirely sure this is a good idea, but even if it were, this is something that needs to be implemented at the MediaWiki level, as this is certainly not the only Wikimedia site that is harassed by Tor vandalism. If the bot is approved, then the need for coding the extension is reduced, so other wikis would be affected by this. Tito xd^(?!?) 22:26, 5 October 2006 (UTC)
Oppose for now, as per my original neutral I'm not so concerned about bots having admin status, but it should be carefully controlled. In this case items such as the dynamic IPs, the option to use the existing mediawiki proxy block and provide the devs a weekly (or whatever) list or the mediawiki extension, the tumbling number of nodes to block (it's actually less than suggested since many already are on the mediawiki proxy list) seem to suggest this hasn't been thouroughly worked through. As I note above we have some very capable people who want to develop bots to solve every problem when there maybe better solutions out there (If the only tool you have is a hammer, every problem looks like a nail sort of stuff). No one seems to have suggested that we have a very current problem with TOR, more so than last month or the month before, so I can't see the great rush. So oppose purely because this seems premature. --pgk 06:06, 6 October 2006 (UTC)
Oppose per User:Xaosflux's concerns - it bothers me that a member of WP:BAG isn't fully on board yet with these proposals. --Mcginnly | Natter 11:52, 6 October 2006 (UTC)
There are a number of them on board now. I myself was not onboard until I had the time, which wasn't until late in this vote. -- RM 15:30, 7 October 2006 (UTC)
And there are also a number who are not on board. --pgk 16:40, 7 October 2006 (UTC)
~~Weak~~ Oppose per Xaosflux and Chacor. At least it has an emergency block function now, so ~~in the event that this bot does gain adminship status (quite likely),~~ it can be kept under control. Comment struck out since that doesn't seem to be the case now. --Core des at (talk) 13:57, 6 October 2006 (UTC)
Oppose - As per whoever above, I don't like the idea of a bot having admin status. I tried reading and understanding the blurb above, but it was too technical - surely the user Tawker can use the Tawker account for any admin-bot task. --Dangherous 15:27, 6 October 2006 (UTC)
See my comments in the neutral section below. There are three reasons (at least) for why this type of bot should have its own account for an admin-bot task: oversight, ease of use, and previous precedent. -- RM 15:30, 7 October 2006 (UTC)
Oppose per Xaosflux and several others above. Jonathunder 15:29, 6 October 2006 (UTC)
(vote changed) Oppose Good idea, but I'm worried about the technical aspects of this (Xaosflux mentions hackability, for instance), and the fact that even after all this debate it seems that many people are still not clear what's going on, so I can't be confident I understand it either. The situation just doesn't quite seem transparent enough for me to be able to support this. --ais523 15:52, 6 October 2006 (U T C)
I'd rather not have bots as administrators, no matter how trustworthy their creators and operators are. Errabee 16:31, 6 October 2006 (UTC)
Oppose per Splash. Tim! 16:52, 6 October 2006 (UTC)
Oppose. Prevents users in the PRC from taking our own advice. ergot 19:00, 6 October 2006 (UTC)
Oppose. I trust Tawker, but I don't think admin bots are a good idea in general. BryanG^(talk) 21:18, 6 October 2006 (UTC)
Oppose: After careful consideration, I am opposing based on the following:
The schema is not developed sufficiently: I do not believe this idea for the bot has been fully fleshed out. While in abstract I believe due consideration has been given to the general concept of what this bot will do, considerable work still needs to be done on the specific requirements of what exactly is to be done. I asked a number of questions above, and the answer I received was basically "approve of the concept and the admin flag and we'll work out the details later" (see [3] second comment). This is a backwards way of approaching this notion. I am not comfortable with this approach.

Considerable discussion generated: following on the above It is apparent from the 12 pages (yes, I checked) of discussion generated above in the General Comments and Discussion sections above that the idea has not been fully discussed and debated on its merits. Bringing this RfA before this discussion was attempted, say at WT:RFA, was premature.

Unclear need: The proponents of this bot have indicated this is a stop gap measure to fill a need until a MediaWiki extension is put in place that is apparently in development already. Yet, the numbers of blocks we are talking about, if given to the admin corps in toto, is rather modest. A few hundreds blocks is nothing. I could wipe that out in <30 minutes by myself. If a list of IPs needing to be blocked was generated periodically and notice of it posted to WP:AN, I think the admins could knock it out very rapidly and this situation can be handled thus until the extension is in place.

Further unclear need: The assumption is we're getting lots of vandals through Tor. Yet, nothing I've read above seems to provide any evidence that this is the case. It is inherently anti-wiki to block access to the project simply because someone might vandalize. I recognize our policy with regards to open proxies, but would like to see evidence that there is massive vandalism coming from this source before applying a 100kg sledgehammer where a chisel might do.

No adequate response to Essjay's comment: Maybe I missed something above, but I've yet to see any real defense against Essjay's comment regarding prior attempts in this vein.

