This is an archive of past discussions on Wikipedia:Oversight. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
Just as easily as the oversight permission can be granted, it can be revoked."
After the earlier discussion about tight control of the Oversight capability, the first part of this sentence sounds like a joke. I don't know if that was the intent.
the person has to personally identify themselves to the Foundation, typically by sending a copy of their ID,
then a Steward has to set the permission for that user.
Now, let's say someone abuses oversight and someone (typically a developer, steward, or another oversighter) sees this:
An emergency request is filed at Requests for Permissions, the stewards check the oversight log, and if obvious and blatant abuse, remove the oversight permission and leave a nasty note, or,
They contact the user and/or ArbCom and mention the abuse, in which either it will be stripped from them immediately (see the steps noted in the above example) or ArbCom opens a case regarding the abuse and then requests the removal or not.
This is, of course, all theoretical at this point, considering it's not yet (to my knowledge, anyway) come up. There are no permissions that you can be granted that can't or won't be stripped at a moment's notice if you patently abuse them.
I suppose the only supporting statement for "as easily as it can be granted, it can be revoked" is that the stewards are fairly responsive to permissions requests. If you're an admin, visit special:userrights and just pretend there's a line for "oversight"... that's all that's involved, really.
Lastly, if you're curious, there are screenshots for various things at commons:MediaWiki bureaucrat (for the Steward screen. I'll take one showing all the current rights on Meta soonish and update things) and commons:MediaWiki oversight for the screens that oversighters themselves use.
For those interested, there is no consensus to do this. Firstly, usernames and article titles may be, themselves, removed from the database if need be for oversight concerns. This isn't directly changed by those with the +oversight permission, but they are the ones who request that our developers remove the information.
If making the logs available to admins were good enough, we wouldn't need oversight at all, as the admins could simply delete an individual revision (admins: don't know how? contact me, I'll explain)... quite frankly, oversight is used for those revisions that not even the admins should see and such information can easily be carried in the username or edit summary as it can the actual edit.
We have a very large pool of those with oversight permissions, and they're not all buddies with eachother. If one of them abuses the permission, they'll certainly hear about it from the others. Too much of this, and they're guaranteed to have the permission removed. ~Kylu (u|t) 01:51, 29 January 2008 (UTC)
Clarify article please
Hi - Neither this or WP:RFO mention whether uninvolved 3rd parties can request oversight. I noticed someone had published an email address in WP:RS and raised it on ANI - [1] who asked me to request oversight. Thanks -- John (Daytona2 · talk) 14:55, 26 November 2007 (UTC)
Anyone can request an oversight, even non-editing members of the public, but we request that you not send non-oversight-related email to the list. If you're not sure, you can send email to the info-en OTRS queue (info-en@wikipedia.org) and let the OTRS team determine if a normal deletion can handle the situation, or to escalate it to the Oversight list. ~Kylu (u|t) 04:44, 16 January 2008 (UTC)
How do you mean "in the projects" exactly? If you check WP:RFO, you'll notice that there are instructions for non-Wikipedia-gurus to submit the requests. ~Kylu (u|t) 06:51, 17 January 2008 (UTC)
Having looked again I just mean WP:RFO. There is nothing that specifically says that anyone (eg uninvolved 3rd parties) can request oversight, leading, in my case to the assumption that I couldn't. -- John (Daytona2 ·Talk ·Contribs)11:53, 17 January 2008 (UTC)
Oversighting of User:Defender 911's edits on 11 August
Can anyone give the community a rough count of how many edits of User:Defender 911 were oversighted on or about 11 August? He was indef blocked on 11 August and has come back asking forgiveness, but the argument is being made at WP:AN#User:Defender 911 that no one can judge the extend of his first-offense block as some massive number of edits were removed. Just want to make sure the community isn't being confused here by people who could be misremembering these four month old events. -- Kendrick7talk11:15, 16 December 2007 (UTC)
I checked the logs for several days before and after that date, and found no edits by this user. If you want a broader search, you'll need to provide the page names; I'm not going to go hunting for a needle in a haystack.
Thanks for the replies. This was in the context of one admin ask for diffs from another admin and being told the incriminating diffs had been eliminated. Although, memory probably played a role here as I suggested. But there's other strangeness about the way this users block became a ban without community input in the first place. I'll continue to try and convince someone to lift this four month old block, but other's are suggesting there may instead be an ArbCom case in your future here. -- Kendrick7talk17:23, 17 December 2007 (UTC)
What about personally identifiable information such as real name, home address and phone number in an article's title?
It's not handled via the same means, but go ahead and send those requests to the oversight list. ~Kylu (u|t) 04:40, 16 January 2008 (UTC)
User:Oversight for "email this user"
This taken but unused account has now been usurped by the Arbitration Committee so as to allow alerts to be emailed by the "E-mail this user" functionality from inside the wiki. For those who don't have their email set up on the computer they're using, and as another useful alert hotline :-) - David Gerard (talk) 22:10, 6 February 2008 (UTC)
Messages sent via emailuser to User:Oversight bounce
(Well, it's being held for approval, but my understanding is that this does not happen with email sent directly to oversight-l?)
Your mail to 'Oversight-l' with the subject
Wikipedia e-mail
Is being held until the list moderator can review it for approval.
The reason it is being held:
Message has implicit destination
Either the message will get posted to the list, or you will receive
notification of the moderator's decision. If you would like to cancel
this posting, please visit the following URL:
https://lists.wikimedia.org/mailman/confirm/oversight-l/XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
It's not a bounce, it's just pending, since oversight-l is a moderated list. When a non-oversighter mails the list, the mod approves it (that way they don't end up with spam). :) ~Kylu (u|t) 00:47, 12 March 2008 (UTC)
Amendment of "Revocation" section and title box text
The old revocation section has turned up to be flawed by ambiguity. This was never spotted, mostly since it has never been needed. So I've redrafted it, and present reasoning here for communal review. Background:
The 'revocation' section was added very early without dissent, but also without discussion. It looks like a section that has been there from the start (July 2006), and never looked at because it's never been needed, in all that time. What is ambiguous is this. The text stated: "If a Wikimedia Steward feels that the editor has abused oversight by hiding revisions which do not qualify under one of the above criteria, they will immediately remove the permission from the editor." This question was raised by a Steward who felt it misleadingly suggested stewards (even if not granted the permission) might be expected to self-authorize for English Wikipedia Oversight.
It is unclear - and has probably never been thought of - was the original edit thinking of stewards in a role of investigator, or decision-maker? That is, does that sentence signify
"Stewards are authorized to randomly check the actions of local oversighters and de-oversight them if they have concerns"? or,
"If a question arises as to whether a matter was proper or not, a steward who feels the user abused oversight may make the decision to remove that right"?
Or was it in fact, a sentence that is unnecessary (but never spotted) since we handle all abuses of rights via arbcom, and stewards have never been asked to do more than enact formal requests to modify permissions?
The latter seems far more likely, since we routinely separate such matters (sysop rights/crat, etc), and there is no precedent of Stewards randomly checking up on other similar rights such as CheckUser, unless they actually have the CU bit. Also the local oversighters routinely check the logs.
The amended text addresses that. Stewards don't act as police or investigators on WMF, as a rule, and apparently seem to actively avoid doing so. WMF policy requires oversighters to patrol each other, as with CheckUser. The current policy and practice is that any question of misuse of a significant right (Sysop, Oversight, CheckUser) is passed to Arbcom who will investigate and request its removal by a steward if there is a problem.
I also clarified two other areas - emergency requests (rare but could happen) and primacy of WMF policy. I note that in reality, none of this has ever come up so it may all be pointless in a way.
Summary of edits:
Title box - corrects the implication that this page mirrors meta and cannot be changed without approval. What cannot be done is to override anything in WMF policy. We can always add more, such as process and other information.
Clarify "patrolling stewards" implication. If there is doubt, refer to Arbcom as with all other elevated permissions.
Add paragraph on emergency requests. Unlikely to be needed, never has been, but one of those things you want for the one time it may happen.
If there is a need to further amend or clarify any of these, or the edit made does not reflect communal best practice, then it will need further discussion and more eyeballs.
I raised this question and I endorse these edits as moving in the right direction at the very least, if not sorting the issue completely. Thanks for the edits, FT2... ++Lar: t/c18:37, 11 March 2008 (UTC)
This user list is out of date -- however, I'm not sure about where to put the new oversight users so I'm not adjusting the list.
Isaac (talk) 19:28, 15 May 2008 (UTC)
I'm confused as to whether oversighting images can now be done by any user with this right, or whether it still requires a developer. The relevant bugzilla thread suggests this is fixed and anyone with oversight can do it, but comments from oversighters suggest this is not the case. Can anyone clarify? WjBscribe16:27, 2 July 2008 (UTC)
I think there's some confusion on bugzilla as to oversighting the image log (the upload, for instance) versus the actual image itself. I just took a peek at an image in oversight and can't find a method to immediately remove the image using that function. ~Kylu (u|t) 00:31, 6 July 2008 (UTC)
Feel free to state your position on the renaming proposal by beginning a new line in this section with*'''Support'''or*'''Oppose''', then sign your comment with~~~~. Since polling is not a substitute for discussion, please explain your reasons, taking into account Wikipedia's naming conventions.
Offensive vandalism isn't currently a reason for deleting a revision, but suppose the offensive material appears in an edit summary in the article's history, as here? (Take a look at the article history with popups). I can see that anyone who seeks out a particular past revision of an article can't really complain if they come across offensive vandalism, but anyone who mouseovers the history link? Or should popups not preview images in edit summaries? Philip Trueman (talk) 20:08, 16 July 2008 (UTC)
Offensive vandalism is a WP:IAR reason for revision deletion. If you're discussing the article Penis, then vandalism by replacing the page with a picture of a penis may be ignorable, yet the same vandalism on Pokemon or Mickey Mouse may warrant it. In addition, many "offensive" edits lately have incorporated client-heavy encoding or features that disable the easy navigation features of our pages: Revision deletion and oversighting of these edits are understandable and expected. Popups can be configured to not preview images, but that's a user's decision, not one that rightly belongs to the vagaries of our opinion. Kylu (talk) 22:38, 26 July 2008 (UTC)
I find Raul654's behaviour objectionable
It seems to me that individuals like Raul654 should not use their position or status within the Wikipedia bureaucracy to override Wikipedia community policy on original research [4][1], or community rejection of certain procedures. I would appreciate if someone reminded user Raul654 of this--mrg3105 (comms) ♠♥♦♣ 05:27, 20 July 2008 (UTC)
It would be best to do so yourself, on his talkpage, then. If this is unacceptable, consider using the pre-existing dispute resolution system on Wikipedia. This is the wrong venue for that discussion. This page is for discussing the documentation of the Oversight right (the project page this attaches to) and, tangentially, related questions. Kylu (talk) 22:27, 26 July 2008 (UTC)
A new and exciting type of vandalism
So yesterday I discovered that User: 74.205.209.250 has come up with possibly a new way of vandalising articles. He made a non-edit (adding a space) in an article, and used the edit summary to make some personal attacks [5]. Now, the problem is that I could have deleted the revisions, but that would not have removed the edit summaries from the article histories (and since there was no vandalism in the articles themselves, I didn't bother). This person's comments will still be seen by everyone who checks the article's history. However, the edits still don't seem to qualify for oversight, based on the current rules. Do the criteria need to be expanded? Or is there some other way of dealing with this? Exploding Boy (talk) 15:48, 15 August 2008 (UTC)
I asked essentially the same question here [6] (the bad edit summary referred to there has gone now - thank you whoever), and got a very reasonable answer. And it's not new - but there seems to be more about now. Philip Trueman (talk) 16:07, 15 August 2008 (UTC)
Wiping out all contributions of a vandal-only account
I am getting tired of seeing ▄█▀█▄ █▄█ █▀ █ █▄ ▀█▀ █▀ ██ █ █ and nasty urls in the histories of articles. This is done by socks of an editor known as Grawp. There are sysops who are much better Grawpologists than myself; i just have this simple question: his edits are definitely not constructive, and some of them target articles with hundreds and thousands of edits, which may be intentional, as it makes them hard or even impossible to delete. Is it possible to automatically wipe out all contributions of an account that has been positively identified as vandal-only and which badly contaminates edit summaries, such as this one?
