Talk:Global rollback/Archive 2

RFC: removal of global rollback due to inactivity

Previous discussions on the subject:

• 1
• 2
• 3

In the last discussion it was mentioned the possibility to establish an inactivity policy for global rollbackers as well. By looking at the list of users with global rollback rights I see a lot of users that at some time were active in the SWMT and helped over there, but now they do not engage in such activities. Since for granting this right the policy says that the user must be demonstrably active in activities where this permissions might be required, I think that users that hold this rights should have, at least, some minimum use of this right or it should be removed.

In the past threads where this was discussed there was agreement that an inactivity policy would be fine, but a criteria was never set. This RFC intends to seek such criteria and enact it in the policy if the community agrees with it.

I'd say that if any user with global rollback rights does not perform a number of rollbacks outside the wikis where they have rollback or admin rights in a whole year will be demoted, with the posibility of re-applying through the regular way.

Thanks. -- MarcoAurelio (talk) 14:05, 18 April 2013 (UTC)

Do we have stats on how many go missing after one year and do not return within two? What is the means that we would be looking to measure when they last used the tool? Or is it just an editing presence?

In light of the proposal around advanced administrative rights, I am wondering whether you would consider a two year period. We prod after one year of inactivity, and remove no further notice after the second. I am wanting something low maintenance but effective. There is not a lot of risk of the right being held for too long, but it is needed to be removed rather than left drift forever. — billinghurst sDrewth 15:38, 18 April 2013 (UTC)

What problem is this proposal going to solve? Rollback only allows users to do an action, which any user can do with two click of mouse, with just one click. Rollback is hardly an "advanced right". Ruslik (talk) 18:31, 18 April 2013 (UTC)

It is not always technically possible to recognize whether a certain revert has been performed using rollback or not, as there are some scripts to customize the edit summary of rollbacks, so it would be impossible to consequently implement an inactivity policy for global rollbackers. I also agree with Ruslik. Regards --Iste (D) 19:57, 18 April 2013 (UTC)

I don't see any reason why rollback needs to have inactivity standards. It doesn't do much damage, it's granted to trust people and if they go rogue we can always remove it. Almost all global rollbackers have much more sensitive permissions anyway, and if they want to wreak havoc they'll use those. It is not an advanced right and there's no major harm in letting them keep it. Snowolf ^{How can I help?} 20:11, 18 April 2013 (UTC)

I do not think that users that went inactive in the activities that lead them requesting this permission —primary SWMT— should continue to have it, as with any other additional rights. We even have formally retired editors holding this global rollback permission, and it respectfully makes no sense to me. If they returned they would be great but I think that additional rights of any sort are to be used, and when the need of them ceases, they should be resigned. Yes, it makes no harm to keep inactive people populating that list; but the same could be said of practically any other permissions.

Regarding the point Iste Praetor mentions, I agree that measuring the inactivity just based on the use of rollback would be difficult in some cases. I do not think that should be the standard. I was thinking more like checking if the user has still some activity in SWMT. I think that it would be very simple and easy with tools like "luxo". I think that there are reasonable doubts of inactivity in SWMT if the user has not made a single edit outiside his/her regular wikis in a year or in two years. It is unlikely that users that have some activity in the SWMT doesn't have a number of edits outside the wikis they're regulary active on IMHO.

I tend to agree with Billinghurst above that we do not need a complicated system but something simpler and effective. The discussion is open to propose methods. I'd propose that rather than setting hard inactivity standards like X number of rollbacks on Y time, I would, for example, message the users that seem to be inactive on their talk pages asking if they are still interested in doing SWMT and keeping the permission. If they do not respond in one month or say that are no longer interested, the permissions are removed; otherwise if the user say that is still interested in helping and keeping the permission the permissions can be kept. Result: in the worst case we will have an inactive user returning from inactivity and joining the SWMT sporadically just to keep the permission (which is help after all) or, in more happy cases, we will encourage users returning to activity because they have discovered the passion of doing SWMT activies again ;-).