Respectfully submitted with no intent to comment on either Tawker or Werdna, --Durin 23:14, 6 October 2006 (UTC)
Ok, I'm gonna try and avoid flood here and respond on Durin's talk -- Tawker 05:34, 8 October 2006 (UTC)
Oppose. I don't think I have ever opposed an RfA before, but a bot? No. And before my opposition is discounted for lack of rationale, I share misgivings about this in principle, but also its necessity, its potential for misuse and for mission creep, and other concerns expressed by those with a more technical bent above and below. -- ALoan (Talk) 00:27, 7 October 2006 (UTC)
Firm Oppose Not on the idea of a bot blocking tor proxies, but on the person who runs it. Tawker today just deleted the Wikitruth article and salted it completely out of process. A person who abuses their power like that should never have admin powers. I'm willing to assume good faith and consider that Tawker's account only did that because someone hacked itand if so then his bot would have security issues, too. Anomo 01:26, 7 October 2006 (UTC)
Ok, this whole Wikitruth thing is about to pull my hair. I had a msg on IRC w/ someone asking if they thought it was notable and I took a look, did the google test, saw a grand total of 50 hits top one being Wikipedia and figured it failed WEB and hence A7 and hit the button. No AfD notices or deleted notices or anything on talk so boom, I hit the button and moved on. How I had not heard of that kind of site before amazes me, but thats basically what happened there. -- Tawker 05:34, 8 October 2006 (UTC)
This: [4] is the last version of the talk page, before you speedied it. There is a clear box mentioning the previous AfD. Also, salting the page is not what I would expect if this was a an almost accidental, regular deletion. --Cpt. Morgan (Reinoutr) 19:24, 8 October 2006 (UTC)
Oppose per Durin's very well made rational argument. I'm not 100% sure giving a bot admin powers is a good thing. ALKIVAR™File:Radioactive.svg 01:39, 7 October 2006 (UTC)
Oppose unstoppable admin bots - however, if devs gave us an option to prevent all acces to the sites to members of the bot user group, a simple bot could be left to do things that require admin buttons. Zocky | picture popups 01:59, 7 October 2006 (UTC)
Hmmm.... Actually, if the bot is blocked and desysopped, it's pretty much shut out, AFAICT. So the principle is not fatally flawed. However, given the arguments that this bot isn't really needed, and that its function may be provided by the software soon, I think we should sit this one out and not set the precedent of giving adminship to bots at this time. Zocky | picture popups 02:58, 7 October 2006 (UTC)
It is "stoppable", BUT it takes a steward to do it, it's not like stopping an editor. — xaosflux ^Talk 05:01, 7 October 2006 (UTC)
Yeah, that's true. It would be a good idea to wait until devs can enable admins to completely block bots before allowing unattended bots to be given admin access. Zocky | picture popups 20:43, 7 October 2006 (UTC)
I've been assured that the bot will not run unattended, but will be manually run for ten minutes every week. While this takes care of some fear, it makes the case for establishing the precedent even weaker. I suggest running this bot from Tawker's own account. Zocky | picture popups 14:23, 9 October 2006 (UTC)
Oppose There would have to be a compelling case that it is necessary to give a bot admin powers before we should let it happen. This is no more than a convenience. These could be blocked by hand, or scripting could be done to set up the blocking edit but wait for a human to say, "Yes, I choose to make this block." It is not necessary to give a bot admin powers to accomplish Tor blocking. I also generally endorse the bulk of Durin's opposition above. GRBerry 03:18, 7 October 2006 (UTC)
Oppose For once I don't have to worry about hurting the candidate's feelings (unless it's quite a bot), but I really feel uncomfortable with this solution. I'm not sure that we can compare this -bot to any other or draw conclusions from any previous case. I understand that the author has attempted to fix a problem, but, for reasons of precedence (yes, I know about Rambot, but this is a different type of action in the works) as well as the consequences of any bug, any failure, I cannot support a script. Geogre 03:27, 7 October 2006 (UTC)
The rambot was mentioned, I had to chime in. The rambot had more consequences than this bot every will. It flooded RC, caused an article count bug to be exploited, and majorly hosed up the usefulness of "Random Page". It affected a lot of users who got very very upset. This *is* a good example. With respect to the consequences of bugs, I have introduced errors to thousands of articles that required the bot to update them. That's 30,000+ bugs in articles, but it wasn't a big deal, because the bot fixed them. The rambot also caused the article count bug to come to light, so if you look at historical edit count charts, you'll see a spike up and then a spike down when the bug was found and fixed. Problems occur, but they generally are not as severe as people make them out to be. -- RM 15:20, 7 October 2006 (UTC)
Oppose - For reasons already enumerated above, particularly by Splash and Aaron Brenneman. -- Ξxtreme Unction 03:47, 7 October 2006 (UTC)
Oppose I had originally supported the request but it seems as if there are better alternatives (i.e. other scripts that admins can run) hoopydink^{Conas tá tú?} 03:53, 7 October 2006 (UTC)
Oppose Abuse of powers. Quatloo 04:06, 7 October 2006 (UTC)
Oppose per Durin's well-articulated reasons, and Thatcher131's arguments posted on the talk page. I don't think the risks of creating precedent for a sysopped bot on enwiki outweigh the problem, if any, we have with vandals coming through Tor. - Tangot a ngo 05:19, 7 October 2006 (UTC)
Oppose per the above comments, those made by Durin in particular. Silensor 05:37, 7 October 2006 (UTC)
Regretfully oppose per Durin and what some of the WP:BAG members had to say. - Mailer Diablo 05:40, 7 October 2006 (UTC)
The BAG group members are divided. Some support, oppose, and are neutral. I've updated what I had to say in defence of the bot, now that I've had time. I don't know if it makes a difference or not. -- RM 15:20, 7 October 2006 (UTC)
Agreed, WP:BAG has not come to a consensus on this; complex bot requests often go through several rounds of questions and answers, refinement, testing rounds, more refinement, more testing, etc, prior to approvals. On complex bot requests it can be a lenghty process, but it usually works pretty well, delivering useful, non-damaging bots to the community. Given more time in WP:RFBOT many of these issues may have gotten weeded out. — xaosflux ^Talk 16:26, 7 October 2006 (UTC)
I have no problem with reasons (like yours) for opposing because more time could be taken to figure out the details. This is quite reasonable to me. I just have a problem with anti-bot mentality that this would *never* be a good idea (which is obviously not your opinion). All that said, I think it is less critical that the BAG members agree here, given the community wide input is ideal over a few people in BAG. Still, I respect your concerns, and it is clear that many others do too. -- RM 19:01, 7 October 2006 (UTC)
Actually I'm pretty surprised that the WP:BAG members are split on this RfA - it indicates that development of this bot and its plans are premature. Perhaps it may be there wasn't much discussion about it until this RfA, but regardless of outcome I suppose the bot owner now has sufficient issues and feedback to work on. - Mailer Diablo 19:21, 7 October 2006 (UTC)