The Arbitration Committee is considering granting Oversight to user:Jayvdb. This is an important decision. For more information, including background, and comments by email only, please see WP:AN#Proposed granting of Oversight to Jayvdb.
I was just randomly checking some things and noticed that the oversight users list on this page is out-of-date (I think it was wrong even when the last updates were made on 27th October - [7] and [8]):
User:UninvitedCompany should be moved from the "arbitrators" list to the "former arbitrators" list
User:Newyorkbrad should be on the list on the "arbitrators" list
Looking further, I see AGK moved UninvitedCompany to the right list, but the change got lost when FT2 reverted AGK's introduction of table format. I also tracked down the point where Newyorkbrad was removed to this edit. I've checked the software-generated list at Special:ListUsers/suppress, and checked the list of Arbitration Committee members at Wikipedia:Arbitration Committee, and based on that I'm going to put Newyorkbrad back in the list, and redo AGK's move of UninvitedCompany to the "former arbitrators" list, and update the "accurate as of" date. Carcharoth (talk) 00:36, 2 November 2008 (UTC)
These changes seem fair, Carcharoth. Thanks for meticulously documenting your change, too—that's always appreciated, in the interests of good communication. AGK16:27, 2 November 2008 (UTC)
Your participation is needed! The historic first-ever CheckUser and OverSight election run by the Arbitration Committee has just started. It's taking place here. Editors are needed urgently to scrutinise the candidates so that those appointed are the best possible people for the job. Your participation here is important to make the election a success. Thanks in advance, --ROGER DAVIEStalk00:37, 6 February 2009 (UTC)
Oversight's effectiveness
Isn't it reasonable to say that if something has been up on Wikipedia for a while then oversighted later then it will be still up on Wikipedia mirror sites? --DFS454 (talk) 12:30, 6 February 2009 (UTC)
Hi! Could you please tell me which bug ID on bugzilla enabled global permission "hideuser" to local oversights? That feature has still bugs (like bugzilla:18182 or bugzilla:18185) which should be fixed first, so that stewards will be able to keep on testing before it will be enabled locally. Thanks, —DerHexer(Talk)10:49, 1 April 2009 (UTC)
Proposed policy revision
So, an IP made some nasty comments about a real person on various pages, and someone who knew about it asked for oversight. Most of the comments were not within the strict definition (although a few were). However, they were certainly deletable, to keep them out of the page history and protect the people involved. And Selective deletion has been around for a long time and is non-controversial. But this raised a new problem. The article had been deleted several times before to remove bad vandalism edits and libel. Deleting it again (1500+ revisions) to knock out 4 bad edits had the additional problem of making sure that the previously deleted bad edits were not accidentally restored. And it occurred to me, why use a chainsaw when we now have access to a scalpel?
The ultra-strict no-oversight-without-a-good-reason policy was necessary because edits removed with the Oversight Extension can not be restored--ever. So we have been properly careful and limited in how it was used. But edits hidden or suppressed with RevisionDelete can be restored, so I propose to loosen the strings on the use of RevisionDelete.
Before I continue, it is important to note how RevisionDelete actions are logged. There is a checkbox in the interface, "Suppress data from administrators as well as others". If this box is checked, the action is logged in the Suppression log. Users and admins can see that an edit was removed, but they can not see the content or who removed it. If this box is not checked, the action is logged in the deletion log--the content is still hidden from all users including admins, but you can see who removed it. It is also important to note that it is possible that RevisionDelete could someday be handed out to all admins. Admins could delete single revisions containing libel or copyright violations, without having to delete and restore the whole page. Other admins could review and reverse the deletions. Users in the "Oversight" group would have the additional ability to suppress edits so that admins could not see or reverse them. At the present time, admins can not do this, but if the "suppress data from admins" box is not checked, then at least other admins can view the action in the log and will know whom to ask for more information. It is a semi-transparent solution, and much more transparent than the oversight extension used to be.
The bottom line is, it makes little sense to treat RevisionDelete as cautiously as we treated the oversight extension. I propose to allow users with the oversight permission to use the RevisionDelete tool (but not the oversight extension) to remove edits that would otherwise be acceptable for deletion using Wikipedia:Selective deletion, provided that theSuppress data from administratorsbox isnotchecked, so that the action will show up in the regular deletion log.
Proposed change
The ability to permanently hide (using the oversight extension) or suppress (using RevisionDelete) edits, edit summaries, and log entries, is permitted under the following circumstances:
Removal of non-public personal information, such as phone numbers, home addresses, workplaces or identities of pseudonymous or anonymous individuals who have not made their identity public. This includes hiding revisions made by editors who were accidentally logged out and thus inadvertently revealed their own IP addresses.
Removal of potentially libelous information, either: a) on the advice of Wikimedia Foundation counsel; or b) when the subject has specifically asked for the information to be expunged from the history, the case is clear, and there is no editorial reason to keep the revision.
Removal of copyright infringement, on the advice of Wikimedia Foundation counsel.
In addition, the RevisionDelete tool may be used to delete vandalism and libel that would otherwise be suitable for Selective deletion in situations where:
The deletion has been requested on the oversight-L mailing list or by other communication with a user with Oversight permission, but is not strictly within the above definitions.
Posts of a kind that would be reverted or deleted by an administrator if they appeared in an edit, but require RevisionDeleted since they appear in a log entry that can not be hidden by an admin.
Deletions involving pages with large numbers of revisions. Pages with more than 5000 revisions can not be deleted by admins, and deleting pages that have close to that number puts a large strain on the database servers.
When the RevisionDelete tool is used to remove edits that qualify for Selective deletion but not Suppression, the Suppress data from administrators box should not be checked.
The oversight extension is deprecated and use of oversight (instead of RevsionDelete) should be avoided unless specifically called for under the circumstances.
Agree with the changes to the policy as stated by Thatcher. Since the action can be reversed and is semi-transparent, then expanding the wording of the policy to cover the needed (and already used) exception is appropriate. Wording the policy to match the best practices is needed to give all users guidance about when to the tool. FloNight♥♥♥19:37, 5 May 2009 (UTC)
Sounds very sensible to me. It would be good to give administrators the right to see these revisions, presuming they still can't. [[Sam Korn]](smoddy)22:15, 5 May 2009 (UTC)
I proposed similar, but more generalized, on the grounds that RevisionDelete with "no admin access" is akin to oversight, but with admin access it's similar to ordinary traditional admin delete in scope and effect. It's also a lot more easily reversible. My suggestion was therefore a brief list similar to CSD, which would allow updating by consensus as time passes:
Use of RevisionDeleted:
An edit, a field within an edit, or a log entry may be removed from public view using the RevisionDeleted tool if it falls into one of the following categories:
RD1 - Purported or actual privacy breach: personal information of an editor or third party. †
RD2 - Blatant defamation, threats, or harassment (or links to the same) against an editor or third party. †
RD3 - Hiding the IP of a "logged-out" edit where the editor has since rectifed the post "logged in".†
RD4 - Breach of copyright or intellectual property rights ††
RD5 - Posts of a kind that would be reverted or deleted by an administrator if they appeared in an edit, but require RevisionDeleted either since they appear in a log entry or because RevisionDeleted would be less disruptive or intrusive than the usual deletion tool.
† These actions may be performed with the administrative interface locked. All other usage should usually allow administrators access to review the deletion. †† may be admin locked in some cases, if based upon the advice of the the Wikimedia Foundation counsel.
The oversight extension is deprecated and should not be used. The sole exception is where complete removal of the revision is specifically called for. Typically these will be rare, and will relate either to serious harassment cases, or to edits whose removal via lesser means would itself probably foster a breach of privacy.
Is there a substantive difference between your proposals and Thatcher's? My preference would be for his presentation as its flow is simpler (no footnotes). Furthermore, my personal preference would be for this tool to be used as little as possible, so explicitly accommodating for later additions wouldn't be my preference. [[Sam Korn]](smoddy)22:51, 5 May 2009 (UTC)
Yes. Thatchers works too, but it is taking the stance of RevDeleted primarily as an oversight replacement tool. That's unlikely to be the case - the tool is designed for admins as well as oversighters to use, and will surely become a general deletion tool for non oversight matters (as Thatcher himself has anticipated).
On that basis it makes sense to design the wording not as "oversight policy tweaked" but as "deletion policy tweaked". In other words, word it from the premise it will be a general deletion tool that has some features only to be used in certain situations, rather than an oversight tool. FT2(Talk | email)23:01, 5 May 2009 (UTC)
Currently, RevDelete is an oversight replacement. We don't know when or if RevDelete will be extended to admins. When it is, then the oversight policy goes back to the original policy (that suppression should be used for limited listed cases), and Wikipedia:Selective deletion gets completely re-written. As a matter of practicality, the oversighters will still do a lot of single revision deletions, because the oversight mailing lists attracts such requests, but they would be entirely ordinary admin actions at that point. The purpose of this rewrite is to bridge the gap and allow sensible use of the tool without causing technical violations, such as the one that was debated on the mailing list today. Bearing in mind that the policy will be reverted (more or less) if RevDel is extended to admins, I prefer to keep the structure as nearly similar to the current structure (but I wrote it so of course I would say that...). I do think the point about log entries is a useful addition. Thatcher00:49, 6 May 2009 (UTC)
Partly true, but not quite. Even if admins in general aren't yet able to operate the tool, oversighters are also admins, and they can use and are using RevDel in circumstances where it's an admin level action and not an oversight action.
For example, as an admin, a user with access to RevDel might delete just the contents of a revision but not the whole revision entry in history/contribs, or might delete the edit summary only, or might delete a single revision rather than having to do a mass-delete-and-partial-restore. All of these are less disruptive than normal deletion but not oversightable.
If you look at my own logs, you'll find the same. As an oversighter I use it to remove defamation/libel, and privacy breaches, and those actions are locked against admin viewing or change. But as an admin I also use it where it's only one part of the edit that needs deletion and reasonable criteria for deletion are met but not criteria for oversighting, or when mass-delete-and-restore isn't possible.
(Even if right now the only admins with access to admin-usage are other oversighters, that won't be so in future. In future my having chosen and used that setting correctly will indeed affect who can see or modify the edit.)
So it isn't some theoretical possibility for future. Its in the clear design of the tool now (albeit we're in rollout/debug stage and it's being tested with oversighters only right now), and oversighters are indeed using it in their role as admins, for admin level deletion where it's preferable and less disruptive than usual deletion. So a policy that only views RevDel as an oversight tool is missing the point and already out of date. It is right now being used as an admin tool as well, with admin access to edits and deletions enabled (retrospective when admins get tool access) for material that is admin deletable but not within the remit of oversight. FT2(Talk | email)02:07, 6 May 2009 (UTC)
For want of a better page, I have collated here some initial findings on the use of the RevisionDeleted package as a deletion tool from its rollout around Jan 30-Feb 1 to today (just over 3 months). Since RevDel can be used both as an admin level and oversight level tool, and has been used as both, I hope to summarize for both kinds of usage, as well as Oversight (as a comparative, which RevDel is often likely to supersede).
As a first step, below is a table of Admin usage of RevDel. (I'll add suppression use data shortly.) This reflects all RevisionDeleted actions to date that were not tagged as "admin suppressed". In principle, these would be expected to relate to deletions that are effectively "normal admin deletion", ie not oversightable matters such as privacy and defamation removal, since they appear in the normal deletion logs and if admins were given access to revDel they could view them. Since RevDel actions are currently not accessible to admins anyhow (see collapse box), some oversighters took note of the distinction and checked the "admin suppression" box for oversightable material, and others didn't.
