Wikidata:Requests for permissions/Bot/MusikBot II
- The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
- Withdrawn — MusikAnimal talk 19:12, 9 September 2021 (UTC)[reply]
MusikBot II (talk • contribs • new items • new lexemes • SUL • Block log • User rights log • User rights • xtools)
NOTE: This BRFA is paired with Wikidata:Requests for permissions/Administrator/MusikBot II. Both must pass for either to be valid.
Operator: MusikAnimal (talk • contribs • logs)
Task/s: Automatically protect highly-transcluded templates and modules based on community-maintained configuration
Code: GitHub
Function details: This is proposing bringing English Wikipedia's TemplateProtector to Wikidata. This is in response to a recent wave of template vandalism. The same bot task also runs on Meta. The enwiki documentation should explain everything, but the basic idea is that you set thresholds for the number of transclusions that, once exceeded, the bot would apply protection. You can run a script occasionally to find high-risk templates and protect them manually (as was recently or will soon be done), but at any time a template could jump from just a few transclusions to a many thousands. This is why automation is needed. For Wikidata, I might recommend configuring semi-protection after 500 transclusions, and sysop protection after 5000. Any admin will be able to change this via the on-wiki config.
The bot obviously will need admin rights to do this task. I do not have local admin rights on Wikidata, but I am a steward. I am hoping this is enough to buy your trust, since technically I already have admin rights here, I just don't use them except in emergencies (which would not include anything related to this bot). I have filed an RfA at Wikidata:Requests for permissions/Administrator/MusikBot II for further discussion.
A side note: I'm requesting as MusikBot II because that's the same account that this task runs as on other wikis. There are no tasks running as MusikBot on Wikidata.
— MusikAnimal talk 04:32, 17 August 2021 (UTC)[reply]
- Support BrokenSegue (talk) 15:42, 17 August 2021 (UTC)[reply]
- Yes, I think you would need a request at Wikidata:Requests for permissions/Administrator.--Ymblanter (talk) 19:46, 18 August 2021 (UTC)[reply]
- @Ymblanter After filing this BRFA, DannyS712 mentioned we might need consensus for preemptive protection of templates. I would think this is uncontroversial seeing as it is standard practice on nearly every wiki (i.e. the same reason you protect the Main Page, etc.), and also we've actively been experiencing template vandalism. Do you think an RfC of sorts is needed? If not, should I go ahead open a RfA for the bot now? Thanks, — MusikAnimal talk 20:42, 18 August 2021 (UTC)[reply]
- I do not have a strong opinion on this issue, but probing it at the Project Chat could probably help.--Ymblanter (talk) 20:46, 18 August 2021 (UTC)[reply]
- I asked at Wikidata:Project Chat (permalink), between the admin that comented there and this discussion, it seems a proper RfC isn't needed. So I've gone ahead and filed an RfA for the bot. — MusikAnimal talk 03:24, 19 August 2021 (UTC)[reply]
- I do not have a strong opinion on this issue, but probing it at the Project Chat could probably help.--Ymblanter (talk) 20:46, 18 August 2021 (UTC)[reply]
- @Ymblanter After filing this BRFA, DannyS712 mentioned we might need consensus for preemptive protection of templates. I would think this is uncontroversial seeing as it is standard practice on nearly every wiki (i.e. the same reason you protect the Main Page, etc.), and also we've actively been experiencing template vandalism. Do you think an RfC of sorts is needed? If not, should I go ahead open a RfA for the bot now? Thanks, — MusikAnimal talk 20:42, 18 August 2021 (UTC)[reply]
- Do you have samples for the perceived need? Could this simply be solved by semi-protecting template namespace? (we have done this for other namespaces as it is thought that these shouldn't be protected one-by-one. How is it planned that admins cope with edit requests for needlessly protected pages (given that there is already a serious backlog for admins in the field)? --- Jura 11:35, 19 August 2021 (UTC)[reply]
- Probably semi-protection as the highest protection level for this task would be sufficient in Wikidata. I don't see a need for full protection, regardless of the number of transclusions. —MisterSynergy (talk) 11:44, 19 August 2021 (UTC)[reply]
- We have recently seen some vandalism on e.g. these module pages. Not sure, though, whether it is a systematic problem; all but one other edits from recent changes in Template and Module namespace of the past 30 days were clearly damaging as well. (Highly used) template and module pages do not need to be edited by IP users and newcomers anyways. —MisterSynergy (talk) 11:56, 19 August 2021 (UTC)[reply]
- So we should just semi-protect the two namespaces? --- Jura 12:11, 19 August 2021 (UTC)[reply]
- That's however not possible without having an adminbot for that @Jura1. I think that semiprotecting only highly transcluded pages should be tried first, as it's a smaller hammer than semiprotecting everything. Martin Urbanec (talk) 22:06, 19 August 2021 (UTC)[reply]
- I think you are mistaken. In any case, the proposer hasn't really demonstrated the need for the bot. --- Jura 22:37, 19 August 2021 (UTC)[reply]
- The need follows an ongoing wave of template vandalism. The assumption was you'd rather take the least admin action necessary to prevent abuse. If you as a community are comfortable semi-protecting all of the template and module namespaces via title blacklist, let's say, that will largely solve the problem. However some of the templates with very high transclusion counts (I can provide a list to sysops), should probably be sysop-protected, seeing as autoconfirmed is easily attainable. Bear in mind that it's possible any template can also suddenly become highly visible too, if it were transcluded on another highly visible template. Humans generally do not think to check these things, which is why automation is helpful.