I'm not sure it should be that big an issue, this is an exceptional situation and I think it's important to note that I don't think any of the objections are on the basis of doubting the individuals involved or the requirement to solve the underlying problem. There is also a chicken and egg scenario, which one comes first? Do all the work on BRFA to have it knocked back straight off at RFA? On the other hand if it got succesfully through RFA what if the requests group wish to reject it? Personally I think it should probably have gone to discussion elsewhere such as WP:VP to thrash out some of the issue first, then here, BRFA should then have been more concerned with some of the technicalities rather than anything substantive (Bot approval isn't and never will be a way to bypass community consensus). --pgk 20:52, 7 October 2006 (UTC)
Yes, this is an interesting RfA because it unprecedented (In terms of a role bot with sysop powers). And RfA is more of guage of a consensus of social acceptance on giving sysop powers, rather than looking at it in its pure technical merit. - Mailer Diablo 05:42, 8 October 2006 (UTC)
Weak oppose. I don't think than adminbots should be strictly prohibited, but this particular bot will flood logs an Special:Ipblocklist, and also I think that such measures against proxies adn Tor should be taken at the MediaWiki level, prohibiting IPs from editing every Wikimedia project. MaxSem 12:53, 7 October 2006 (UTC)
There is perhaps a flaw in your logic: These blocks should be done whether or not the a bot is doing them. The problem right now is that there is too much to be done for humans to do, but they should be done anyway. No one is saying the task is a bad one, which implies group consensus that this should be done. So whether a bot floods the log or a user does so there is no logical difference. A bot can be programmed to run over long periods of time to spread out the log messages, satisfying your concern. -- RM 15:07, 7 October 2006 (UTC)
On the contrary though: the discussion up top indicates that, once a month, about 200-300 blocks need to be made. Brenneman has volunteered to do this, and I'll help him out if you like. I can handle 100-150 blocks on one day a month with barely batting an eyelid. Could have 'em all done in about...15 minutes maybe. -Splash - tk 15:31, 7 October 2006 (UTC)
Ok, if humans could do it and have such a desire, then that's fine with me. I still don't have a problem with a bot saving people the time (there are plenty of other useful tasks). Still I'm not sure how your point addresses the issue of the block log. While a bot could spread out the changes, it won't apparently do so (I don't see the problem if it does spread them out). But this is still no different than humans doing it. To oppose on the basis that it will flood the log just doesn't make sense. Whether humans or bots do it, the logs still are there. -- RM 15:58, 7 October 2006 (UTC)
Yes, these blocks should be done anyway. But performing them with bot serves as a ersatz of a real solution - centralized blacklist. Running this bot makes development of the blacklist less crucial task. MaxSem 16:14, 7 October 2006 (UTC)
This seems like some sort of a slippery slope argument to me, but I do understand your point. I don't think anyone disagrees that a WikiMedia implementation of this would be ideal. -- RM 19:01, 7 October 2006 (UTC)
Probably you're right, but I can't say I like this implementation. MaxSem 07:37, 8 October 2006 (UTC)
Oppose I trust Tawker and Werdna here, without a doubt, however I just cannot imagine a bot having a mob and bucket. Things could easily go wrong instantly. Humans make mistakes, but not on the scale of bots. // Pilotguy (Have your say) 13:06, 7 October 2006 (UTC)
Oppose. Adminship should be limited to humans. The need for making an exception in this case has not been demonstrated. I also agree with users who have pointed out that this solution is a kludge that needs to be developed further before it could even be given consideration. Good intentions, yes, but poor execution. SuperMachine 14:59, 7 October 2006 (UTC)
Oppose per Durin's analysis and several others above. 250 blocks a month does not require a bot to handle. Provide the IP list on a centralised page and >1000 admins watching, problem solved. --Cactus.man ✍ 15:26, 7 October 2006 (UTC)
Oppose. We're not voting on whether or not to give adminship to a bot; we're voting on whether or not to give adminship to a human being who says he wants to use that adminship to run a bot to do job X. The problem is, once he's got the adminship, he can then use it to run any sort of bot he wants. Until and unless some agreement is worked out whereby this account's admin bit would come up for renewal on a regular basis (once every 4-6 months?) or every time any major change is made to the bot's programming (per accusation of, well, just about anyone), I can't support this. Nothing personal. --Aaron 18:59, 7 October 2006 (UTC)
Tawker is already an admin, so we are not voting on a user becoming an admin. If that were the case, he could just run it from his main user account. No, the purpose here is twofold: 1) Seek an admin bot flag to do this task, and 2) To get community input for the task itself. The Bot Approvals Group members (who are divided) have already pretty much stated that this task will not change without another RfA, and this has been firmly demanded by much of the community anyway. I don't know what other assurance you'd like: The task can't change without approval and the user is already an admin. Also, as soon as the feature is implemented in the MediaWiki software, the admin priviledge can and should be removed. -- RM 19:07, 7 October 2006 (UTC)
Oppose, per Sjakkalle, Durin, and [5]. Although I admire both werdna and tawker, I don't think that this is such a great idea. In fact, I think that it should have gone through bot approval before coming here, so that we could see a more fleshed out, final version before !voting. --Storkk 20:11, 7 October 2006 (UTC)
Oppose, sort of Ehhhh, Bot Approvals Group's nervousness makes me want this to be better planned out and accurate. The fact that the emergency stop mechanism is not well formed until just now makes me wonder. Why not just have the bot maintain a large list of Special:Blockip links of IPs which are foul? Also, violating other websites robots.txt file, which this one specifies no access, makes me not feel so good. We'd ask the same that robots not violate ours. The theory I like. But this one needs some more planning and better adjustments. I would suggest that if it ends up getting implemented, the blocks link to a detailed description page of "Why your IP is blocked for tor". Kevin_b_er 22:14, 7 October 2006 (UTC)
I think the robots.txt stuff is a bit of a red herring. The servers being talked to are actually tor servers, and we are requesting the directory. If you look at the tor source code in directory.c is the code to always return that for robots.txt, yet at the same time config.c contains a few servers for the client software to connect to as diectorys and get a list of nodes. i.e. We are doing exactly what the tor client itself does and ignoring robots.txt --pgk 22:48, 7 October 2006 (UTC)
That's true enough until you reverse the situation. Our bots and our scripts ignore our robots.txt all the time. We still don't want other people's bots and scripts to ignore it. Zocky | picture popups 00:02, 8 October 2006 (UTC)
OK, but the point was that this is very much expected to be read by computer, not human. Regarding the desires of the players here, then I'm sure that the suppliers of Tor also don't want us to blocked them from editing, should we consider that also? Simiarly we don't want tor being used to vandalise us, "they" seem quite happy to ignore that desire... --pgk 08:29, 8 October 2006 (UTC)
Oppose Too much power for one user, this is really just an rfa for a "super-tawker" it seems. Could be the beginning of a "sysop powered robot" arms buildup. Yankee Rajput 23:40, 7 October 2006 (UTC)
o_0 - David Gerard 02:12, 8 October 2006 (UTC)
Oppose Because I don't want someone with 2 sysop accounts, even if bot, or semi bot or whatever —Argentino (talk/cont.) 01:48, 8 October 2006 (UTC)
Bots are not considered sock puppets. -→Buchanan-Hermit ™/?! 02:06, 8 October 2006 (UTC)