(Comment: - There was discussion of this on the functionaries-en list. It is not considered to be a formal "lapse"; rather it is an intrinsic reflection of the ongoing rollout, the current restriction to oversighters, and the initial lack of policies and norms which are developing as the tool becomes more used. In other words, mostly due to the initial teething period.)
About RevDel (if you're new to it)...
(Caveat - this is a personal understanding only)
RevisionDelete (RevDel) is an enhanced deletion method that at present co-exists with normal deletion and oversight. It allows the selective deletion of any or all aspects of an edit or log entry - ie any or all of the user's name/IP, the revision text, and the edit summary. Users with RevDel access can see the deleted information, and alter or undo the deletion; the revision itself shows up in all contributions and histories as normal but with the deleted information inaccessible to the public.
A further setting allows this deletion to be "moved up a level" and to exclude administrators as well, only allowing oversighters access to the view and modify page. A revision that is locked from admin viewing ("suppression" or "removal") in this way is still visible to all users (eg in page histories, contribs, etc), but viewing of the suppressed/removed data, and changing the deletion setting to allow admin access or reverse the deletion, is limited to oversighters while that setting is checked.
At present, RevDel is being rolled out, tested, refined, and debugged. During that process the permissions are restricted to Oversighters only, as a kind of "test group". Although admin-access is available as a setting for oversighters while using RevDel, admins haven't yet got the necessary permissions/groups added in the configuration file, to allow them to use the tool and view or modify edits marked as "admin deleted" via RevDel. The tool therefore functions more like Oversight in that all actions are effectively suppressed even if marked as "admin viewable". When the extension is ready and the config file is updated, all existing actions not tagged as "suppressed/removed" would become admin accessible and admins will be able to view and modify them via RevDel (subject to any items where suppression has been explicitly added).
Admin-level RevDel use to date
To date (Jan 30 - May 6, approx 3 months and 1 week)) there have been 382 uses of RevDel as an admin level tool in the "delete" log:
Was material oversightable?
Reason given / number of actions
Total
No/ probably no (ie, would usually be admin viewable)
threats against a user (57)
grossly offensive/gross personal attack (34)
inappropriate links (eg to malware sites or other highly inappropriate websites) (32)
personal info (often combined with harassment/libel) (162)
defamatory (including defamation posted as vandalism) (41)
personal info - user's IP (edited logged out) (10)
unspecified "oversight-l request" (2)
215
Maybe (Need to review edit to confirm)
"unsourced and potentially damaging info" (9)
"defamatory" (including posted as vandalism) (7)
16
Unknown/other
unclear from reason (13)
no reason provided (5)
test use (6)
24
Analysis by oversighter of the above:
detail...
User
Admin deletion applicable
Oversight applicable
Oversight may be applicable (check needed)
Reason unclear/ missing
Test
Alison
21
71
1
Cary Bass
1
3
4
1
Casliber
3
Coren
1
Daniel Case
16
Deskana
4
12
3
3
Dominic
5
24
1
FloNight
9
1
Fred Bauder
1
7
2
FT2
11
1
Jayjg
1
2
Jayvdb
32
22
3
1
Luna Santin
4
Mailer diablo
2
Risker
38
26
6
4
1
Rlevse
2
1
Roger Davies
3
5
3
Sam Korn
2
3
Taxman
1
Voice of All
1
Wizardman
12
YellowMonkey
1
TOTAL
127
215
16
18
6
Summary and interpretation
As well as being used in place of Oversight, RevDel is also being used as a form of admin deletion by users able to access it, both to delete log entry deletions, and also to allow less disruptive admin deletions (eg by removing the offending aspect of an edit leaving the rest visible in public history, or by selectively deleting the offending revision/s only).
The tool continues to be used quite often for removal of oversightable material ("suppression") but with the "admin view/modify" setting not selected. If admins are given access in future to the RevDel function for admin-style deletion work (where oversight is not required), there'll be a backlog of deletions that should have been suppressed, which will need to be tweaked first. There are around 215-230 of these right now, but it would be sensible for oversighters to make sure they set oversight-style removals as suppressed rather than deleted, to prevent the backlog growing.
Reasons and usage of the tool seem broadly reasonable. (I haven't yet checked the suppression log but would expect that to be to a high standard.)
There is a proportion of actions (~ 10%) where the reason or validity cannot easily be determined from the narrative supplied (though it may well be visible from the edit itself). A readily understandable reason will help others who use or patrol the logs.
These stats are only part of the picture. Lacking guidance, some oversighters always check the "hide from admins" box, so we don't know how many of them might have been more suitable for administrative deletion rather than suppression. (I am not going to check them all but I might check a random sample.) What we do know from your stats is that RevDel is being used for deletions that are not strictly under the current criteria for oversight, and that admin suppression is being used inconsistently. Hence the need for a revision to the Oversight policy. Thatcher11:15, 6 May 2009 (UTC)
Nod. The good news is that unlike the old Oversight, which took an act of BRION to undo, a review of all of these can be carried out and corrections can be made, so the fact that everything wasn't gotten right from the get go is fixable. I for one am glad that the error seems to be mostly on the side of being more cautious than needed, rather than less. That's the right direction to err, in my view. ++Lar: t/c13:18, 6 May 2009 (UTC)
It is very easy to mistakenly forget to check a box, and then go back and fix it. In the meantime, that will have been logged in the normal deletion log. If I understand correctly, this list isn't distinguishing between true deletion and deleted revisions that were subsequently fully suppressed? I think nearly all of my uses of the tool have been for suppression, and I have several times had to fix ones that I saved too soon without checking the "suppress" box. I certainly haven't been using the tool for much of any admin-level deletion at this point. Dominic·t14:00, 6 May 2009 (UTC)
FT2, I am concerned that, when analysing this information, you may be using a different standard of what should or should not be oversighted than others with oversight permissions. Since this cannot sensibly be discussed on this page because of the nature of the edits. Can I ask that you please send your full analysis to the Audit Subcommittee, particularly including what actions you sorted into which category? Thanks. Risker (talk) 18:59, 6 May 2009 (UTC)
One of the running issues is the relative strictness of interpreting the oversight policy. It appears that FT2 is saying that a lot more edits should have been oversighted/suppressed than actually were - the table above addresses only those edits that were revision deleted, not those that were suppressed. For those of us actively using the tool, we need to know what he thinks should have been suppressed that we didn't think should have been at the time we did the revision deletion. It may be that his interpretation of the policy is radically more liberal than that of other oversighters. Risker (talk) 02:58, 7 May 2009 (UTC)
The purpose of this summary is to look at what RevDel is being used for, when it's not flagged as "admin locked". That's for three reasons:
Checking that accidental disclosure of oversightable material is not a problem building up - To see whether privacy breaches and defamation are inadvertantly being hidden in a manner that admins might see in future, when under the old system they would usually have been oversighted.
Documenting for the community and to inform the policy debate - To document for the community what kinds of things RevDel is being used for, when the action is not performed as an oversight equivalent and the deleted data is not permanently hidden from admins. There is discussion on a potential policy and the usage the tool might get, so it's useful to see how oversighters are presently using the tool to inform the discussion. We find it's often being used in "admin viewable mode" to remove non-oversightable but clearly deletable material.
Scrutiny of usage - At this time deletions flagged as admin viewable, are not admin viewable. They are also not subject to review by other oversighters using the suppression log. They are buried in the delete logs which have thousands of entries a day. They are also not subject to the strict criteria of WMF oversight policy. So it's worth confirming for the wider community whether the tool is being reasonably used, even though no formal policy yet exists.
The standard I've used in compiling this data is the logged reasons themselves, and the usual standard for deletion vs oversight. If a deletion gives a reason, then any admin can in principle say whether that reason would prima facie be sufficient for oversighting or not. The reason should be sufficient to explain a delete action; if the reason is misleading that's a whole different question. For around 90% of the logged actions there is a clear stated reason. This will either relate to a breach of privacy or defamation or not.
The table shows the 382 uses of RevDel that appear in the ordinary delete log, sorted by the reason given. Where the reason is unclear, that's noted. If the reason indicates a deletion took place to protect a person's privacy, personal information, libel/defamation, and the like, then the presumption is that it's more likely on the surface to have been an oversightable matter (ie any oversighter asked to remove a person's name, phone number etc would be fine oversighting it). These 382 revisions will all be retrospectively visible to all admins if admins get single revision delete in future, unless corrected, so it's quite important to save future problems, that the "prevent admin viewing" box was correctly checked. FT2(Talk | email)15:14, 7 May 2009 (UTC)
I have to say, I disagree with your reasoning to a certain extent. Every edit must be assessed on its own merits. There are different levels of "personal information", some of which qualify for suppression; some of which do not. Nevertheless, it may be appropriate for some of those in the latter category to be deleted anyway. I think assuming a priori that all revisions deleted as "personal information" should be hidden from admins is potentially misleading. Yes, we need to have some review; no, I do not think relying upon summaries alone is the right method.
I am inclined to be more conservative with this tool than the average, I think. I feel strongly that we should be as open as possible. That is to say, we should prefer leaving revisions fully visible wherever we can; by first preference we shouldn't hide from administrators. Now, I'm aware that we don'ttrustadministratorsabsolutely; nevertheless, hiding revisions from them should be an absolute last resort.
One would not assume automatically that every "personal data" should be removed to oversight levels, that's quite right. But log reasons such as "Nonpublic personal information", "Attempted outing", "Full name of an editor, non-public, used for harassment", "Potentially libelous statement about third party" are classic oversight reasons, and it seems likely that a significant number of actions are thought of as being removed to oversight levels but haven't been. It's worth checking and awareness.
But beyond that, the point of the exercise isn't to identify that "X many items should have been oversighted and weren't", even if that's what some have drawn from it. An email to the mailing list would do that job just fine, if that was the intended purpose.
The purpose here is to simply document usage, because 1/ this is a new and significant tool for most users; 2/ there is (or will be) public debate about appropriate usage in "admin access" and "admin lock" modes; and 3/ there is almost certainly zero scrutiny of the deletions that are flagged as "admin level" nor any guiding policy, even though these revisions are in fact deleted from the view of all users and admins except oversighters. That's 3 good reasons to more fully explain the tool, and summarize it's actual use (when not used as an oversight replacement) to users in the wider community, now that it's had some 3 months of rollout. FT2(Talk | email)19:34, 7 May 2009 (UTC)
I do, of course, generally agree with all this. Nevertheless, your presentation strongly implies "these actions were mistakes", and, while that may be true in some cases, I do not feel it is universally the case. I feel your method is flawed and as such will not give the enlightenment that you seek. Of course I applaud your effort in producing this summary, but I question what I feel is the rather leading nature of its presentation, which appears to favour a policy position that I personally do not hold in its entirety. [[Sam Korn]](smoddy)19:47, 7 May 2009 (UTC)
The unfortunate situation is that currently, all items hidden with the RevisionDelete tool are suppressed from admins, whether or not the box is checked, but checking (or unchecking) the box changes how the change is logged. But, we anticipate that someday, RevDel will be available to all admins, meaning that events logged in the deletion log will be visible to admins. These leads to a very important question, Are there suppressed edits that are logged in the deletion log because the box was unchecked, that will become visible to admins someday, that need to be re-suppressed? And it would not hurt any oversighter to page through their deletion log and see if anything needs to be re-suppressed. That can be handled separately from other issues. Thatcher20:58, 7 May 2009 (UTC)
Agree - that's one point arising. Because of the sheer volume of the deletion log, sifting the full deletion log is not easy. A log of an individual oversighter's deletions back to late January would be practical. I've sent a spreadsheet showing all actions that would be admin-visible to the functionaries list for oversighter attention, to make it easy for oversighters to review RevDel from commencement up to end April.
