mic_none

Wikipedia:Bots/Requests for approval/FA Template Protection Bot Source: en.wikipedia.org/wiki/Wikipedia:Bots/Requests_for_approval/FA_Template_Protection_Bot


The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section. The result of the discussion was Approved.

FA Template Protection Bot

[edit]

Operator: Chris

Automatic or Manually Assisted: Auto

Programming Language(s): PHP using Cobi's classes

Function Summary: Adds all the templates on today's and tommorow's fa to Wikipedia:Main Page/Protection which is cascade protected

Edit period(s) (e.g. Continuous, daily, one time run): on the hour

Already has a bot flag (Y/N): N

Function Details: The bot works like so:

  1. Parses the name of today's fa from here (with some error checking)
  2. Get's all the templates on the fa and removes those listed here
  3. Formats the templates into a nice list
  4. Places the formated text on Wikipedia:Main Page/Protection

The bot will only run if User:Chris G/FABot Run is set to "Run" (cases insenstive)

Discussion

[edit]

Since this is a request for an admin bot I have transcluded this brfa onto Wikipedia:AN so we can have wider community input. Per disscusion at Wikipedia:BN and the Adminbot rfc I am testing to see if this brfa alone can gather enough consensus for a 'crat to be willing to +sysop the bot. Also I would like to make it clear that I am and have been for some time running this bot on my main account. The bot's source code can been seen here --Chris 04:42, 16 August 2008 (UTC)[reply]

From an management side of things it's editing a protected page, big stinking deal, it's almost an instaapprove. It's the community acceptance of a +sysop flagged bot which will raise hell.

Can't exactly trial it without the flag... its a community acceptance thing, on a personal note, I have no problems with it -- Tawker (talk) 05:01, 16 August 2008 (UTC)[reply]

We could do something like have it mark it's edits with a unique flag (or link to this BRFA) in the edit summary, and, have Chris run it under his account, I can easily filter those edits via an SQL query, and, list them on a subpage of here for review. SQLQuery me! 05:13, 16 August 2008 (UTC)[reply]

On a sidenote, why not put the images used in it on the protection page as well? SQLQuery me! 05:15, 16 August 2008 (UTC)[reply]

Because then we'd have to rename the bot :P. If there aren't any objections to protecting images as well it should be a fairly trivial change to the code to get it done --Chris 05:34, 16 August 2008 (UTC)[reply]

Would the bot remove the templates from the protected page the following day? WODUP 06:11, 16 August 2008 (UTC)[reply]

Yes, it removes the old templates and adds the new ones --Chris 06:13, 16 August 2008 (UTC)[reply]
I figured as much, so far, to me, this sounds good. SQLQuery me! 06:23, 16 August 2008 (UTC)[reply]
Okay, sounds good. :) WODUP 06:25, 16 August 2008 (UTC)[reply]

I've updated the source so that it adds images as well --Chris 06:42, 16 August 2008 (UTC)[reply]

Question: Will this have to go through RfA like User:RedirectCleanupBot did? I can't seem to find any consensus on Wikipedia:BN. NuclearWarfare contact meMy work 19:40, 16 August 2008 (UTC)[reply]

Copying Iridescent's post from Wikipedia talk:RFA to this BRFA. WODUP 22:43, 16 August 2008 (UTC)[reply]

Support, but a beansy thought that someone will need to watch for; presumably if it's run every 10 minutes, a vandal can vandalize a template, add it to the article after vandalizing, and the bot would then lock the template into the vandalized state. Not a major problem but it will happen. Also, will it unprotect all the templates once the article's delisted, or only those that were unprotected to start with?