I wrote that line to be descriptive, not prescriptive - David Gerard 02:12, 8 October 2006 (UTC)
Oppose, bots should not have admin status. Note that my oppose vote has nothing to do with the individuals involved, whom I have total faith in personally. User:Zoe|(talk) 03:38, 8 October 2006 (UTC)
Oppose, I agree that a bot should not be an admin. I do also want to remind some of those opposing to assume good faith; I see no reason to think Tawker and Werdna are acting in bad faith. Heimstern Läufer 06:16, 8 October 2006 (UTC)
Oppose I for one, do not welcome robot overlords. I do see a benefit of the bot, however I do not think that sysop powers should be given to things that do not have discretion.--Toffile 06:55, 8 October 2006 (UTC)
1. Admins are not overlords. 2. From the endless mention to sci-fi, I think a lot of the "oppose" voters here misunderstand the whole concept of "bots". As far as I can see, this bot is hardly a "robot". If you are concerned about "robots" Tawker's User:Tawkerbot4 and its clone User:AntiVandalBot clones do a much more perilous task regularly and with less supervision - reverting mainspace articles automatically using, as I understand it, a set of rules learned using machine learning, which is a branch of the feared AI. And somehow I don't hear too many people complaining about having robots "controlling our articles". From the description above, this bot is a puppy compared to these "dangers" posed by the anti vandal bots. As far as I understand TawkerbotTorA doesn't even have this "AI" thing that everyone is afraid of (for no good reason), and it would be operated manually by two trusted Wikipedians, where is the rogue robot overlord in that?--Konst.able 07:13, 8 October 2006 (UTC)
Please don't assume that everybody voting against is an idiot, even if they make a goofy introduction to their vote. This bot will not be run manually AFAICT, and admin accounts can't be stopped without a steward, even if they're bots. It's a rational position to be against allowing unattended bots to run on admin accounts, especially if they do things that could be manually run as scripts by an admin. Zocky | picture popups 12:30, 8 October 2006 (UTC)