If it was intended for any revision on that list not to be admin visible then that revision will need to be re-deleted with the box checked. As Thatcher says, that's separate though. The main purpose of this table is more to ensure we have useful communal information on the use of RevisionDeleted -- if we're discussing a suitable RevDel policy, it helps to have a good idea the kinds of issues on which the tool's being used. FT2(Talk | email)21:47, 7 May 2009 (UTC)
I had made back in March a bug request for a special hide/suppress log, T19806. That would be very useful right now, and necessary if the non-oversight permission is granted to admins. The difference between oversighting and deleting should be very clear in the interface, and expressed in the log too. Cenarium (talk) 00:13, 9 May 2009 (UTC)
Actually, the process that oversighters are currently using to hide revisions (single revision delete) is publicly viewable now and any such edits are marked. It will likely become the norm for all admins to use in the future, so I am not sure there will be a need for a separate log; the actions show up in the "deletion" log now. The edit suppression button that is associated with it, however, is intended as the replacement for oversight, and the suppression log is not publicly accessible any more than the oversight log is now. Risker (talk) 00:40, 9 May 2009 (UTC)
That's precisely the problem, the deletion log is way too large and those actions are completely lost there. It needs to be separated for analysis, transparency and review purposes. Cenarium (talk) 00:45, 9 May 2009 (UTC)
Fair enough, I'll leave it to the devs to decide whether or not it's needed. I don't see it as being any less transparent within the deletion logs than separate from them; I suppose there is just enough difference in the function of the tools to make a case for a stand-alone log. Risker (talk) 00:56, 9 May 2009 (UTC)
I don't understand the need for this. Once RevisionDelete is made available to admins, they will have the ability to delete single revisions of articles and hide edit summaries from regular users, but other admins will be able to see and restore them. It does not strike me as being any different than the deletion log now. That is, if I wanted to hide an edit because the edit summary is offensive or the content is bad vandalism, I could delete the whole page and restore all but the bad edit, or I could singly delete the bad edit. Either way it goes in the deletion log and I see no reason to log such actions in a separate place. The additional ability to hide edits from admins as well as regular users will only be available to the oversight group and is listed in a separate suppression log. Thatcher01:11, 9 May 2009 (UTC)
First, from a technical point of view: the actions are of a completely different type, one is about deletion (and undeletion) of an article, the other about visibility of a revision. The name revision deletion is slightly confusing in this regard (it can already be tested for admins on testwiki, by the way). I suppose that it was easier to log those actions in the existing deletion log than creating a new one, but this can still be done.
If the new system is made available, the practice known as selective deletion would surely be deprecated for that purpose, probably frowned upon. If the hide log is not separated, it will not be easy to find the actions to hide (to any level) revisions, thus decreasing the transparency, ease of analysis and review of the system. As for the suppression log, what kind of data does it contain ? isn't it restricted to oversighters ? Cenarium (talk) 02:43, 9 May 2009 (UTC)
Then if developers have created a - private - log for suppressions, I suppose they won't object to create a - public - log for admin-level hide actions, it looks natural. The terminology is somewhat confusing ... Cenarium (talk) 03:19, 9 May 2009 (UTC)
(unindent) I've commented at bugzilla (#17806). Basically, once RevDel is at least read-accessible by all admins (even if they can't modify it) then there is probably no need for a separate log.
You might not be aware fully how RevDel works. In brief, it's more (not less) visible than usual deletion. A revision will still show up in contribs, or page history, which it doesn't for deletion, and any user let alone any admin can see it exists. What revDel does is strike out various fields, ie, it won't show their contents to someone who isn't (in this case) an admin. For admins, a link appears next to every revision, allowing them to see the hidden data and check what was removed from public view, and that the removal was reasonable. That link is visible in every users contributions, every diff page, every page history, for every revision, so any admin can see any revision whenever they want, and check it was reasonably removed or what it contained.
Compared to usual deletion, it's more transparent. Anyone can see that something's been removed (usually only admins can), and it's clearly visible in the contributions and page history (usually deletion isnt shown in sequence and you have to click separately for any deleted contribs or revisions). FT2(Talk | email)09:51, 9 May 2009 (UTC)
I already had the occasion to test revision deletion on testwiki a few months ago. The matter here is (public) logging, not admin access to the hidden content. Every user can see those actions in the deletion log, but they are literally lost among deletions, so it's difficult to find them as a whole, and analyze and review them, live. Let's say, you can't review what you can't find... Cenarium (talk) 13:02, 9 May 2009 (UTC)
Agreed but with the caveat, that it's a very temporary problem. It's only a problem because admins can't view them even if they want to, and oversighters don't patrol the usual deletion log per se. Once admins have the ability to review them, then it won't be an issue - there are thousands of admin level deletions but the important thing is every one of them can be viewed or reversed at will by any of almost 2000 admins. The situation where these can't be, is likely to be temporary until revDel is fully ready for admin use. FT2(Talk | email)17:29, 10 May 2009 (UTC)
The problem is not temporary. The only way to find an admin-level hide action (essentially, for review purposes) is to search the deletion log (or an admin's, or a pages's, when we already know where it has been done or by who), but it is too big to be parsed properly. This will also be the case if/when admins can view admin-level hidden content. So the need for a split. Cenarium (talk) 19:12, 10 May 2009 (UTC)
Incorrect. You seem to be presuming some special need to identify and patrol these deletions separate from other admin deletions, but no such need exists. Does the fact there are thousands of deletions a day bother you? No; they are patrollable, reviewable, and if a query arises, any of them is checkable by any admin who can verify what it contained. When admin access is enabled these edits are more publicly visible (not less) compared to existing deletions, and compared to existing deletion the full details will be completely equally viewable and checkable by any admin with equal ease. In fact much of the reason RevDel isn't yet admin accessible is to allow for refinement of the deletion system, ensuring logging is easily trackable, the interface is convenient, and so on. FT2(Talk | email)19:44, 10 May 2009 (UTC)
Those actions are not deletions, from a technical point of view, it's entirely about the visibility of the content of a revision (the name revision deletion is misleading), or of a log entry (even less related to deletion). They are logged in the deletion log because it's easier than creating a new one and, albeit remotely, related (because deleting has also the consequence to make the content of revisions not visible, except to admins). So separation would be justified on a technical level. From a practical point of view, making the logs separate is beneficial as most admins who would be interested in reviewing those actions would not be, or to a much lesser extent, interested in reviewing the massive amount of deletions. Only a few of them are actually checked by other admins, and at times, many are automatic, thus cluttering the log. Trying to find at least one of those actions in the deletion log would be akin to trying to find a needle in a haystack. The fact you seem to elude is that this matter is controversial, more so than deletion, or on a different level, and making the review of those particular actions as easy as possible will be necessary if we want to convince the community to enable this for admins (which is not guaranteed, in view of the discussions on this). Cenarium (talk) 00:45, 11 May 2009 (UTC)
Again incorrect. Redacted edits are logged in the delete log because once admin access is enabled, the effect is identical -- an edit by a user is partly or completely not visible to the public but only to admins, and their removal from public view and reinstatement is also under admin control -- and proliferation of logs isn't desirable for essentially the same kind of subject matter. One log for "edits fully removed from view" and one log for "edits partly removed from view"? Pointless. It's hard to see why admins being able to strike out an edit summary (say) would be controversial when their ability to strike out the entire edit is not. It's hard to see why their striking out in a way that any user can see the struck out material would be controversial, when striking it out in a way nobody except admins can see, is not. Can you explain why you feel that admin striking out of (say) an edit summary would be controversial when full deletion is not? Or how improper redaction would be more difficult to spot in a page history or contribs or delete log than improper deletion? I'm afraid I'm not seeing it just yet. What I do see on the discussion you linked to is that we need a policy on usage, but that's a given and exists for all admin tools. FT2(Talk | email)08:53, 11 May 2009 (UTC)
No; the effect is absolutely not identical, when you delete a page, it becomes a redlink, history cannot be accessed, it's detected as a non-existent page by the software, etc. Hiding a page revision restricts the visibility of the content, summary or username. But the revision still appears in the history, the diff link still works, etc, it's not like so-called selective deletion. You can also hide a log entry, for example an entry in the move log, but you cannot delete the move log... We're not talking about creating many more logs, just one for a type of actions that is very different from deletion, and which has an interest on its own. It should be obvious why it can be viewed as controversial... This is discussed in the linked discussion. Cenarium (talk) 12:01, 11 May 2009 (UTC)
Sorry, I'm still not seeing any "big deal" about it. A need for communal agreement on appropriate usage, yes, but this tool is far less capable of abuse than existing deletion, and yet affected revisions are more (not less) visible. So its potential to be a problem would strongly seem to be much less. If normal deletion is not a problem, then this will be even less of one. If you disagree, can you give me a realistic scenario (assuming "admin access" was enabled) where deletion of say an edit summary of revision text this way would be more of a problem, or with less ease of scrutiny, than a normal deletion of a revision by an administrator? FT2(Talk | email)12:34, 11 May 2009 (UTC)
Final call for voting in Checkuser/Oversight election
Removal of vandalism. Suppression may be occasionally used to remove vandalism that can not be removed by normal administrative measures. Such cases should be handled with suppression, rather than with the Oversight tool, so that they may be reversed if needed, and should be discussed in advance on the Oversight mailing list unless they are urgent or time-sensitive.
This is already happening. "Policy is what people do, then eventually write it down." Either it is acceptable to use suppression to remove certain kinds of vandalism that can not be removed using regular admin tools (pages with >5000 edits, certain other types of vandalism), in which case we need rules on how, when, where and why; or suppression should not be used for vandalism no matter what, in which case there needs to be a strong consensus on this page which is communicated to Arbcom and the Oversighters. Leaving the policy as was (stating it was never used this way when it sometimes is) is unacceptable either way. Discuss.Thatcher01:37, 12 August 2009 (UTC)
If we're going this route, I'd strongly prefer that we keep a tight, tight, tight leash on it -- this criterion, more than any of the others, strikes me as dangerously open to interpretation and creep. As you mention, it already happens, in some rare circumstances; might be handy when faced with a rash of users restoring a specific oldid (think of 4chan rushes). If some clear precedents emerge, this might grow on me, but in general I turn a stern eye toward any expansion of oversight's remit, as a matter of principle. The current wording allows for suppression without prior discussion if the matter is "urgent" -- if the matter really requires oversight, it might always be urgent -- would it be prudent to require or recommend prompt reporting of such action to the list, for discussion? – Luna Santin (talk) 04:54, 12 August 2009 (UTC)
Sure, discussion beforehand if not urgent, and prompt reporting and discussion afterward if urgent. Although I tend to think that vandalism which does not reveal personal information or contain potential libel is unlikely to be urgent, and this is certainly the case for at least two specific examples I am aware of that were removed with Oversight or Suppression. What I would like to make clear is that if we are going to occasionally act in this way it must be discussed and suppression must be used, not oversight, in case the discussion decides that it should be reversed. Thatcher11:34, 12 August 2009 (UTC)
For the record, the majority of "suppressions" along this line are revisiondeletion of inappropriate edit summaries where the edit itself may or may not be left in place (e.g., "F***you idiot!" which, while disruptive and jarring to a reading audience, don't meet the oversight criteria, and most of which are only revisiondeleted but not actually suppressed), removal of otherwise admin-deletable edits on pages too large for admins to carry out deletions (rarely fully suppressed), and hideusername actions (the use of this tool always fully suppresses). The last is often carried out in concert with stewards managing crosswiki mass vandalism issues. I'm not really sure all of these situations need to be discussed, either in advance or afterward, on Oversight-L; I dread to think how many hideusernames I'd have to bring up after a bout of mass crosswiki vandalism, as I've sometimes hidden 40 or 50 at a crack. There are several bugzillas outstanding related to revisiondeletion; at least one that I've signed on to would split the view access so that at least admins would be able to see the revisions that have been "deleted" but not fully suppressed. Risker (talk) 23:29, 12 August 2009 (UTC)
I admit I'm curious whether such actions would be allowed under the oversight policy as stated at the time; this is one of a small handful of policies I personally consider to be largely prescriptive in nature, which goes against my usual approach. I know that many of those names included oversightable information, but a large number of others did not -- yet, they continue to be suppressed, apparently within the bounds of consensus, so I think this proposal aims to correct that apparent discrepancy between policy and practice. As those usernames tend to come up in batches, I don't think discussing them in batches would add too much volume (though I'd understand if others disagree)... if it's problematic, perhaps there could be some very specific precedents in which discussion is not so necessary. – Luna Santin (talk) 00:01, 13 August 2009 (UTC)
We probably don't need to discuss every case of harassing user names, as there is already consensus on oversight-L that such suppressions are OK. But it should be documented here, and people given a chance to object and discuss on this page. Other forms of vandalism should probably be discussed. Thatcher02:05, 13 August 2009 (UTC)
The hideusername actions are in the logs and can be reviewed there, if needed. Aside from having oversight-L discussions about the rest of the stuff, perhaps it would be more useful to vote for the applicable bugzillas so that they can come under the review of non-oversight admins. I'll try to pull up my archives there and post the bugzilla number here. Risker (talk) 02:25, 13 August 2009 (UTC) Here are the relevant open bugzillas (the second one has a list of related bugzillas at the bottom). This is a good time for people to identify any other issues. Risker (talk) 04:11, 13 August 2009 (UTC) Thanks to MZMcBride for finding these for me.