Best regards and thanks for your participation. -- MarcoAurelio (talk) 13:25, 19 April 2013 (UTC)

Is this a vote? If so, I  Support an inactivity policy for global rollbackers. No need to have them forever if people are inactive imho. Trijnstel_talk 20:17, 21 April 2013 (UTC)

I think this is a solution in search of a problem. What kind of damage or inefficiency can an inactive rollbacker cause? Besides, we can't have an inactivity policy without a defined expectation for activity. As Iste Praetor said, it can be difficult to tell what is a rollback and what is not. PiRSquared17 (talk) 20:30, 11 May 2013 (UTC)

The question I have is, what if a global rollbacker leaves Wikimedia entirely? --Rschen7754 07:37, 28 May 2013 (UTC)

• I asked global rollback mostly for six small wikis I check every day (lbe, xal, sah, crh, os, cv). Some of them have very little or no activity, and vandalism is not that common. I believe I use my rollback there may be once in several months. I can easily imagine a situation when I have not used it for a year, and then according to the proposal I just lose it, despite the fact that I am active, and I believe I did not abuse the rollback. Fine, then I will have to roll back in two clicks, but I just do not understand what this proposal aims to achieve.--Ymblanter (talk) 19:09, 28 May 2013 (UTC)
  • I believe this proposal is for totally inactive people, not for those who are still active somewhere. Trijnstel_talk 10:49, 30 May 2013 (UTC)

• I'd say that if a global rollbacker is still active (edits within the last six months) anywhere then they can keep the bit. Rollback isn't the only right in the package, there are other useful ones, and unless there is some reason that we don't trust someone anymore then we shouldn't be removing rights from them that they could still be using. Ajraddatz (Talk) 22:14, 12 June 2013 (UTC)

• I would say that we could remove global rollbackers who left the Wikimedia projects entirely for more than 2 years in order to prevent a GR list full of retired Wikimedians. Anyway, I agree that the GR right doesn't particularly need an inactivity policy which governs the use of the right itself as it is a quite minor right and also not community-serving. Vogone talk 19:47, 6 February 2014 (UTC)
• I agree with Vogone. --Rschen7754 21:01, 6 February 2014 (UTC)

Any progress here? Since there hasn't been any new input for almost 1 year, I'd suggest to close this with the result to apply inactivity removals to global rollbackers who haven't shown any editing activity (total inactivity) for the last 2 years. That seems to be a solution almost every participant here could agree on, at least that is my impression from what I read. Vogone (talk) 22:17, 17 January 2015 (UTC)

It's been some time indeed, and I don't see strong consensus in adopting this, so I think it's better if I withdraw this proposal for now. Thanks. -- M\A 18:02, 25 January 2015 (UTC)

Autoreview

Related topic: Talk:Global sysops#Additional flags and name change

Hi all, I think it would be logical to include "autoreview" flag to GR permission as well as "autopatrol", since many times we perform "undo" function and "autoreviewrestore" is not enough. What do you think?--Syum90 (talk) 08:27, 23 June 2015 (UTC)

I think this was previously discussed, but rejected because being a gr does not necessarily mean that all edits should be reviewed on those pages. GR doesn't imply any editorial knowledge, especially since it applies everywhere. Ajraddatz (talk) 08:46, 23 June 2015 (UTC)

I've had doubt about that which does not include this permission to GR. I think it is useful for us to use GR permission including "autoreview" function as a patrol activity.--Infinite0694 (Talk) 08:49, 23 June 2015 (UTC)

@Ajraddatz: as I mentioned above, it's the same case than "autopatrol" flag, I think, and this flag is included into the permission. Maybe "autopatrol" flag has to be removed from the permission.--Syum90 (talk) 10:45, 23 June 2015 (UTC)

According to Talk:Global rollback/Archive 1#Rights, throttle, it mentions that RC patrollers don't need to patrol our edits. So, I reckon that what Ajraddatz says about that and what "autopatrol" flag means are two different stories.--Infinite0694 (Talk) 12:34, 23 June 2015 (UTC)

@Infinite0694: I don't think so; on wikis where "patrol" is enabled for edits (not only new pages) each edit has to be checked, and it's the same on wikis where flaggedrevs are enabled, where each edit has to be checked too. They are two different tools to do the same. I think it's a matter of consistency, I think the two flags or none should be included, but not one yes and the other no, I see no reason for it.--Syum90 (talk) 13:19, 23 June 2015 (UTC)

Again, this is just a memory, but it makes sense: I seem to remember autopatrol being included so that all reverts are marked as patrolled, since there is no "autorevertpatrol". Since the review right is split up that way, global rollback has the one that has to do with reverting, but not the one that deals with editorial control. Ajraddatz (talk) 18:10, 23 June 2015 (UTC)

Patrol and FlaggedRevs are not the same. Patrol is merely for countervandalism purposes while FlaggedRevs also allows judgement of content quality. Vogone (talk) 19:53, 23 June 2015 (UTC)

@Vogone: "review" flag is only for check the revision, the flag for judgement of content quality is "validate". ajr's last comment has much sense.--Syum90 (talk) 07:22, 24 June 2015 (UTC)