Zocky nailed just about every reason why I'm oppposed. The actions are not easilly stoppable, there is the shutoff code, but in the case that it fails, someone has to get one of the Stewards. In addition the Bot approval for TawkerBotTorA does list this as being an automatic bot, so this discrepancy in the requests unnerves me a little. As I said, I can see use for the bot, however the other methods have not been fully explored in order to deal wiht this problem sufficiently to convince me that this is the only solution. --Toffile 18:08, 8 October 2006 (UTC)
Sorry for misinterpreting you Toffile, but there are a couple of other users who have been mentioning sci-fi. Neither was I referring to all opposers in general, Zocky, I do understand that there are valid resons for opposing.--Konst.able 03:04, 9 October 2006 (UTC)
Oppose, after careful consideration and with no ill-will towards Tawker and Werdna. Basically, this feels like a solution in search of a problem. I absolutely agree with brenneman's point on software development: if this is a serious enough problem to be automated, then the correct solution is to build in TOR filtering to the MediaWiki software, not to run an ad-hoc admin bot with too big of a hat from IRC. I'm not so worried about the slippery slope arguments, but I do think it is good to uphold the principle that adminship is a role with responsibilities as well as rights. From that perspective, limiting it to actively human-managed, non-scripted activities, carries certain benefits with it. I think this bot (or any bot) is incompatible with that philosophy. Nandesuka 13:03, 8 October 2006 (UTC)
Oppose, per Durin's comments above - I might support this idea if it was possible to give the bot blocking powers only, but I'm a bit uncomfortable with the idea of a bot having fulll sysop access. Plus, the whole thing seems a bit like a sulution in search of a problem - as others have pointed out above, performing about 600 blocks once a month can easily be done manually by a couple of admins. -- Ferkelparade π 14:32, 8 October 2006 (UTC)
Oppose on principle. -- No Guru 15:42, 8 October 2006 (UTC)
Oppose, as other options appear to be equal or better solutions (adding the IPs to the regular proxy block list, manual blocking, or running the script from the main account). --Cpt. Morgan (Reinoutr) 19:26, 8 October 2006 (UTC)
Oppose, bad idea. Accountability problem. Duke of Duchess Street 20:20, 8 October 2006 (UTC)
Oppose per many above. This bot doesn't seem to have had a very thorough "once over" by the community prior to this RfA, and frankly the potential problems assosciated with its use are to great to justify giving admin rights. Markovich292 21:43, 8 October 2006 (UTC)
strong oppose: Already have seen too many blindbot accidents. Even if bots could be programmed adequately 'to do no harm', there is no reason to further dehumanize an outpost (the Wiki) of nominal sanity in the weird wide world of internet spam. Is kinda reminiscent of black box voting, to boot. Ombudsman 02:41, 9 October 2006 (UTC)
At least, until someone with a clue that is also an admin steps up to run and maintain it. More than a few things made me think WTF? looking at the code and it was easy to make up test cases it would fail. The list of tor exit nodes does not change that much any given month that this really needs a bot to take care of (graph). "In all seriousness, though, there are just too many blocks for one human to do them all." yeah right, several admins regularly pull in 1000+ blocks a month... and this is a wiki, one person doesn't have to do them all. Kotepho 03:24, 9 October 2006 (UTC)
Please note, that the proposed bot operator is already an admin. — xaosflux ^Talk 04:28, 9 October 2006 (UTC)
I don't have a clue? Last time I checked I have the sysop flag around here. This isn't a human rfa, it's a proposal to sysop a bot. Now, I'm taking a look at the backlog on WP:OP and hence the bot proposal... some things just seem to be in backlog forever -- Tawker 06:34, 10 October 2006 (UTC)
Oppose. I have no problem with Tawker himself (in fact, I'm a friend of his), but I'm just not comfortable with a bot being an admin. What if there's a mistake and it fucks up the site? Can't let that happen. (I really don't care if it's "idiotproof", accidents do happen.) 1ne 16:36, 9 October 2006 (UTC)
Oppose. I don't think it's a good idea to give that much power to a bot, and I worry about "mission creep". -Hit bull, win steak^(Moo!) 17:41, 9 October 2006 (UTC)
Oppose I thought long about this and while giving this bot "the bit" would definitly make the lives easier of those that vigilantly block out open proxies - including Tor exit nodes - I have four major concerns. Firstly, as said above while the bot would lighten the load I do not think this is the best solution for the Tor/open proxies issue and using it would reduce the need to look for a better one. Secondly, I kinda fear mission creep - once a it got "the bit" its tasks might be expanded - in best faith - above and beyond the blocking of Tor exits. Thirdly I fear precedent - so far only humans have gotten "the bit" and giving it to this bot would create a precedent so that other bots - with task that has more danger for unintentional sideffects - would have it easier to gain it. Lastly I am not entirely comfortable with giving admin powers to even a manually started bot - the possibility that something might go wrong after all is always present. Charon X/talk 21:36, 9 October 2006 (UTC)
Oppose per Durin and the great firewall--Dakota 00:30, 10 October 2006 (UTC)
Oppose until and unless it is recoded to be stoppable quickly by any admin. While the circumstances this bot could run amok in are quite limited it sets a bad precedent to have an unstoppable bot. Eluchil404 03:34, 10 October 2006 (UTC)
Strong Oppose. Bots are fundementally unsuited to perform certain tasks; Namely anything requiring JUDGEMENT, an understanding of CONTEXT or EDITORIAL DECISIONS, should not be done by bots. The adminship involves all three. Also giving the mop to a bot is essentially, allowing Tawker to have a second account with sysop powers. This sets a dangerous precedent. Besides, we already have too many non-writing admins, who don't really understand how to work with the actual editors and are thus seen by us as lacking the moral authority to police a community of writers. Adding a Robocop to the mix will only make matters worse and further erode respect for the admin corps. Oh, also OPPOSE as per WP:BULL--R.D.H. (Ghost In The Machine) 10:34, 10 October 2006 (UTC)
Query: Since the bot would just be processing TOR lists into blocks, what judgement is required? - CHAIRBOY (☎) 15:26, 11 October 2006 (UTC)
Oppose. While the writers of this this bot etc. are certainly well-meaning and have the best interests of the encyclopedia at heart, I still oppose on principle and for the same reasons as most above. Miss Madeline | Talk to Madeline 01:31, 11 October 2006 (UTC)
Oppose. Wikipedia is the free encyclopedia that anyone can edit. This includes anonymous users. Bsmntbombdood 15:56, 11 October 2006 (UTC)
Comment. This bot is not designed to ban anonymous users, it's designed to block Tor nodes. Open proxy use has always been bannable, this would merely automate a small portion of the process. --tjstrf 16:07, 11 October 2006 (UTC)
Strongest oppose. Wikipedia relies on admins being accountable for their actions, which a bot obviously wouldn't be. Are we goin g to block the bot's creator every time the bot violates blocking policy? Until we are, this is a very bad idea. Also per Essjay's evidence of feature creep. Cynical 16:05, 11 October 2006 (UTC)
As the bot would only be blocking the entries on the TOR list, what policy violations do you anticipate? BTW, the bot's creator and maintainer is obviously accountable for its actions, no less so than the driver of a car is. - CHAIRBOY (☎) 16:23, 11 October 2006 (UTC)