As I've said internally, I think that using the tool to remove highly offensive vandalism that it is not possible or practical to remove otherwise is fine if it is done with suppression. The same stuff that we have discussed on Oversight-l and agreed was fine can now be added to policy on site because it will document what we can do now that we have the newer suppression tool. FloNight♥♥♥02:18, 13 August 2009 (UTC)
Although in practical terms there isn't a massive difference between supression and oversight, I fully support the removal of things like offensive edit summaries using the supression tool, as has been happening before this addition to the policy was made. --Deskana, Champion of the Frozen Wastes03:14, 13 August 2009 (UTC)
First, "Policy is what people do" -> "Policy is what the general community thinks is best". Second, what kind of vandalism are we talking about here? Just edit summaries? Do this only if they expose private information. Defamatory comments in edit summaries may be not sufficiently visible. If it is entire versions of pages, don't do this, it screws up vandalism stats tools. M10:38, 13 August 2009 (UTC)
It's not even possible to supress every revision in a page. And frankly, if there's offensive edit summaries, I don't think anyone cares if "vandalism statistics" are a little off. I certainly don't. Also, though the oversighters ask for input from the community, it is ultimately their decision what they supress and what they do not, and ArbCom can remove them if they disagree with their actions. So, telling us what to do and not is not a constructive way to get things done. --Deskana, Champion of the Frozen Wastes15:09, 13 August 2009 (UTC)
Some are vandal user names that are gross, disgusting and harassing, but do not reveal personal information. These vandal user names are blocked as soon as they are created, but suppression can be used to remove them from the user creation log. Another recent example is user:Wit-o-pedia, who constructed a message using his contributions (several edits have been removed so the message does not make much sense any more). To remove the message, or at least to hide the parts that made it personal vandalism and harassment, requires either deleting or oversighting the individual edits, and deletion only works on pages with less than 5000 edits, and has other complications. There are also cases of gross, offensive and harassing page moves. Entire pages are not suppressed unless they qualify under the personal information, libel or copyright clauses. Thatcher15:22, 13 August 2009 (UTC)
The word "offensive" is awfully vague, which is going to make discussion of it difficult and enforcement of it worse; would it behoove us to get more specific, here? Maybe offer some more examples? I don't think oversight was ever intended to be used on vanilla vandalism -- aside from the ever-increasing amount of work it would create for us, there's no clear stopping point. I don't know about you folks, but I sure as heck don't want an unlimited license to hide any edit I just don't like. Display-breaking names and malware links are one thing, perhaps, but really, what's the enduring harm to the project if somebody says "fuck" in an edit summary? Or even if a string of edit summaries creates a cute sentence which happens to include some bad words? Is it really so dangerous that we need to abuse page history to hide it? It could be that I'm misunderstanding the intended scope, here, but if I'm having the slightest trouble following this, just imagine how people without access to oversight-l will read into it. – Luna Santin (talk) 19:40, 13 August 2009 (UTC)
Well, I can go through the logs, but most people can't. :) Those examples you mentioned are quite helpful for the sake of discussion (thanks!), and that was about the caliber of thing I had in mind when this talk started. I've realized that I approach those from more of a "harassment" angle than an "offensive" one: they mention and attack specific people, again and again, in such a way that those people might well feel harrowed. The way discussion was going, I was concerned we're leaving ourselves open to eventually suppressing more harmless names like "User:Joe is dumb" and the like -- thus my concern over interpretation of the word "offensive". Sounds like we're more or less on the same page, in retrospect, but hopefully there's no harm in checking. – Luna Santin (talk) 20:20, 13 August 2009 (UTC)
But "harassment" is not part of the big 3 original criteria. The important thing, I think, it to make sure that these things are done with Suppression and not Oversight, and that the oversighters talk about it, either before or after, to work out the boundaries. And the written policy should reflect this. Thatcher20:35, 13 August 2009 (UTC)
True, but given the new fourth criterion, I think it would fit in. Where harassment is involved, I'd be that much more inclined to consider suppressing some bit of vandalism. As far as the new addition, I'm in support as of a few comments ago -- I just realized I forgot to say so, in my "let's be careful, now" dogma. Just feeling out the edges to hopefully make early guidelines somewhat clear, both internally and for the sake of public discourse. – Luna Santin (talk) 20:55, 13 August 2009 (UTC)
Luna Santin, we are not stretching the written policy to do more that what we have discussed in private as functionaries and agreed was sensible given the limits of the software and the manpower needed to clean up the offensive vandalism. By making the changes to the policy on site, I want to give us the freedom to get rid of the highly offensive vandalism that includes names without dwelling on whether it is outing or harassment. Since many of the user name vandals make large attacks, the easiest way is to clean them all up at once without handwringing over whether the named person is real and is going to feel harassed, or whether MrVandal is playing games with us. Also, with a large number of new people with the tool after the election, I think that it is good to give them clearer polices. And we need to emphasis that other uses need to be discussed on the mailing list PRIOR to doing them so that we can stop the tool from being routinely used in ways that were never intended without understand why it is happening. We want to keep the policy as tight as possible but meet the needs of the Community as new problems arise. FloNight♥♥♥21:13, 13 August 2009 (UTC)
(Outdent) I just came across Luna's advert for outside eyes. I must say that this strikes me as so sensible as to be a no-brainer. I fully agree that routine vandalism should not be grist for the suppression or oversight mill, but I don't see that anyone advocates that. The type of vandalism I see as being implicated here, both within the proposed change and in current practice, is the kind that specifically seeks to take advantage of seams in our deletion policies and technological limitations to leave a lasting mark. Suppression seems to be a sufficiently fine-edged tool for those cases when normal deletion isn't up to the task. XymmaxSo let it be writtenSo let it be done16:29, 17 August 2009 (UTC)
Clarification to this section
Suppression should not used for material that's merely "grossly offensive". Administrative deletion should be enough, where no actual personal information or plausible defamation is involved. As I understand it, the reason for allowing suppression is purely that we don't have any better tool at admin level, and hence this criterion is temporary until (if/as/when) a better tool is available for admins.
Note that it's implicit that it will become redundant if a better admin tool did exist ("...vandalism that can not be removed by normal administrative measures...") and reversal is implied if suppression is no longer necessary ("... so that they may be reversed if needed...") but the policy presently reads that it's an enduring addition. I have added a brief note to ensure this is clear.
Several months ago, I drafted a policy for the use of the revision deletion function for administrators on the English Wikipedia. After consultation with a small group of users, I made modifications and changes (with the help of FT2) to better address the suggestions of these people. I have waited a while for the policy to become more stable before consulting the wider community, because it is in my belief that there is nothing worse than discussing a policy draft that still does not have the consensus of its drafters.
The policy in its current state is quite similar to the Criteria for Speedy Deletion policy, in that it defines very specific circumstances in which the revision deletion functionality can be used. The policy is defined so strictly to help allay some of the fears of potential misuse of the functionality, with deviation from the set criteria resulting in whatever sanctions are decided upon by the community.
I would invite all users to read the statement and FAQs that I have written at Wikipedia talk:Revision deletion#Community consultation regarding the functionality and then discuss on the talk page the merits of ratifying this policy, and subsequently enabling the feature for administrators on this project.
Thanks for listening and happy editing! ~fl05:45, 26 September 2009 (UTC)
Three Audit Subcommittee vacancies: Call for applications
The process to appoint the three non-arbitrator members of the Audit Subcommittee is underway, with the election itself starting on 30 October. If you think you may be suitably qualified, please see the election pages for the job specification and application arrangements. Applications close 22 October 2009.
I've just noticed this happen, though all three edits made by the user in question were reverted in such a way that their name was mentioned in the reverter's edit summary. Should something be done about those? Miremare18:49, 12 October 2009 (UTC)
Three Audit Subcommittee vacancies: Urgent call for applications
The process to appoint the three non-arbitrator members of the Audit Subcommittee is underway. If you are suitably qualified, please see the election pages for the job specification and application arrangements. Applications close 22 October 2009.
Audit Subcommittee elections: Urgent! Final call for applications
Time is rapidly running out. The closing date for completed applications is 23:59 (UTC) 22 October 2009. If you are interested in becoming one of the three non-arbitrator members of the Audit Subcommittee, see the election pages now for the job specification and application details.
Criterion #4 allows extreme vandalism to be removed from logs via suppression.
At that time, RevisionDelete (equivalent to suppression) was the only tool available to remove egregious vandalism in usernames or log entries, where admin tools wouldn't work.
Vandalism in log entries per Criterion #4 can still only be removed by oversighters (eg on admin request), but now the deleted items can be easily seen and discussed by all administrators in the usual way, they don't need suppression.
The only area of Criterion #4 where suppression is still needed, is username vandalism. Alone of all the RevisionDelete functions, "hide username" is now the only one that doesn't have an admin-level version.
Criterion 4 can now be narrowed right down to this one area:
4. Removal of grossly offensive or disruptive usernames. Suppression may be occasionally used to remove grossly improper usernames of little or no project value, that cannot be removed by normal administrative measures, and where renaming is not immediately available. (Note: This criterion is an interim measure, due to limitations of present-day administrator tools.)
Everything else can be handled by admins, or at worst by admin-viewable actions, and doesn't need suppression.