- However, unless I'm mistaken, these templates aren't ever used in statement values, meaning it won't propagate to the wikis where most Wikidata content is seen by readers. So perhaps it doesn't matter much. My goal was to put a stop to template vandalism here. Whatever means to do that is fine by me, and if you're simply not interested and think you can manage, that's fine too :) — MusikAnimal talk 15:34, 20 August 2021 (UTC)[reply]
- Do you have samples of edits you consider meriting protection? (please provide diffs). Your explanation of the need is somewhat vague. --- Jura 16:55, 20 August 2021 (UTC)[reply]
- As noted above, there are a number of examples in recent changes. I noticed this myself as I locked a number of these accounts (note that depending on the account, SUL details may only be visible to stewards). On further investigation I discovered Wikidata has quite a few high-risk templates with no protection whatsoever, and others probably deserve sysop protection. I asked around, and multiple admins expressed interest in the bot, so here I am. I'm not sure how to make the situation more clear. If you don't agree with the bot, that is fine :) It's just the best solution we have right now until such a feature is available in MediaWiki (phab:T237814). — MusikAnimal talk 18:02, 20 August 2021 (UTC)[reply]
- Can you provide specific samples? If you can't demonstrate the need with diffs, it's ok, but it's just a sign that maybe you shouldn't be operating the bot. We actually already have the possibility to semi-protect the namespace. --- Jura 20:52, 20 August 2021 (UTC)[reply]
- I'm confused why you need specific diffs when recent changes gives ample examples. The diffs are listed there. Every edit shown has been reverted. Many of these such as Special:Diff/1482254867 are long-term abuse and the revision history I think clearly shows this is an ongoing problem. I am fully aware you can semi-protect the namespace(s) without the need for the bot, as I mentioned above. I feel like I am unduly being given a hard time. No, I am not a regular at Wikidata, if that's what you want me to say, but I know what I'm doing when it comes to counter-vandalism and I certainly am capable of operating this bot. I'm more than happy to withdraw this BRFA and the associated RfA if this is the kind of treatment I can expect when I am trying to help. — MusikAnimal talk 21:18, 20 August 2021 (UTC)[reply]
- It seems that local admins disagree with the need for protection for the sample you provided. --- Jura 21:35, 20 August 2021 (UTC)[reply]
- I'm confused why you need specific diffs when recent changes gives ample examples. The diffs are listed there. Every edit shown has been reverted. Many of these such as Special:Diff/1482254867 are long-term abuse and the revision history I think clearly shows this is an ongoing problem. I am fully aware you can semi-protect the namespace(s) without the need for the bot, as I mentioned above. I feel like I am unduly being given a hard time. No, I am not a regular at Wikidata, if that's what you want me to say, but I know what I'm doing when it comes to counter-vandalism and I certainly am capable of operating this bot. I'm more than happy to withdraw this BRFA and the associated RfA if this is the kind of treatment I can expect when I am trying to help. — MusikAnimal talk 21:18, 20 August 2021 (UTC)[reply]
- Can you provide specific samples? If you can't demonstrate the need with diffs, it's ok, but it's just a sign that maybe you shouldn't be operating the bot. We actually already have the possibility to semi-protect the namespace. --- Jura 20:52, 20 August 2021 (UTC)[reply]
- As noted above, there are a number of examples in recent changes. I noticed this myself as I locked a number of these accounts (note that depending on the account, SUL details may only be visible to stewards). On further investigation I discovered Wikidata has quite a few high-risk templates with no protection whatsoever, and others probably deserve sysop protection. I asked around, and multiple admins expressed interest in the bot, so here I am. I'm not sure how to make the situation more clear. If you don't agree with the bot, that is fine :) It's just the best solution we have right now until such a feature is available in MediaWiki (phab:T237814). — MusikAnimal talk 18:02, 20 August 2021 (UTC)[reply]
- Do you have samples of edits you consider meriting protection? (please provide diffs). Your explanation of the need is somewhat vague. --- Jura 16:55, 20 August 2021 (UTC)[reply]