My point is that a human admin has an incentive not to violate policy - ie that his personal account will be blocked (even temporarily) and his reputation on Wikipedia will suffer. A bot does not. Also see Essjay's comment (which I linked to) about function creep - Tawker has previously given assurances (according to Essjay's comment) that his bots would not have admin functions. Therefore if he is willing to break those assurances, why wouldn't he break the assurances that it will only block entries on the TOR list? Cynical 18:00, 11 October 2006 (UTC)
This issue has been address elsewhere in this discussion, but the point is that Tawker runs his own bot, and it will not have admin priveledges according to his own assurances. This is a different bot, and his comments were not related to that purpose. Because of this, he brings an RfA for a new bot, requiring full community approval. This isn't a broken promise for his bot. It's a totally separate and new request. For instance, would it make a difference if I ran the bot instead because I never gave such assurances? The reference to Essjay is taken out of context and frankly shouldn't be used since it has nothing to do with this particular RfA and bot. Under the reasoning given, the only reason to oppose is if you don't believe Tawker to be a decent enough bot operator to have an admin bot, which is an acceptable position. But if you believe that he *is* decent enough and trustworthy, then it makes no sense to oppose because he made assurances that a different bot would not be given admin access. -- RM 18:35, 11 October 2006 (UTC)
My point was that, from my impression of Essjay's comment, it appears that Tawker previously made promises about a different bot which he then failed to keep. On that basis, since Tawker appears to have broken a promise about a different bot, there is no reason to believe promises he makes about this bot. Cynical 19:15, 11 October 2006 (UTC)

