Page MenuHomePhabricator

Add 'deletedtext' right to user group 'researchers'
Closed, ResolvedPublic

Description

The "researcher" user right was originally created to allow researchers the ability to process the history of Wikipedia more effectively. A lot of research that is relevant to Wikimedians requires access to deleted content. For example, building machine learning models to catch spammers, track copyright violations, and otherwise model bad behavior requires access.

Right now, we work with the WMF Trust and Safety to to help researchers get access to deleted content (in combination with a Non-disclosure agreement and a Memorandum of Understanding), but the user-groups that we can give them are somewhat inappropriate since they allow for much more than access to deleted content. E.g., sysop and staff rights are overkill. This "researcher" user right is a good opportunity to grant temporary and well-scoped access to these researchers.

See the "researchers" user group and some details about its origin:
https://en.wikipedia.org/wiki/Special:ListGroupRights
https://en.wikipedia.org/w/index.php?title=Special:ListUsers&group=researcher
https://en.wikipedia.org/wiki/Wikipedia:User_access_levels#Researcher

Event Timeline

Halfak created this task.Fri, May 22, 9:16 PM

Change 598553 had a related patch set uploaded (by Ammarpad; owner: Ammarpad):
[operations/mediawiki-config@master] Add 'deletedtext' permission to researcher group

https://gerrit.wikimedia.org/r/598553

Approved, this can be added to the user rights group. Making it available to researchers will, as always, be administered through the Legal department's review process case by case as appropriate, including the MOU reviews Aaron references above.

Approved, this can be added to the user rights group. Making it available to researchers will, as always, be administered through the Legal department's review process case by case as appropriate, including the MOU reviews Aaron references above.

Thanks for the quick sign-off. This can proceed to SWAT then.

Change 598553 merged by jenkins-bot:
[operations/mediawiki-config@master] Add 'deletedtext' permission to researcher group

https://gerrit.wikimedia.org/r/598553

Mentioned in SAL (#wikimedia-operations) [2020-05-26T11:14:37Z] <awight@deploy1001> Synchronized wmf-config/InitialiseSettings.php: SWAT: [[gerrit:598553|Add 'deletedtext' permission to researcher group (T253420)]] (duration: 01m 06s)

MarcoAurelio closed this task as Resolved.Tue, May 26, 11:21 AM
MarcoAurelio assigned this task to Ammarpad.

Note that @FaFlo and @Capt_Swing (Jtmorgan) are not (or at least no longer) WMF Staff, so their status with the user group will need some review as we extend the user right. I'd be happy to work with @Capt_Swing to get him on an MOU if he doesn't already have one with us. I'm also interested in talking to @FaFlo if he'd like to retain the right. @JanWMF, would you like to coordinate that?

@Halfak: @Capt_Swing is disabled. You might want to ping him somewhere else/on a working personal account.

FaFlo added a comment.EditedTue, May 26, 3:07 PM

Greetings. I do not really require it for any specific research, so I'd fine with being removed. Iirc, back in the day, it was meant mostly for retrieving more than 50 revisions with full content via the API (nowadays mainly needed to keep the WikiWho API running smoothly). But I probably can achieve the same with a normal bot account, right?

Greetings. I do not really require it for any specific research, so I'd fine with being removed. Iirc, back in the day, it was meant mostly for retrieving more than 50 revisions with full content via the API (nowadays mainly needed to keep the WikiWho API running smoothly). But I probably can achieve the same with a normal bot account, right?

@FaFlo Yes, the bot user group does have the apihighlimits permission allowing you to retrieve more API results than normally you'd be able to. Researcher would be required if you need to access deleted content over en.wikipedia.

JJMC89 added a subscriber: JJMC89.Tue, May 26, 6:31 PM

Greetings. I do not really require it for any specific research, so I'd fine with being removed. Iirc, back in the day, it was meant mostly for retrieving more than 50 revisions with full content via the API (nowadays mainly needed to keep the WikiWho API running smoothly). But I probably can achieve the same with a normal bot account, right?

@FaFlo Yes, the bot user group does have the apihighlimits permission allowing you to retrieve more API results than normally you'd be able to. Researcher would be required if you need to access deleted content over en.wikipedia.

API high limit requestors global group might be a better option for WikiWho since it covers multiple projects.

@Halfak et al. happy to sign an MOU :) I have found the existing researcher rights useful in the past, for both WMF and non-WMF work (e.g. debugging why HostBot sent an invite to someone it shouldn't have, and also real research). So would like to keep it as long as it doesn't create any headaches for others.