Presumably with a few bugs in the hopper getting resolved the right way (hopefully) even this part can be narrowed away, I think... once hide username has an admin-level version. ++Lar: t/c17:01, 27 October 2009 (UTC)
The stewards routinely hide user names that are merely harassing of other editors, even if they are not gross or vulgar, as well as user names that are mildly vulgar but which attack no particular user. This policy holds local oversighters to a much stricter standard. That's fine if you want, but the looser standard applied by stewards should be noted somewhere, either here or at meta. It is de facto policy because it is being done every day, no matter what the written policy says. Thatcher17:04, 27 October 2009 (UTC)
Concur with Thatcher. It's my observation that a significant number of the hideusername actions on this project are carried out by stewards, usually addressing global block issues. On a fairly regular basis, our own project's oversighters are working in concert with stewards to deal with cross-wiki issues. If a username is being hidden globally, why would we retain it here, and why would we deliberately write a policy that places our oversighters at risk of sanction when meeting global standards? As Thatcher says, de facto policy may be more liberal than some may like, but there's something silly about placing our own oversighters at risk of sanction for tool use and then having a steward carry out exactly the same action. Risker (talk) 17:24, 27 October 2009 (UTC)
Yes... but, there's a proposal on the strategywiki to review this policy and make global changes to align with current Steward practice (by some stewards anyway) to make allowances for hiding (at least from non admins) usernames that are merely vulgar or insulting. en:wp policy should align with that. I'm not sure I see the confict between that and narrowing the scope of this though. ++Lar: t/c17:55, 27 October 2009 (UTC)
Reality
The past policy said only that edits which revealed personal information or libel were suppressible. So "Thatcher is a pedophile" is suppressible under the old policy but "Thatcher is an asshole", "Thatcher takes it up the ass" and "Thatcher picks his nose" are not suppressible, even though they are all clearly intended to attack, harass or annoy Thatcher. Recently someone suppressed User:ZSCOUT370 is NAPOLEON DYNAMITE which, while probably intended to annoy or harass, could not remotely be considered personal information or libel. Also suppressed was User:Zscout370, there's no point in blocking these unused accounts, which is just plain trolling (and 4 year old trolling at that). Someone also suppressed User:Baby covered in semen, which is not an attack on any user. If it is current practice to suppress of any user name that is intended to harass, annoy, irritate, or troll another user, as well as user names that are offensive without attacking a specific user, then say so. Vandal user names, including but not limited to those that attack, harass or annoy other editors, may be hidden on request.Thatcher14:25, 30 October 2009 (UTC)
I'm coming around to thinking that this is where we should be. But in any case I don't think we should be holding people to stricter usage standards locally than globally if we can't require local-global congruence. ++Lar: t/c20:38, 30 October 2009 (UTC)
No comment on that specific case, but there is precedent to this: the checkuser policy on enwp is more restricted than the global one. I don't think that ever caused significant problems? — Coren(talk)12:49, 11 November 2010 (UTC)
Urgent! Last call for votes: AUSC October 2009 elections
There's only one day to go! The Audit Subcommittee election,
using SecurePoll, closes at 23:59 (UTC) 8 November. Three community
members will be appointed to supervise use of the CheckUser
and OverSight tools. If you wish to vote you must do so
urgently. Here's how:
The old criterion #2 for oversight that allowed removal of defamation if it was "upon request", has been modified at Meta. The criteria for removal stay the same but there is no requirement in the policy for a formal prior request.
This reflects best practice for many years and across many wikis, where oversightable material is removed on sight, not merely when the subject asks (if they do) a long time later. The proposal was passed without dissent, checked with WMF, the global policy changed, and the local policy updated to match.
There is no practical effect of this, for this project, because it hasn't been a norm to hold back oversighting on this wiki until a formal request is made.
This is a reminder that voting is open until 23:59 UTC next Monday 14 December to elect new members of the Arbitration Committee. It is an opportunity for all editors with at least 150 mainspace edits on or before 1 November 2009 to shape the composition of the peak judicial body on the English Wikipedia.
IMO, "Username or IP removed" by itself is unhelpful. My first question was, "why was it removed?" I had to guess and hunt to figure out I was looking at an oversighted accidentally revealed IP address. Wouldn't a link to this policy in the edit summary be helpful? --Geniac (talk) 17:16, 23 January 2010 (UTC)
Gah!! :( Sorry about that. We could bug one of the devs about it, maybe, but Mediawiki isn't going to let us fix this - Alison❤19:47, 24 January 2010 (UTC)
I would like to see WP have a Reversibility policy: any procedure, tool, or bot that can edit, delete, or revert edits in articles, logs, etc., must be reversible. For example, suppose someone edits an article to include a private cell phone number for the President of the USA. Clearly, this should be expunged immediately, whether the number is genuine or not, just in case it is genuine. There should be an easy-to-find link, button, or page that allows anyone to request emergency intervention, then any of a dozen or so trusted individuals should have to power to edit out the phone number (even if the page is fully protected). Suppose that later it turns out that publication of the phone number was deliberately requested by the President in a public telescast. The same small group of trusted people should have the ability to reverse their emergency action. For any apparent malicious action, there might be a subsequent reason to reverse. For this reason, all urgent edits, deletes, and reversions must be reversible. David spector (talk) 02:31, 31 January 2010 (UTC)
Proposal for new 'public interest' clause
This is a proposal for adding an extra category of instances in which oversighting and/or selective deletion may be used:
Removal of non-public information that could cause significant harm to a person, property, or endangered wildlife.
To invoke this clause, the following three criteria would all need to be satisfied:
a piece of information has been added to WP which does not appear in a reliable source
there is a serious risk that the presence of that information on WP may lead to significant harm to a person, property, or endangered wildlife
the potential harm would not be in the public interest.
Such a clause would be helpful in the relatively rare instances in which the publication of information is not illegal but is clearly not in the public interest, and is not covered by an existing clause.
A recent instance on WP prompted the drafting of this proposal and illustrates the kind of situation in which it might be useful. The location of an orchid which is extremely rare in Britain, with just one genuinely wild plant left, was added to a couple of articles. This information has always been kept out of the public domain by the relevant conservation bodies, and had not previously appeared in a reliable published source. The plant is accepted to be at risk from illegal plant collectors, while any seedlings it produces would be at risk from trampling by careless visitors. Thus there was a risk that publication on Wikipedia could have led to the extinction of the species as a native British wild plant.
There are likely to be few circumstances in which this clause would be justifiably invoked. If preferred, wording could be added to make this clear, and to make it clear that a high level of proof would be needed for a request for oversight/selective deletion to be accepted.
As a counter-example to the above, there is currently another rare British orchid whose location details appear on WP without a reliable source been cited. However in this case there are multiple plants at three different locations, and so the loss of one or two would not be catastrophic. Thus while removal of the information is justified unless a reliable source can be found, the potential harm may not be serious enough to warrant selective deletion.
Such a clause would both safeguard the person, property or wildlife concerned, and safeguard WP from accusations of being responsible for causing such harm.
This proposal goes beyond information about endangered wildlife in order to allow for any other 'public interest' cases that might arise. However a case could be made for a narrower clause, or series of clauses.Jimi 66 (talk) 22:13, 16 February 2010 (UTC)
I support what this proposal is trying to do. The wildlife issue is a good point -- we would not want to be found to have been instrumental in the extinction of a species. Harm to property is also a good point to raise -- if some joker posts the security code for a building or access details for a corporate computer system on Wikipedia, it would make sense to oversight such an edit. These cases don't seem to be covered in the present Wikipedia:Oversight#Policy listing. --JN46622:40, 16 February 2010 (UTC)
I also agree that we need to give this matter serious consideration. Wikipedia is not censored, but when we originate inappropriate and damaging information we are being irresponsible if we do not remove it completely. SilkTork *YES!09:52, 17 February 2010 (UTC)
I would support what this is trying to do, but have a big objection to the wording. The "public interest" clause isn't needed and is in fact the biggest problem. It would require Oversighters to judge "public interest" which is something they are not qualified to do, and sets a precedent that can be railroaded by anyone with a claim that this or that "isn't in the public interest" in future. Also the demand to "not be in a reliable source" is a problem, because the moment the orchid's location appeared in a reliable source (which does not have to be widely available) that defeats the whole idea of it.
Also bear in mind the New York Times "kidnapped journalist" drama last year which this clause would also have been used for - is there a clear consensus that we would have communally agreed to suppress all mention of the case if Oversight norms had allowed it back then, or should do so if it happens in future?
The better wording would be a narrow and objective one:
Removal of non-public information that if published would cause clear direct harm to a person, property, or part of the natural world. The criterion is that there is some form of formal protection granted by society to the person or object at risk (for example by recognizing it as an endangered species, witness protection, court reporting blackout, or the like), that this appears to be for its own protection and welfare rather than due to censorship, and the information published would directly undermine or negate that protection.
Good wording. The only quibble I have is that it might not cover the scenario where someone posts instructions on how to break into someone's house, building, computer system etc. (unless ordinary property rights are understood to be included in "formal protection granted by society"). --JN46612:07, 17 February 2010 (UTC)
We do host numerous articles on notable material that might be considered "instructions how to..." For example the famous AACS key where we legitimately give the encryption key in the article; articles on topics like lock picking or key bumping, etc. I would trust existing process that if someone posts "here is the security code to break into XYZ", we'll deal with it. But when society has formally created a block on something for the subject's genuine protection (and not for censorship) it's rare, not covered by oversight norms, and usually only in very serious matters of survival and welfare. FT2(Talk | email)20:24, 17 February 2010 (UTC)
I support this idea, and I agree with Dweller, it would be good to remove as much subjective judgment from this as possible (consistent with keeping it a useful and applyable policy). Good idea, needs elaboration. ++Lar: t/c11:53, 17 February 2010 (UTC)
Maybe "non-public" would be more clear and stringent as "nonpublished."
I would change "if published would cause" to "if published, could lead to." The publishing is not directly damaging, but just a potential factor. Maurreen (talk) 12:49, 17 February 2010 (UTC)
This is a terrible idea. First, Jimi66 is the author of the failed Wikipedia:Sensitive wildlife locations proposal. As far as I can see, this is just an attempt to smuggle his proposal through against the consensus on that page, by wrapping it up in a stronger proposal with feel-good language. Ignoring that, though, oversight is not a tool for feel-good proposals in the first place, but only a last resort for problems that can't be solved any other way - and only for circumstances that everyone would always agree is a problem.
Asking oversighters to perform such a function is asking them to make content rulings via the oversight mechanism, and once that's done there's no way back. If everyone agrees the oversighter was wrong to make such deletions, then they necessarily are going to lose their bit over it. If there's some agreement that the oversighter was right to make such content deletions - which the above commenters are surely assuming would be true most of the time, and there can't be any harm in something uncontroversial, right? - the harm would be much worse. Once an oversighter has made such a content decision, it will not be possible for real discussion to occur on the subject, because at least one side will have to wonder exactly where the line is that would cause oversighters to nuke things from orbit on the same principle. The oversighters would have to do this, too - if they have once said that the content is oversightable, they had better defend it forever, or else they were wrong and they will necessarily lose their bit over it. This would be entirely destructive of normal editing anywhere it happened.
Jimi66, if you are concerned about a few British orchid species, there's nothing wrong with that - but Wikipedia is not an orchid protection covenant that happens to have an encyclopedia lying around - we are first and foremost an encyclopedia, and secondly, just as important, a collaborative effort to write an encyclopedia. We can't have policies like this because we wouldn't be a collaborative encyclopedia anymore. Our longstanding policy in this is WP:NOT#SOAPBOX, and it's not going to change. Please accept this. — Gavia immer (talk)13:26, 17 February 2010 (UTC)
Because we would still be holding onto damaging information which could be restored either deliberately or accidentally. Essentially we are talking about material that already meets the deletion policies (as it is not reliably sourced), but which is additionally harmful in some manner, and if harm does occur, the responsibility for that harm could morally if not legally be traced to Wikipedia. The principle is sound, though the wording and implementation of the principle needs working on. SilkTork *YES!14:58, 17 February 2010 (UTC)
Material that is obviously immediately harmful and inappropriate to leave visible to normal administrators is already covered by the oversight policy, as it should be. This proposal is to extend oversight to things that we merely should not have, or things that some editors think we should not have, such as the location of specific orchids. Oversight is not appropriate in such cases - we cannot have a lengthy debate about whether to oversight something, because that would destroy any benefit of oversighting it. Consequently, we cannot extend oversight to cases where any debate at all would be appropriate before deletion. In particular, the proposer's hope to make mention of those orchids forbidden under penalty of oversight is explicitly a hope to avoid debate about something where actual debate went against him. — Gavia immer (talk)15:17, 17 February 2010 (UTC)
Do you consider that the person who nominated this is important to the principle? It was my idea to nominate this as an Oversight clause, but as it was Jimi 66 who initially brought the concern to the community's attention I asked him to draft the proposal and set it up. Would you like diffs of all that, or shall we just concentrate on the proposal? SilkTork *YES!15:30, 17 February 2010 (UTC)
"Important to the principle", no. "Important to the context of the proposal", yes, because this tells us what the proposer hopes it will be used for. I think that's relevant information, but I certainly don't need to mention it further than I have already done. — Gavia immer (talk)15:35, 17 February 2010 (UTC)
The proposal is too broadly worded and a considerable expansion of oversight use. For instance, if I wrote at Talk:Thylacine, "Blog X has reported that a Tasmanian tiger has been spotted by the east banks of River Kumaon. Does anyone know if anyone reliable source has reported anything similar ?", my edit is eligible to be oversighted under the proposal; on the other hand if someone simply rollbacked my edit, they'll probably loose their rollback privilege. Think about that! Do we really want the bar for oversighting to be lower than the bar for rollback ?