Update' Tawker's message on my talkpage only serves to convince me of my position. I've quoted the final sentence only for reasons of shortness. 'They are totally unrelated technically, TB2 is in python, TBTA in C#. I just thought I'd try and set the record straight.... you seemed a tad facts confused'. Charming. Cynical 19:21, 11 October 2006 (UTC)
Oppose, get bot approval for the main account instead like User:Fvw did. JYolkowski // talk 01:41, 12 October 2006 (UTC)

Neutral

I have issues with making this go through an RfA. We generally have been very worried about giving bots adminship status and I would oppose on that ground alone were it not endorsed by Jimbo. However, if Jimbo endorses it, it isn't then clear to me why we need to go through this RfA. I'd also like to add that even as someone who opposed Werdna's last RfA, I don't see why that doesn't mean we can't trust Werdna with the password to the bot. The main reason for opposition was incivility. I don't think anyone seriously would think that Werdna would abuse the bot's admin access. JoshuaZ 00:20, 5 October 2006 (UTC)
The bot will only run the given code. Any revisions to the code will be tested thoroughly on my own sandbox wiki. I hope this serves to allay your concerns. — Werdna talk criticism 00:24, 5 October 2006 (UTC)

Jimbo is showing constraint. He wants the community to decide this sort of thing, even though he approves of it. He has more than enough influence anyway, just from stamping it with approval. This isn't a life or death issue for Wikipedia, so there is no reason to directly intervene. -- RM 15:35, 7 October 2006 (UTC)
I object to the bot running automated until its actions have been tested on a large scale. Testing on a local wiki is not enough to iron out all bugs. If the bot only updates once a week, it would be feasible to human-confirm a list of pending (un)blocks before performing them. —Quarl ^(talk) 2006-10-05 08:07Z
Who will be the one checking the block/unblock list against the list of approximately 1000 Tor servers, a large percentage exit nodes? — Werdna talk criticism 10:13, 5 October 2006 (UTC)
There may be on average 1000 exit nodes, but what is the weekly turnover? If it's 50, and some admin just has to click once per exit node, we have more than enough administrators to do it. —Quarl ^(talk) 2006-10-05 21:14Z

I don't have an issue with bots doing admin tasks if those bots are very well defined in their actions (and tightly constrained to them), though personally I'd prefer it to be run under the users own admin account to keep accountability focused (In this case doing so is something excluded by Tawkers RFA promises). I do however think the issue of dynamic IPs and no longer TOR nodes needs to be resolved before this should be moved forward. --pgk 12:33, 5 October 2006 (UTC) switched to oppose --pgk 06:06, 6 October 2006 (UTC)

Switched to neutral from provisional support, leaning oppose. I've not been convinced about a bot having admin tools, and Splash does bring up a very good point about Tawker and his bots. I'd much rather see this bot operated by more than just Tawker. – Ch acor 13:05, 5 October 2006 (UTC) Switched, again, based on a review of some archives. – Ch acor 13:26, 5 October 2006 (UTC)