- I think you are mistaken. In any case, the proposer hasn't really demonstrated the need for the bot. --- Jura 22:37, 19 August 2021 (UTC)[reply]
- That's however not possible without having an adminbot for that @Jura1. I think that semiprotecting only highly transcluded pages should be tried first, as it's a smaller hammer than semiprotecting everything. Martin Urbanec (talk) 22:06, 19 August 2021 (UTC)[reply]
- I think that semi-protection should be sufficient for this task, full protection should not be needed. --Ameisenigel (talk) 18:33, 23 August 2021 (UTC)[reply]
- So we should just semi-protect the two namespaces? --- Jura 12:11, 19 August 2021 (UTC)[reply]
- By my quick calculation, currently less than 200 templates are protected. [1] How many will the bot protect in addition to that in an initial batch? --whym (talk) 13:13, 27 August 2021 (UTC)[reply]
- However many meet the thresholds established by the Wikidata community. I recommend applying semi-protection after 500 transclusions, which at the time of writing would effect 187 templates and 80 modules. — MusikAnimal talk 00:27, 31 August 2021 (UTC)[reply]
- Can you explain what you refer to the "thresholds established by the Wikidata community"? I find it rather disturbing that you make such claims, can't provide a single template illustrating the need for protection due to vandalism, offer no remedy for editing templates after protection whereas protected edits already have a backlog. I'd expect a bit more from WMF staff. --- Jura 10:23, 7 September 2021 (UTC)[reply]
- I am baffled by your animosity towards me. It's very clear you oppose the bot, which is more than fine. In conveying this I would appreciate polite, constructive criticism and discourse rather than personal attacks. Anyway, the intent was to get approval for the bot, then allow the admins and broader community decide what the thresholds are. So the thresholds I speak of have not been established yet (though they could be, if there's clear consensus). I offered my recommendation for them but in reality my role would not extend beyond maintaining the bot itself. I thought this was clear by now, but the idea is for preemptive protection, not necessarily in the response to preexisting vandalism. Myself and others have provided examples where there was prior vandalism (and indeed that's what prompted this whole effort), but those examples are only scratching the surface. I will not explicitly disclose any loopholes here in a public forum, but I will say it would seem it is currently possible for anyone to make visible changes to Wikidata:Main Page. For that I recommend simply using cascading protection, but perhaps this is not a concern.
In hindsight, the concept of preemptive protection probably isn't that important for Wikidata, anyway. This project is not subject to a PR disaster like content wikis are given most readership of the content does not occur on Wikidata itself, and templates are not used in statement values. At any rate, this effort is not worth the stress it has brought. When multiple admins expressed to me their interest in the bot, I was expecting smooth sailing in getting it approved. I kindly withdraw. — MusikAnimal talk 19:12, 9 September 2021 (UTC)[reply]
- I am baffled by your animosity towards me. It's very clear you oppose the bot, which is more than fine. In conveying this I would appreciate polite, constructive criticism and discourse rather than personal attacks. Anyway, the intent was to get approval for the bot, then allow the admins and broader community decide what the thresholds are. So the thresholds I speak of have not been established yet (though they could be, if there's clear consensus). I offered my recommendation for them but in reality my role would not extend beyond maintaining the bot itself. I thought this was clear by now, but the idea is for preemptive protection, not necessarily in the response to preexisting vandalism. Myself and others have provided examples where there was prior vandalism (and indeed that's what prompted this whole effort), but those examples are only scratching the surface. I will not explicitly disclose any loopholes here in a public forum, but I will say it would seem it is currently possible for anyone to make visible changes to Wikidata:Main Page. For that I recommend simply using cascading protection, but perhaps this is not a concern.
- However many meet the thresholds established by the Wikidata community. I recommend applying semi-protection after 500 transclusions, which at the time of writing would effect 187 templates and 80 modules. — MusikAnimal talk 00:27, 31 August 2021 (UTC)[reply]