Like no one expects the Spanish Inquisition it seems nobody understands the Flagged Revisions. It can be tweaked according to the wishes of the particular community. For example the Finnish Wikipedia has "Editors" who can "review" changes to any level, and while they have a job title that is inferior to "Reviewers", they have more power than those "Reviewers" who only can review and cannot "validate". While I understand the point and favour the arguments, I still oppose the idea of granting a global right that simply does not mean the same everywhere. --Pxos (talk) 13:40, 26 June 2015 (UTC)

2016

Hi, look this page - Requests for comment/Inactive Global rollback users--6AND5 (talk) 14:13, 16 February 2016 (UTC)

suppressredirect

I suggest we clarify that use of suppressredirect for anything other than non-obvious pagemove vandalism is also subject to "local policies". Any objections? xaosflux ^Talk 19:25, 28 February 2016 (UTC)

All user rights included are subject to local policies. --Vogone (talk) 19:35, 28 February 2016 (UTC)

Thank you, I tweaked the wording - to specify that ALL of these permissions are primarily subject to local policies not just the "rollback" permission - what led me here in the first place was a discussion about specific suppressredirects on enwiki made by a globalrollbacker that was not also an enwiki sysop. xaosflux ^Talk 03:05, 5 March 2016 (UTC)

Users removed due to inactivity

As per the result of Requests for comment/Inactive Global rollback users, the following users are removed from the global rollback group as, not having editted on any project for at least 2 years:

• Jorunn
• Werdan7
• Gaeser
• Erwin85
• JenVan
• Reder
• Iste Praetor
• Hu12

The users will receive a note on their talk pages. This list is not closed, and might increase in the next days. Best regards, —MarcoAurelio 17:46, 14 August 2016 (UTC)

Hmm, I see RadiX handled those already. New data in progress though. —MarcoAurelio 17:49, 14 August 2016 (UTC)

Proposal (patrolmarks)

I propose to this global group to add the right patrolmarks. I think it can help.-MrJaroslavik (talk) 12:19, 27 November 2019 (UTC)

@MrJaroslavik: I look forward to your reasoned proposal. "I think that it can help" is an opinion, it is no rationale for a change. I would suggest that in your rationale that you include a link to mw:Help:Patrolled edits and also what would be the considered benefits, rather than just focusing on the feature.  — billinghurst sDrewth 13:12, 27 November 2019 (UTC)

patrolmarks rights IMO, is useless without patrol right. Masum Reza^☎ 22:32, 14 December 2019 (UTC)

@MrJaroslavik: On second thought, it would be helpful to filter out non-autopatrolled edits from RC feed. Not all global rollbackers have patrol rights on every wikis. Any comments here? billinghurst? Masum Reza^☎ 05:21, 15 December 2019 (UTC)

@MrJaroslavik, Masumrezarock100, and Billinghurst: For my experience of Global Rollbacker add the patrolmarks rights would be very useful because it allows you to focus your efforts on changes that are unpatrolled (on RC and Special:NewPages). And it allows you to improve the results to find vandalic changes. For example I here on meta am patroller to be able to see the changes unpatrolled. --Samuele2002 (Talk!) 20:59, 19 April 2020 (UTC)

Indeed, patrolmarks makes it possible to filter for only unpatrolled edits; autopatrolled or manually patrolled edits are very unlikely to be vandalism --DannyS712 (talk) 21:09, 19 April 2020 (UTC)

@Samuele2002: Yes, i meant this. I am aware that this is not enabled on all projects, but why not use it where it is enabled.--MrJaroslavik (talk) 21:27, 19 April 2020 (UTC)

I cannot imagine that this right will really help anyone. I personally use this only in Wikidata, in all other projects do most of my reviews and rollbacks with tools (most SWViewer) GUC for User Control. I really only look into the last changes in small wikis with a few edits, everything is clear and you don't need any additional filter.-𝐖𝐢𝐤𝐢𝐁𝐚𝐲𝐞𝐫 👤💬 20:32, 31 July 2020 (UTC)

I understand that not everyone would find it useful, but is there any drawback? We trust GRs, why not let them know which edits are unpatrolled on wikis where that information is available? Support --DannyS712 (talk) 07:26, 1 August 2020 (UTC)

OK, I Support per DannyS --𝐖𝐢𝐤𝐢𝐁𝐚𝐲𝐞𝐫 👤💬 09:39, 3 August 2020 (UTC)