~~Per Chacor, Xoloz. I need to think over more carefully. - Mailer Diablo 15:11, 5 October 2006 (UTC)~~ Changed to oppose. - Mailer Diablo 05:40, 7 October 2006 (UTC)
I don't know enough about bots to have a fully informed opinion, but Essjay's negative view of bots with sysop rights has a strong influence on me. Since the hard work is parsing the tor list, why not make a non-admin bot that will parse the tor list and update a page such as WP:OPP/TawkerbotTorA with the proxies formatted with the Template:Tl template. Then any admin could go in and easily knock them down. Thatcher131 15:34, 5 October 2006 (UTC)
Good idea. You just got your sysop bit, feel free to make the six hundred blocks. In all seriousness, though, there are just too many blocks for one human to do them all. — Werdna talk criticism 16:46, 5 October 2006 (UTC)
- How many blocks are we actually talking about? Keep in mind; we have far more than one admin willing to do this kind of work. --Durin 20:35, 5 October 2006 (UTC)
Neutral (moved from Conditional Support) I would've preferred a Jimbo-authorized one-week sysop bit before this RFA to see how it works in action. We're essentially asked to approve an editor with zero contributions. ~ trialsanderrors 04:40, 5 October 2006 (UTC)
According to Werdna, this bot has been testet on his seperate Wiki for three weeks. Is there any reason to doubt that this is enough? D e lta Tango | Talk 15:18, 6 October 2006 (UTC)
A private test wiki does not have the load of en: in terms of the visitors we get daily; false positive blocks would likely be non-existant, as those ip anons wouldn't be trying to edit there. — xaosflux ^Talk 05:10, 7 October 2006 (UTC)
Neutral - I don't understand. Why do you need a separate account under which to run a script? Wouldn't it be more reasonable to run this under your own user account? This doesn't look like a bot per se, but rather, just a way to save some time on running a highly repetitive task. I didn't think anyone had a problem with running manual scripts under your own user account. Obviously, if this was an autonomously running bot, then that would be a different issue. joshbuddy, talk 18:30, 5 October 2006 (UTC)
I feel strange telling another BAG member this. But there *are* a number of good reasons to do this from a bot account. First is oversight by other users. A bot is more visible for the task it is performing. Secondly, emergency shutoff is easier to accomplish without interfering with a normal user account. Third, there is precedent for persons doing very specific major tasks to have their own account, such as User:Dannyisme. -- RM 15:13, 7 October 2006 (UTC)
Comparing the function of this bot to User:Dannyisme seems strange, the reason as I understand it was to differntiate office actions i.e. official capacity, blocking tor is not comparable. Personally I think the reasons you state for it being a separate account are in fact reasons I would perfer them to use their own account. Oversight isn't an issue, if the bot is blocking properly the distinction between will be clear. As for not interfering with a normal user account, to me for admin actions I'd prefer that sort of accountability, the bot screws up and it's your account which loses it's priveleges while it's sorted out. --pgk 16:45, 7 October 2006 (UTC)
I didn't mean to say that the two examples were identical, because they are not. It is just that separating the actions of the bot from the normal user account has specific advantages, and there is precendent for separating certain actions from a main user account to ensure that those actions are isolated. Obviously the bot actions are not directly comparable to office actions, and it was not my intention to imply that they were. Anyway, you do make a good point about the accountability aspect. As for my personal opinion, I prefer a separate bot account, but wouldn't have a problem either way. -- RM 18:54, 7 October 2006 (UTC)
Neutral per JoshuaZ. Also, I trust Tawker completely, just don't want a bot admin. Grand Slam 7 00:58, 8 October 2006 (UTC)
Neutral. Well, bots are really annoying sometimes, but they do make life easier regardless. I'm sure this admin bot will piss some people off, even though it may do more good than harm. File:Aqu01rius.gif (user | talk | websites) 01:38, 8 October 2006 (UTC)
Neutral. Blocking a series of TOR proxies is monkey work that does not require much "human intelligence." Nonetheless, even monkeys are smarter than bots, and a bot without an off switch and with the ability to be quite disruptive if it goes haywire frightens me. Yes, the bots been mildly tested on many sandboxes; however, it has not been tested on any wikis as large in scale as Wikipedia and I worry about certain unexpected things that may occur during the bot's run that could cause it to start misbehaving. I could support this under a few guarantess: 1) It is only run under Tawker's very close supervision - if he gets up to go to the bathroom, he stops the bot. 2) There needs to be a steward on guard, readily available and contactable (on IRC or by phone) at all times that the bot is online to desysop it at the first indication that the bot is acting strangely. (I would ask that this steward's contact info be posted on the bot's userpage). 3) The bot needs to be programmed to stop blocking anytime it receives a message on its talk page or is blocked (if it already does this, I apologize - I've not looked over the code). 4) The bot should not unblock or modify the blocks of any ips on its list - it can leave a note somewhere that one on its hitlist is blocked for only 48 hours, or blocked with account creation enabled, but not modify it. There may be perfectly legitimate reasons that some proxies are blocked differently than policy dictates. Anyway, I think this bot may well serve a very useful purpose and therefore I'm trying to lay my anti-adminbot prejudices to rest; however, I cannot support the bot unless these conditions are met. AmiDaniel (talk) 05:13, 9 October 2006 (UTC)
- It's been suggested (somewhere in this page) that if Tawker is going to ride along everytime this bot runs, why does it need its own account? Have the script parsed into a ready-to-go page (as done already by someone further up the page) and have Tawker use some AWB analog to do the blocks from his own account. - brenneman ^{L} 05:35, 9 October 2006 (UTC)
Neutral until somebody comes up with a decent off switch for this bot. -- Run e Welsh | ταλκ 08:35, 10 October 2006 (UTC)
Neutral - bots are not for admin tasks, though I may give support with more preparation. --Gray Porpoise^{Phocoenidae, not Delphinidae} 10:57, 12 October 2006 (UTC)

The above adminship discussion is preserved as an archive of the discussion. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the talk page of either this nomination or the nominated user). No further edits should be made to this page.

Wiki143:Requests for adminship/TawkerbotTorA

Contents

TawkerbotTorA

General comments

Discussion (for expressing views without numbering)

Support

Oppose

Neutral

Navigation menu

Wiki143:Requests for adminship/TawkerbotTorA

TawkerbotTorA

General comments

Discussion (for expressing views without numbering)

Support

Oppose

Neutral

Navigation menu

Search