Instead of trying to come up with new clauses for use of oversight tool for every conceivable rare situation, it is much preferable to use (1) common sense and (if necessary) IAR instead, and, (2) use tools (like reversion, deletion etc) that are more easily reviewable and reversible than jumping for the biggest gun in our arsenal. Abecedare (talk) 14:10, 17 February 2010 (UTC)
(e/c)This proposal is out of line with a lot of recent consensus, such as the decision to keep the Rorschach test images. We aren't going to remove the page on ANFO or the page on pipe bombs, or pages with verifiable information about the "secret" inner workings of Scientology which I'm sure they would claim damages the public interest. As far as FT2's "special protection" clause... well I guess he hasn't kept up with recent laws. Half the population has special protections these days. In Virginia for example, all police, prison guards, firefighters, EMTs, teachers, principals, guidance counselors, children, and the elderly have special protections under the law with enhanced penalties for crimes against them. I'm sure there's a raft of people who would argue that most nudity should be removed because it does direct harm to minors. Gigs (talk) 14:26, 17 February 2010 (UTC)
I understand the motivation for this proposal, but I don't see why this information would be so sensitive that we'd need to keep it away from the prying eyes of even our own admins (Full disclosure: I once oversighted an edit that included a valid software product key, but only because the article had more than 5,000 edits and I could not go the selective-deletion route). The existing oversight criteria are underlain by a) maximum concern for privacy (two of them) and minimizing libel exposure (the other one). I don't see how information that might be useful in the commission of a crime or other socially harmful behavior needs to be constrained by oversight, unless we were to get a court order requiring this maximum level of suppression (And I can't really see that happening unless the information were classified). Daniel Case (talk) 14:43, 17 February 2010 (UTC)
Oppose as redundant to the failed proposal, which was itself redundant to WP:V: If content is unsupported by reliable sources, it may and should be removed as unverifiable, no matter whether or not it may harm anybody or anything. Oversighting it, in addition to normal deletion, is not normally effective for suppressing harmful information that is even slightly interesting, given the nature of the internet. Sandstein 18:50, 17 February 2010 (UTC)
I agree with those who say the wording would need to be more rigorous, more objective and perhaps narrower than in my draft, and also agree FT's wording is a good step in that direction.Jimi 66 (talk) 21:52, 17 February 2010 (UTC)
Oppose. "endangered wildlife"? This seems an end-run attempt to reintroduce that clause soundly defeated in the previous attempt. Pcapping02:22, 18 February 2010 (UTC)
Oppose. This is just not a matter that requires oversight. I doubt it even warrants revision deletion by admins. Remove it from the article -- and if you couldn't get the community to support doing that, in general, there's no way you're going to get it passed as an amendment to the oversight policy.--Father Goose (talk) 09:13, 18 February 2010 (UTC)
Can we look at a separate question: Should article and talk page revisions containing information of the type we are discussing be deleted rather than oversighted (putting them out of reach of ordinary editors, but still visible to administrators), and is this current practice? --JN46613:09, 18 February 2010 (UTC)
I think that revdelete should be treated the same as oversight in terms of policy. In terms of practice it might include more things just because it's less of a limited resource, but I do not think that we should regularly be revdeleting anything that we wouldn't oversight if oversight had more resources. Gigs (talk) 14:18, 18 February 2010 (UTC)
Strongest oppose Wikipedia should not take positions on political questions. What constitutes "harm" to "nature" or the "public interest" is at the heart of politics. RayTalk21:38, 18 February 2010 (UTC)
Strong Oppose - no need for this. Current oversight policy and revision deletion should suffice. I see no need to expand oversight to to include new poorly defined categories, especially when this appears to be essentially be pseudo forum-shopping an issue. -- Bfigura(talk)22:16, 18 February 2010 (UTC)
Oppose. Simple deletion will remove any such content from search engine scope. Part of this is redundant with existing oversight policy, and the expanded part requires a level of judgment from oversighters that they should not have to make. DES(talk)02:32, 20 February 2010 (UTC)
Hilarious I've been published in an academic journal re: preserving natural resources (yes I'm in favor of it), but this proposal is redundant and goofy. Townlake (talk) 17:02, 20 February 2010 (UTC)
Comment The problem about rare species is a very serious one, especially with plants. In many cases, there will be only one very small place sometimes of only a few square meters where the only specimens of a plant species may be growing--one wanton commercial or amateur collector can extinguish the species forever. There is a very strong scientific consensus that the exact locations of such plants not be published, though it is recorded in archival sources open to serious workers. If it gets in Wikipedia , it could come only from unpublished and therefore not acceptable source, or be based on OR, and could almost always be deleted. The question is whether the revision should be hidden also, and for this specific case I think it might be acceptable. The proposal above is much too broad, however. I think the minimum that would be needed to make it acceptable would be that there is in fact a strong outside consensus among workers in the field not to publish the material. I know this is not consistent with my general principles, but even with the strongest such, in the RW, there are exceptions. I can think of a few others in various areas, but in them I would be speaking as only an amateur. The analogy for some of them is that we limit access within Wikipedia to the exact code of some of our edit filters. DGG ( talk ) 03:52, 21 February 2010 (UTC)
Support. Blanket rules just about always have holes in them, and sticking to a rule that Wikipedia must publish the information because that's our mission is absurd. I get the impression that most of the opposition to this and to the previous proposal has to do with people who see a rule saying that Wikipedia is for publishing information and insist on taking it absolutely literally. Any rule must be moderated by the consideration "is it right to do this?", not responded to by saying "Wikipedia is an encyclopedia".
And to answer one of the objections, there's a big difference between something that exists in an obscure source and something that exists in one of the top sites on the Internet. It's possible something could have a reliable source, and yet republication in Wikipedia would drastically increase the number of people who have access and result in harm that otherwise wouldn't happen.
And to answer another: the correct thing to do when faced with questions like "then why shouldn't we listen to Scientologists who say publishing church secrets causes harm" is to look at the specifics of each case individually and decide that the Scientology secrets specifically are not going to cause any harm, not to make a blanket rule which tries to ignore the need for human judgment. Ken Arromdee (talk) 03:31, 23 February 2010 (UTC)
Comment Several editors have mentioned the Rorschach test. I have no problem with the Rorschach test decision. But endangering a species that is on the brink of extinction by publishing the location of the last surviving specimens is a different matter. It's irresponsible. Rorschach images are not living things. --JN46612:13, 23 February 2010 (UTC)
Devil's Advocate: The people for whom the test is supposedly "spoiled" by seeing the blots and common responses ahead of time are very much alive. --Cybercobra(talk)15:49, 23 February 2010 (UTC)
Oppose I do see where Jimi is coming from, and I think it's good that we're having this discussion so we can arrive at a consensus on this. However, I just don't think it's our job to make moral judgments of the kind that Jimi is suggesting. In doing so, I think we would be drastically changing the character of Wikipedia. Our job is to make information that's already available more widely available. If information isn't available in a reliable source, we already have policies for dealing with that. But if something is available in a reliable source, as the information referred to above is, it shouldn't be off limits (Note that Jimi and I disagree on the reliability of the source - although the source is self-published, it's written by an expert on the locations of Britain's rare wildlife, and the information in it is true, otherwise we wouldn't be having this discussion - so it's definitely a reliable source in this literal sense!). One other pedantic point, we're not talking about extinction (as in "gone forever") in the case of either orchid - both species are common enough across mainland Europe (UK-level extirpation is the correct term). SP-KP (talk) 19:01, 24 February 2010 (UTC)
Oppose If Jimbo, who is on record that "human decency" should trump us naming Star Wars Kid despite RS'ing, can't see the point in this, why on earth is it even being discussed? Jclemens (talk) 23:27, 24 February 2010 (UTC)
Support:this may not be a popular view, but preserving the natural world is of comparable importance to preserving the dignity of any individual human. The concept of harm to an identifiable human being as the worst thing possible, more so than the well-being of our shared ecosystem, is a fallacy: we self-identify, we evaluate it as a threat to our own individual self-importance, and therefore do not make a rational moral judgment. If we would protect the lost dignity of one foolish person, we should protect the existence of a species of living beings. DGG ( talk ) 06:14, 25 February 2010 (UTC)
Fascinating. It's not often we philosophically disagree, DGG, but this is one of those cases. While I agree with the end state (unique flora and fauna kept safe), I disagree that Wikipedia can or should contribute to that effort. "Security through obscurity" is tenuous at best, and, once breached, can never be repaired. In light of the limited and fragile protection, I see no reason to alter the fundamental purpose of Wikipedia: to publish relevant information. The actual protection of limited ecological resources must be left to local authorities, not a bunch of Wikipedians suffering from hubris. :-) Jclemens (talk) 22:22, 28 February 2010 (UTC)
The Oversight Log isn't used anymore anyway, as it's now the suppression log. As for why neither are publically viewable... well, why would they be? The whole idea behind oversight and suppression is that stuff is taken care of "on the down low"; I can't really imagine why the general public needs to see what articles had diffs removed (especially when the problem lies in the titles of the pages themselves). Doubly so for grossly inappropriate usernames; the whole reason we block with "suppress username" is that it doesn't show up anywhere, and that would defeat the purpose... EVula// talk // ☯ //01:07, 22 February 2010 (UTC)
Oversight is visible in histories anyway, so it's not like oversight is secret. A visible log would be good for transparency. Gigs (talk) 03:11, 1 March 2010 (UTC)
A visible log would have to include page titles to be useful, and the page title might be the whole problem. Not to mention that having a visible centralized log would make it easy to write a bot that reads every suppression, looks for the suppressed information in all the places it might have been cached, and compiles distilled evil wherever it can find it. — Gavia immer (talk)03:47, 1 March 2010 (UTC)
In case you haven't guessed from my entry below, I really dislike the lack of transparency involved here. As a suggestion: why not make the suppression log visible to administrators? That way, it's not publicly viewable (which avoids the issues described above), but it is viewable by a reasonable number of (trusted) people. (Disclaimer: I am an admin.) Mike Peel (talk) 07:49, 8 March 2010 (UTC)
Well, bottom line, much of what is suppressed is suppressed in part because it is too sensitive even for the 1700 admins to be viewing it; a lot of it is genuine privacy violation, and no, you don't have the right to know what User 1234's real name is (or his social security number, or his children's names, or where he works) any more than the general public does. Simply put, access to sensitive personal information is not part of the administrator toolkit. Risker (talk) 08:04, 8 March 2010 (UTC)
I agree that admins shouldn't know that information, and I'm not asking for it to be made available. What I am asking for is some way for more people to find out that an oversight has taken place, e.g. "Article X was oversighted on 1 January 1900 by User:Y with rationale Z" (where "Z" might be "contains personal information"). That's the normal definition of a log... Mike Peel (talk) 08:25, 8 March 2010 (UTC)
This is a question that should be asked at the WMF or Board of Directors level; it is their rules by which we operate when dealing with sensitive information. Any WMF project can have stricter rules, but not less strict rules. Something to keep in mind is that it is a free ride for anyone to start looking for where privacy violations have occurred, and their nature, and to watch places that appear in repeat requests. I won't go further for WP:BEANS reasons. Incidentally, it is all one log, including the hidden usernames, which often include personal private information. Risker (talk) 08:32, 8 March 2010 (UTC)
Hi all. Someone (I can't find out who; I presume Big Brother) has oversighted the history of User:JPatrickBedell. That didn't do much good, as it was mirrored by Google when it became interesting to the press (and is still visible at mirrors e.g. [11]), and hence was discussed in the media (see e.g. [12]; Google has the other 4,000 stories).