• Comment Posted at Stewards' noticeboard#Granting Global rollbackers patrolmarks about this --DannyS712 (talk) 05:39, 12 August 2020 (UTC)
•  Support --Novak Watchmen (talk) 21:50, 12 August 2020 (UTC)

---

Done Group permissions were changed. --Martin Urbanec (talk) 12:10, 15 August 2020 (UTC)

Proposal to add deletedtext/deletedhistory

LTAs are often on the go in several projects, and is it necessary sometimes have to look at contributions from multiple wikis to see whether it is an LTA or a simple vandal. Local sysops sometimes delete posts without other actions because they do not know every LTA, which makes it difficult to identify crosswiki LTAs in projects without a GS. That is why I proposal the GR also get access to deleted content.--𝐖𝐢𝐤𝐢𝐁𝐚𝐲𝐞𝐫 👤💬 19:32, 31 July 2020 (UTC)

This sounds like overkill. For rollbacking, you don't need to know whether something is mere vandalism or very bad LTA vandalism. --MF-W 21:42, 31 July 2020 (UTC)

@MF-WThis is for the initiation of further measures, i.e. request on SRG, checkuser request etc.--𝐖𝐢𝐤𝐢𝐁𝐚𝐲𝐞𝐫 👤💬 23:00, 31 July 2020 (UTC)

Sounds like overkill to me too. Rollback should be sufficiently determined by the content, not by the history.  — billinghurst sDrewth 06:21, 1 August 2020 (UTC)

 Oppose Huge overkill, if you want to be a sysop, apply to be a sysop - rollback is meant to be a simple and quick content fixing tool. — xaosflux ^Talk 14:48, 12 August 2020 (UTC)

 Oppose per xaosflux; it would mean an increase in the review and granting; let us keep it simple. If this was to progress I would want it to be via a RFC where the big wikis could see and comment. The impact would be upon them, and I don't think they would favour it. They definitely would want an opinion.  — billinghurst sDrewth 16:25, 12 August 2020 (UTC)

 Oppose This is clearly out of scope of the group. * Pppery * _{it has begun} 14:43, 15 August 2020 (UTC)

Comment if this is approved, stewards please confirm with WMF legal that this is allowed before adding the rights, given that it allows viewing deleted content --DannyS712 (talk) 09:40, 22 August 2020 (UTC)

 Oppose Generally, these features are given to groups such as admin, checkuser, oversighter and member of the arbitration committee in local projects. For some this roles, you must be a trusted user and approve confidentiality agreements. Of course, global rollbackers are also trusted users. However, some deleted content also contains personal information such as home address, credit card number, phone number. At this point, globall rollbackers will also have possible liabilities. Imho, viewing deleted contents isn't a good idea for the global rollbacker group. --Uncitoyen^talk 09:58, 22 August 2020 (UTC)

DannyS712 + Uncitoyen you are confusing the rights with "suppressrevision/suppressionlog" --𝐖𝐢𝐤𝐢𝐁𝐚𝐲𝐞𝐫 👤💬 14:23, 23 August 2020 (UTC)

No, sometimes when the oversighters aren't active, the admins can hide these contributions with revisiondelete on the Turkish Wikipedia. Then, when oversighters are active, they suppress to these contributions. This allows global rollbackers to see this on the Turkish Wikipedia. Perhaps there are other local wikis with such situations. This is a concern for me. --Uncitoyen^talk 14:29, 23 August 2020 (UTC)

No, see the second quote at w:Wikipedia:Viewing deleted content DannyS712 (talk) 22:51, 23 August 2020 (UTC)

 no further action, no consensus to progress  — billinghurst sDrewth 22:38, 26 February 2021 (UTC)

Local policy on jawiki

The jawiki community discussed rollbacker group. At that time, we set the rules for GR. (See also ja:Wikipedia:巻き戻し者#グローバル巻き戻し者) This rule means the local administrator can ask GR to stop using rollback and other additional features.

I suggest adding this jawiki rule to GR#Local policies section.--mirinano (talk) 03:34, 16 August 2020 (UTC)

I will presume that a jaWP admin will come along and add it when they so choose.  — billinghurst sDrewth 13:11, 16 August 2020 (UTC)

 Oppose There is little practical difference between the JAWP policy and the global policy, so there is no need to provide us with further information.--Infinite0694 (Talk) 16:41, 16 August 2020 (UTC)

You're right. I think GR take a casual glance at this policy but it may not be as much as we would add it. I want to withdraw this suggestion.--mirinano (talk) 16:49, 16 August 2020 (UTC)

Later?