— iridescent 16:00, 16 August 2008 (UTC), Wikipedia talk:Requests for adminship#Heads Up
Well I thought of that too -= but presumably they can do that already, and while the vandalism will show either way, it is easier to track which edit caused it (as against adding a useful template and then vandalising - which is a bit more difficult) ViridaeTalk 22:58, 16 August 2008 (UTC)[reply]
To really deal with Xenocidic's concern it would have to constantly monitor recentchanges via IRC, even checking 5 minutes is more than enough time for someone to add a template and vandalize it. I don't think that feature will really be helpful, though that is a concern that should be dealt with, not sure how though. Mr.Z-man 23:17, 16 August 2008 (UTC)[reply]
The sort of timed attack Iridescent describes is a major concern, and it's not just in the realm of Wikipedia:BEANS, but one that's already been attempted at least twice to defeat my own FA protection adminbot. The number of possible attack vectors are increased here since we're dealing an almost unlimited number of templates that could be vandalized, or even new ones created for the purposes of mischief. Chris, you might be able to pick up the solution here. :) east718 // talk // email // 23:43, 16 August 2008 (UTC)[reply]
That method isnt going to go away - it can happen already and it isnt going to be made worse by this bot. As it stands, vandalism to the FAOTD by this method (adding a template that is already vandalised) is easiyer to deal with than adding an innocuous template then vandalising it. Having the runs more often pretty much gets rid of the latter problem and nothing will top the former except protection/eyes on the page. ViridaeTalk 06:41, 17 August 2008 (UTC)[reply]
OK, we're not thinking on the same plane here. I don't care if somebody adds a template to the article later and vandalizes it; that is expected. The real problem is that the robot might lock a template in a vandalized state if a vandal throws it onto the article just before a check cycle. The optimal solution (that's obvious to me at least) is to just not protect a template if it's been recently edited and to report it for human review somewhere. I happen to prefer throwing up the !admin flag in #wikipedia-en-admins, Chris may opt for something else. east718 // talk // email // 01:44, 18 August 2008 (UTC)[reply]

I don't think the timing problem can be solved. No matter what time it is set to run, unless it is set to run continuously, there will be a break where mischief can occur. Since we started this in July, vandalism during the break has only happened once. In that case someone added a template that should have been in the article already and then vandalized it. In response we lowered the duration been protections. Since then I don't recall anymore problems. As for images, we had a discussion about them when we first created the page and decided not to add them unless they became a target. So far they have not. We were primarily concerned with fair use images, but we also asked what protection of an image would really do, and decided not much. Having worked on the protection page and tweaking with Chris, the bot has my support. KnightLago (talk) 15:25, 17 August 2008 (UTC)[reply]

  • On the timing problem: Featured articles should be fairly stable. They shouldn't receive major template overhauls right before they go on the main page. If this attack vector is an issue, only lock down templates which existed in the article when scheduled, or some days before it goes on the main page, and make the period vary between checks so the lockdown schedule isn't predictable. Gimmetrow 04:37, 18 August 2008 (UTC)[reply]

I recommend this bot goes to RFA, I say this as the programmer of the original redirect cleanup bot. (check the rename logs for my old username). As far as operation, the idea that east said above is a worthwhile one. If you are doing checks every 10 minutes, then check to see if the template has recently been edited (in last 10 minutes), and if it has, bring the template up for human review somewhere. Have it say something in #wikipedia-en or the like, or post somewhere where people don't mind watching. The only real way out of this is to monitor on IRC. —— nixeagle 03:36, 18 August 2008 (UTC)[reply]

I'm way out of date, strike the RFA suggestion per adminbots RFC. However rest of advice still holds. —— nixeagle 05:18, 20 August 2008 (UTC)[reply]
The proposed policy is to not go to RfA per the adminbots RFC. BJTalk 03:40, 18 August 2008 (UTC)[reply]

Ok, sorry about the lateness of my reply but I've been having a few computer problems. A few points have been raised that I would like to answer:

  1. RFA - at this point in time I have no plans to run this bot through an rfa
  2. Images - KnightLago has made some vaild points so unless image vandalsim picks up the bot will not protect images
  3. Templates - I think Nixeagle has made a good sugestion about checking when the template was last edited and will work on implmenting it. Gimmetrow also made a nice suggestion about varing the running time but unfortunately the bot runs on a crontab which is hard to randomise --Chris 08:36, 18 August 2008 (UTC)[reply]

Of course, the real solution to the problem is a better way of monitoring changes to an article's templates. — Werdna • talk 08:54, 18 August 2008 (UTC)[reply]

Huggle now watches templates too, so any sneaky vandals should be reverted quickly (if someone is using Huggle at that time...). Maybe we should make ClueBot watch the templatespace too...Soxπed93(blag) 05:36, 19 August 2008 (UTC)[reply]

I've updated the code so that the bot checks if the template was edited in the last 10 minutes if it was the bot will not protect the template and it will send off an alert in #wikipedia-en-alerts. Now I have two questions:

  1. How often should the bot run (every 5 minutes...)?
  2. What should it be renamed to (I was thinking along the lines of "FA Protection Bot" or something like that) --Chris 10:38, 20 August 2008 (UTC)[reply]
Just thought I'd chime in here, guys, and say that I'm not a big fan of bots having admin rights, as they don't have the ability to make judgement calls. There aren't really any calls to make in this particular case, but I'm still not a fan of the idea. Useight (talk) 14:51, 20 August 2008 (UTC)[reply]
Which would you preffer?
  1. Hard to revert template vandlsim on the main page FA
  2. An unapproved admin bot that has not gone under any sort of review protecting the templates before the article goes live
  3. An approved admin bot that has been reviewed by the community and coders for any policy/technical flaws protecting the templates

I understand your hesitation about giving a bot +sysop, but RedirectCleanupBot has proven that we won't have a SkyNet style incident. In this case I think a bot is much more accurate than a human and does a better job at protecting the templates. Also could you please explain why you don't like admin bots? The only reason you gave you dismised in the next sentence --Chris 09:25, 21 August 2008 (UTC)[reply]

I don't really like any of the options you gave me above, so I'll go with option #4, an admin manually protects the templates before the article goes live. As for my explanation as to why I don't like admin bots, I already said it above, they can't make judgement calls. Just because I dismissed it in this case, doesn't dismiss it as my reason for not wanting admin bots. And the RedirectCleanupBot doesn't "prove" we won't have a "SkyNet style incident", one is too small of a sample size to prove anything. Useight (talk) 16:19, 21 August 2008 (UTC)[reply]
Today's TFA has 40 templates and 5 images on it. While many of the templates were probably already indef-protected, several would probably still need to be done. Twice a day (protect/unprotect), every day (actually, the bot might check 100+ times per day). Are you volunteering? To me, this objection doesn't make any sense. This follows the natural order of things. People realize this needs to be done, so they start doing it. They then realize its just mindless repetitive actions, so they make a quick user script to speed things up. The script eventually develops to the point where all they have to do is click the button. Then they realize that with a bot, they could click the button automatically. I'm not certain if the middle steps happened like that, though its likely (Twinkle has a feature that would really help with that)
And you're right, RedirectCleanupBot doesn't "prove" we won't have a "SkyNet style incident," the actual facts behind how simple computer programs like this work proves we won't. Mr.Z-man 21:02, 21 August 2008 (UTC)[reply]

SQL Query for later: 

SELECT CONCAT('* [http://en.wikipedia.org/w/index.php?diff=prev&oldid=',rev_id,']') FROM revision WHERE rev_comment LIKE 'GymCR47%' AND rev_user = 2091313;