A previous case like this, User:James von Brunn, was dealt with by keeping the text but placed behind a template so that it wouldn't be easily Google-able. Although there are 66 deleted edits, those are all post-press edits by others. Why wasn't this dealt with in the same way, and why isn't this the standard approach in this sort of situation? It's much less Big Brother-like, and much more transparent. Thanks. Mike Peel (talk) 07:39, 8 March 2010 (UTC)
Can you link me to the other discussion, please? I'm somewhat disappointed and concerned about the lack of further discussion here about this... Mike Peel (talk) 16:22, 14 March 2010 (UTC)
As a point of order and not necessarily an endorsement of what was done - the page isn't presently oversighted but the edits are hidden from non-admins by Revision deletion. Admins need to click thru the prompts to add unhide=1 to the diff to view it. –xenotalk13:31, 15 March 2010 (UTC)
CheckUser and Oversight Elections
The Arbitration Committee has determined that there is a need for additional oversighters and checkusers to improve workload distribution and ensure complete, timely response to requests. Beginning today, experienced editors are invited to apply for either or both of the Oversight or CheckUser permissions. Current holders of either permission are also invited to apply for the other. The last day to request an application is April 10, 2010. For more information, please see the election page.
Beginning 15 May 2010, the English Wikipedia Oversight mailing list will be migrating to the OTRS mail management system hosted by the Wikimedia Foundation. The primary purpose of this move is to better track requests as they come in, and to ensure timely and consistent responses. This move comes after the German and French oversight lists moved to OTRS in the past year; both have found that it has assisted them in better responding to requests. Over the next week or so, oversighters who have not used OTRS before will be learning the fine points of that system, but the Oversight team will endeavour to maintain adequate responses to the system. The team has also prepared an introductory manual to assist with the transition, which discusses use of both the OTRS system and the Oversight tools.
The major effect on non-Oversighters will be the change in email address to which requests should be sent. When that change is made, we will widely publish the new email address for everyone's information, and we will encourage regular correspondents, particularly recent change patrollers and new page patrollers, to update their contact lists. The current Oversight-L mailing list will remain accessible for approximately two weeks after the changeover; after that, it will become a closed list where oversighters will discuss complex cases or review best practices.
The CheckUser and Oversight election has now opened. Any editor who has made at least 150 mainspace edits prior to the first announcement of the election may vote. The voting will close at one minute past 23:59 UTC on 27 May 2010.
I brought this up on Template talk:Importance, but nobody watching had an efficient idea - there are 118 deleted edits in Template:Importance's history that would be much better placed, undeleted, in the history of Template:Cleanup-importance, since the template was re-purposed. The change itself should not be controversial, but the poor man's process to get them there sounds pretty disruptive to me. Since oversighters have additional powers regarding revisions, maybe they can accomplish the same goal with no fuss? --Joy [shallot] (talk) 18:59, 15 May 2010 (UTC)
You can't do this other than by deleting and restoring — redaction will keep the edits in the history of that template, just invisibly. Stifle (talk) 13:12, 19 May 2010 (UTC)
OK. So my question becomes - will anything explode if I just do it the normal way? It's a template transcluded into over two million pages. --Joy [shallot] (talk) 13:48, 19 May 2010 (UTC)
I am pleased to advise you that, effective immediately, requests for oversight/suppression will be accepted using the OTRS system. Please bear with us as the Oversight team becomes accustomed to this new method of receiving and replying to requests. We will strive to maintain timely service.
If you have found yourself reporting concerns to the oversight mailing list, please take a moment to add the new email address to your list of contacts: oversight-en-wpwikipedia.org
We look forward to continuing to work with the community in protecting the privacy of editors and others.
Good question, and yes, we receive a lot of requests through that route. We're in the process of updating all of the email addresses throughout the project, and this will be one of them. During the transition period, we will be responding to requests from both the old and the new addresses. Risker (talk) 05:18, 24 May 2010 (UTC) Update: Now updated. Risker (talk) 06:07, 24 May 2010 (UTC)
all Wikimedia admins ... at least so I understood. The document says "On the English language Wikipedia, access to the Oversight and RevisionDelete tools is controlled by the Arbitration Committee." I won't challenge, just curious if things go differently here on EnWP and there are some technical differences of settings on other wikis. --Aphaia (talk) 04:00, 30 May 2010 (UTC)
Call for applications for Checkuser or Oversight permissions
The Arbitration Committee invites applications for Checkuser or Oversight permissions effective with the posting of this motion. The application period will close at 2359 hours UTC on 1 August 2010. For this round of appointments, only administrators will be considered. Candidates who ran in the May 2010 elections elections are encouraged to apply for consideration in this round of appointments. Administrators who applied for permissions in the round leading to the May 2010 election may email the Committee at arbcom.privilegeswikipedia.org by the close of the application period, expressing continued interest and updating their prior responses or providing additional information. New applicants must email the Committee at arbcom.privilegeswikipedia.org by 30 July 2010 to obtain a questionnaire to complete; this questionnaire must be returned by the close of the application period on 1 August 2010. The Arbitration Committee will review the applications and, on 13 August 2010, the names of all candidates being actively considered for appointment will be posted on-wiki in advance of any selection. The community may comment on these candidates until 2359 on 22 August 2010.
The Arbitration Committee accepts, with regret, the resignation of Tznkai (talk·contribs) as a member of the Audit Subcommittee, effective immediately. Tznkai has been a community representative on the AUSC since its creation in 2009, first as an interim appointee and subsequently as a community-elected representative. As well, he has been a long-time Arbitration Clerk, and has been active in arbitration enforcement. We thank Tznkai for his services, and wish him well in his future endeavours, with the hope that he may return to be an active Wikipedian at some point in the future.
In addition, arbitrator KnightLago will be filling the slot now vacant as Kirill Lokshin has come to the end of his term on the AUSC, and SirFozzie has agreed to extend his term to December 31, 2010.
The Arbitration Committee, in consultation with the community and with past and present members of the Audit Subcommittee, will be reviewing the activities and processes of the AUSC through its first year, to identify what improvements can be made. This review will be completed by October 10, prior to the next scheduled round of elections for community representatives to the subcommittee.
Update on Checkuser and Oversight appointment process
Following the call for applicants (19 July) and the initial call for comments on the candidates (16 August), this notice is a second call for comments from the community on the suitability of the candidates for the September 2010 appointments for checkuser and oversight permissions. The Arbitration Committee is continuing to review and collate the comments received so far. If you have not done so already, please send in your comments before 23:59 on 25 August 2010 (UTC).
Those actively being considered for Checkuser and Oversight permissions are listed here (same link as above). As the primary area of concern is confidence in the candidate's ability to operate within the Wikimedia privacy policy, comments of this nature are best directed to the Committee's mailing list (arbcom-llists.wikimedia.org).
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.
It appears that suppression criteria #4 and #5 of the policy are no longer necessary, due to the current ability of administrators to apply revision deletion in cases of attack names in logs and vandalism. Nonetheless, there might be some reason to retain the criteria of which I'm unaware. Thoughts? Peter Karlsen (talk) 22:23, 10 September 2010 (UTC)
I don't see any pressing reason to make this change. There may still be cases where it is in Wikipedia's best interest that the material be suppressed. Beeblebrox (talk) 00:47, 11 September 2010 (UTC)
I'd have to agree with Beeblebrox. There are some things that need to be oversighted. There is no harm in leaving those criteria in. —fetch·comms01:54, 11 September 2010 (UTC)
Removed? Probably not. I would, however, feel free to add a note that the admin-level tools should be used in routine cases, and oversight per se restricted to such egregious cases where the general admin corps should not be seeing the edits. Jclemens (talk) 05:05, 11 September 2010 (UTC)
Okay. I've modified the policy to state that oversight/suppression should always be used to remove non-public personal information, but that for other cases, administrative revision deletion should be used when appropriate, and the oversight tool only employed "when justified by the circumstances". This is quite likely the practice of editors entrusted with oversight already; the purpose of the policy change is primarily to encourage administrators to handle many issues using revision deletion themselves, and only e-mail the oversight list in cases of non-public information, or when otherwise necessary. Peter Karlsen (talk) 08:00, 11 September 2010 (UTC)
So basically, what we're doing right now already? If so, I doubt it matters too much. Sometimes, a little ambiguity can lead to useful instances of IAR. OSers should already know when it is appropriate to do this. —fetch·comms01:53, 12 September 2010 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
When someone emails an oversight request, their email address and identity ought to be confidential. Please at least edit the information box at the top of WP:Requests for oversight to guarantee confidentiality. I would guess that some people shy away from making oversight requests because of the confidentiality concern, and they don't want to go to the trouble of setting up a dummy email account.Anythingyouwant (talk) 22:36, 12 December 2010 (UTC)
A couple of points: Firstly, what do you suppose the oversighters should do about genuinely abusive communications with the mailing list, if they were to guarantee confidentiality? And secondly, do you have any evidence that confidentiality isn't being extended? I've brought things to the oversighters' attention several times, and I've never seen a word of discussion about my oversight requests anywhere on the wiki. There's no need to fix what isn't broken, especially if the fix would itself cause forseeable problems. — Gavia immer (talk)23:10, 12 December 2010 (UTC)
First, I did not mean that an emailer's username needs to be confidential, but rather that the rest of their identity (email address and real name which is often derivable from email address) should be confidential. Second, I am not questioning whether Oversighters generally respect confidentiality, but rather am suggesting how oversight requests can be more assuringly explained so that people are less reluctant to make such requests. I personally was reluctant to make my first oversight request during the past 24 hours because the explanation I saw didn't say anything about whether my email and thus my real-life identity would be freely forwarded or shared. It's a very easy thing to fix without causing any possible problem, and I think it would increase use of oversight.
At WP:Requests for Oversight, after the sentence "See below if unfamiliar, for more details on when to request suppression, how to contact an oversighter, and information needed" simply write, "Your email address and identity beyond your user name will remain confidential." This would fix the problem.
I hesitate to point this out because it seems so obvious. Oversighters, who are the ones removing confidential or otherwise private information from this site, are about the last people on earth to violate someone else's privacy by handing out their email addresses. Adding this to the page is redundant at best. Risker (talk) 05:39, 13 December 2010 (UTC)
<Undent> To the extent I can discern your tone, Risker, it suggests you very much dislike my suggestion, which was only meant to increase users' willingness to submit Oversight requests. So I won't belabor the point, and please feel free to have the last word if you want.
The privacy policy linked at the bottom of this page only covers information "stored by the Foundation on its servers", and explains that "Users may also interact with one another outside of Foundation sites, via email...and should assess the risks involved, and their personal need for privacy, before using these methods of communication." (The policy also excludes: "use of the wiki 'email user'" feature.)
Maybe that is why assurances of privacy and confidentiality are very common elsewhere at Wikipedia. Just to mention a very few examples:
Wikipedia:Mediation Committee/Policy: "Parties concerned about a public statement may make their objections known privately to the Chair of the Committee, via email."
Wikipedia:Accessibility advocates: "A note about confidentiality: Information divulged to an advocate via e-mail will be treated as confidential unless expressed permission is given to share it."
So, I'm not quite sure why you think that the privacy policy linked at the bottom of this page is somehow unique. Anyway, cheers to you Risker.Anythingyouwant (talk) 07:16, 13 December 2010 (UTC)
Done [2] ~Kylu (u|t) 00:53, 12 March 2008 (UTC)
wikipedia.org
We look forward to continuing to work with the community in protecting the privacy of editors and others. 
Resolved