Is there some consensus on whether it can be resigned and then continued on at some point in the future? Or, like Commons' notice "stating that they will be away for a period"? ~Lofty abyss 17:34, 26 February 2021 (UTC)

Not from my understanding, you need to file a new request. Leaderboard (talk) 17:37, 26 February 2021 (UTC)

Actually, if it's two years with no edits at all then that seems long enough... ~Lofty abyss 18:23, 26 February 2021 (UTC)

Remove autopatrol?

Previous discussion: Wikimedia_Forum/Archives/2020-06#Why_does_global_rollback_include_autopatrol?

Please forgive me to bring up the conversation again. I believe this right is primarily added to autopatrol GR’s reverts to ease the burden of recent changes patrollers. However, Recent changes patrol is now disabled by default (wgUseRCPatrol) and most (~700+) wikis only patrol new pages, usually as a means of new article quality control, which is irrelevant to identifying vandalism and spam. It really isn't about trust. It's just not relevant. (Side note: English Wikipedia removed autopatrol from admins)

From another perspective, reverting vandalism isn't the norm, content contributions are. Autopatrol is unlike suppressredirect that can be turned off when GRs contribute to the content. For wikis still with recent changes patrol, if a GR is not an active contributor on a project and only occasionally reverts vandalism, their edits wouldn't impose a burden on that wiki's patrol workflow. If they are active enough to impose a burden, they can be granted local autopatrol when eligible under the local policy. (If they are not eligible for autopatrol, why should we override that just by their "occasional reverts"?)

In short, I feel that keeping this right in the GR toolkit can be of little help, and may contradict the local policy on a significant number of projects. -- 94rain ^Talk 00:38, 19 February 2022 (UTC)

Solution in search of a problem. What is the issue of autopatrolling GRs? When the wikis bring up clear issues, and they show that it is a problem then we can address it. It has been this way for years and there has been no evidence nor abuse, so I think this is an argument in search of a purity. I still think that having GRs edits patrolled outweighs these theoretical issues. it is not about feelings. It was discussed and not actioned.  — billinghurst sDrewth 01:21, 19 February 2022 (UTC)

I am proposing this because an article created by me was brought to attention on English Wikipedia, though they didn't complain much about the content and later I unreviewed the article myself. Later there was a serious case when a GS/GR got their permissions revoked. Eventually they approved a bot task to automatically unpatrol pages created by global rollbackers on English Wikipedia. All these incidents and countermeasures can be avoided if autopatrol was not bundled in the GR toolkit in the first place. English Wikipedia can handle this by unreviewing articles with the page curation extension. But other wikis are not able to "turn off autopatrol". 94rain ^Talk 03:42, 19 February 2022 (UTC)

I stick with my original assessment. The numbers of creations in this situation that need the work are orders of magnitude smaller than the benefits of patrolled reversions. An unneeded purity and as the comment in the wiki said, any GR is pretty certain to be granted local patrol with minimal effort. Also many wikis do not have an AP so the burden that is created is higher than necessary.  — billinghurst sDrewth 05:43, 19 February 2022 (UTC)

Obviously the fact that the English Wikipedia cared enough to write a bot to disable that behavior indicates that there is a real problem. * Pppery * _{it has begun} 06:07, 19 February 2022 (UTC)

English Wikipedia is NOT the measure that I would apply for a real indicator of all wikis, especially when one reads some of the commentary. Show me we have a holistic issue for all the wiki sisters. Show me that we have a real problem with the right, and not some perception to which someone applies bot scripts. Show me that the right is abused.  — billinghurst sDrewth 08:37, 19 February 2022 (UTC)

Does User talk:Tulsi Bhagat/Archive 3#Global rollback and global sysop rights revoked count then? * Pppery * _{it has begun} 15:52, 19 February 2022 (UTC)

I came here to write about, its not working anyways. Its working as autopatroll but not as patroller. earlier it used to work as patroller. I was able to mark pages as patrolled.

• Isolated incidents and puritanical cautions by enwiki community or few active editors there, shouldn't be guiding our global policies and technical transformations.
• Global rollback group is for small wikis and wikis where there are less active editors who can perform antivandal work or patrolling in general. We look after small wikis and if needed large wikis like enwiki, but again most of the active cross wiki vandal fighters who have gr also have local rights and importantly good standing on large wikis too. thats how we allow them to have gr/gs in first place.
• I see nothing wrong in having patroller/autopatroller permissions in gr set.

QueerEcofeminist [they/them/their] 05:12, 4 May 2022 (UTC)