Well, after reading this whole thing over a couple times, and, then giving it another day, I believe that this bot is ready for a trial run, and, that there is consensus for this task to run. For the time being, I would like to restrict this bot to templates only (no images). Being an adminbot, and, this being unusual circumstances, I am going to ask for a thirty day trial, under your administrator account, linking to this BRFA ( Wikipedia:Bots/Requests for approval/FA Template Protection Bot in hopes that more comments may be generated ) adding the string "GymCR47" at the beginning of the edit summary so that this processes edits may be more easily identified. With your modifications, I do not believe the timing attack mentioned above will be an issue. Please consider pinging #wikipedia-en-alerts on that condition, on IRC. Approved for trial (30 days). Please provide a link to the relevant contributions and/or diffs when the trial is complete. SQLQuery me! 05:37, 22 August 2008 (UTC)[reply]

People should be aware that Special:Relatedchanges includes changes to templates. FlaggedRevs would also be an excellent alternative to this. — Werdna • talk 06:12, 22 August 2008 (UTC)[reply]

Trial started under my account --Chris 09:24, 22 August 2008 (UTC)[reply]
Edits during this trial run can be found here SQLQuery me! 08:01, 24 August 2008 (UTC)[reply]
  • My only question is that when an RfA was run on a prior ProtectionBot, it was later said that the bot was unnecessary because the protection of the templates had been coded to work automatically ("cascading protection") ... does this do something differently from that, or has that functionality been dropped, or what is new here? Please pardon this question from a person with no technical skills at all. Newyorkbrad (talk) 02:34, 22 September 2008 (UTC)[reply]
    ProtectionBot would have protected all the templates used on a protected page, since cascading protection did not exist at that time. This proposal is for a bot to add templates used on the Featured Article to an already cascade protected page, so that they also become cascade protected. This is because the featured article is not full protected to permit open editing, and therefore the templates used on it are at a greater risk of abuse than normally permitted. Best way to simplify it: ProtectioBot protected pages, FA Template Protection Bot will only edit a protected page. MBisanz talk 02:41, 22 September 2008 (UTC)[reply]
  • I pretty much agree with MBisanz's post of 02:41, 22 September 2008 (UTC). I think there is general community support for this bot and BAG approving admin bots outside of RFA. The issue of those already existing admin bots that have never been formally approved in any way is thornier however. — Rlevse • Talk • 02:51, 22 September 2008 (UTC)[reply]
I'm going to go out on a limb here and add  Approved.. I feel that there is community consensus for this bot to run on a separate flagged account. This doesn't seem like a precarious task either, as it has been running under Chris G's account for quite a while. To comments about the concerns, here's what I say: There is little to none chance that a vandal can cause a vandalized template to be protected due to the safety function that has been added. In addition, those people who believe that this will be a skynet incident: I think that we have a trustworthy operator here. Chris G has the state of mind to know what a great responsibility he has here. If he violates it, we will definitely take action. I am copying this over to Wikipedia:AN, Wikipedia:VPM, and Wikipedia:BN to get the word out. If anyone wants to stab or trout me for this, please remember to be civil. Good luck, Chris. Xclamation point 03:03, 22 September 2008 (UTC)[reply]
Reopening for comments. Xclamation point 04:03, 22 September 2008 (UTC)[reply]

Flags

[edit]

I am minded to the grant the +sysop and +bot flags in accordance with Wikipedia:Bot policy#Bots with administrative rights, which seems to enjoy a growing consensus, once that ceases to be "proposed" policy. To facilitate this, could BAG members please check that the requirements of that section have been complied with and sign off below if they are in agreement that this bot can be given the appropriate flags? WJBscribe (talk) 00:46, 29 September 2008 (UTC)[reply]

Bureaucrat comment

[edit]

I have determined that this proposal meets the requirements set down at Wikipedia:BOT#Bots with administrative rights and am therefore approving the bot to run with sysop and bot flags. Note that this approval is strictly conditional on there being a continued consensus for the bot to operate: should there be a consensus of either BAG or the community that approval should be withdrawn, a request will be made to a steward to remove the sysop flag (bot status can, of course, be removed locally). Any further tasks must be approved separately. WJBscribe (talk) 00:18, 30 September 2008 (UTC)[reply]


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