Wikipedia talk:Flagged protection and BLPs

Frequently asked questions

Q:Why the fuss about biographies of living persons / Why are they more special than any other article?

A:The essays in Category:User essays on BLP give an idea of other users' take on this issue. The problem essentially boils down to the potential to do harm to real people. Wikipedia articles tend, for one reason or another, to rank quite highly in internet searches, and a libellous/defamatory and inaccurate statement about someone can cause harm to their reputation for each person reading it - there are even reports of people missing out on jobs because of inaccurate information. Wikipedia serves millions of readers a day who are not editors, so it is important that what they read is accurate when that information can affect other people's lives adversely. In terms of reversion being the answer, there are two counter-arguments:

If even a few people read the article before reversion, it can have a damaging effect.
The BLPs of less notable individuals are not heavily watchlisted and so the vandalism lingers for longer

Q:Why do we need a separate usergroup / Isn't this just more bureaucracy / elitism ?

A:Opening the project up to allow everyone, anonymous and otherwise to contribute to currently protected articles is a good thing in many people's eyes. The separate issue is protecting BLPs, for which FlaggedRevs has often been touted. In previous discussions, it has proven difficult to distinguish between the differing roles and perceived responsibilities of someone reviewing a flag-protected non-BLP article and someone reviewing a BLP article. This largely arises from the autoconfirmed threshold being relatively low, which is because it is intended as a balancing act between deterring vandals and allowing new good-faith editors to easily establish themselves. Rather than making Flagged protection more stringent by changing this threshold, creating a new BLP usergroup separates the problem out so that it can be considered separately.

Q:Won't this make it more difficult to review BLPs because promotion isn't automatic?

A:Promotion could be made automatic if some suitable editing metrics could be determined. The separate group would still be required to distinguish the abilities of the usergroup. Editors in good standing with no recent history of vandalism should have no problem rapidly acquiring the permissions from the well-frequented WP:PERM.

Q:Why not just give it to Rollbackers?

A:Rollback is a tool for reverting obvious vandalism. The conditions for obtaining it may differ from those of a BLP reviewer, and, perhaps more significantly, the conditions for removal. As an analogy, it makes sense not to have to remove someone's toolbox if they are only incapable of using the hammer properly.

Q:Why should Administrators automatically be BLPReviewers / have more power?

A:This proposal doesn't change the status quo, but there is an understandable perception of cause for concern. As such, the proposal makes it clear that the sysop group will not be granted the power as part of the sysop package of tools, but be granted them separately. It shall be assumed that Admins can be trusted, since they are appointed by the community into a position of trust. This will, however, allow this particular tool to be removed from an admin perceived to be abusing it, without the horrendous litigation of an Arbcom decision to have them desysopped.

Q:What about backlogs? Won't the backlogs be enormous?

A:We don't know. The inevitable comparisons to the German Wikipedia are irrelevant since they flag all of their articles with a much smaller userbase. This proposal will flag a small minority of our articles, much smaller in magnitude to the German implementation, and we have a significantly larger userbase.

Q:Flagged revisions in any form are bad, because it drives contributors away

A:Ok, not really a question, but a common concern. There is no evidence that this happens. All wikis are currently seeing a trend towards a drop in contributors, not just those who have implemented FlaggedRevs, so it is an error to directly link the two. The flagged protection aspect of this proposal should actually mean that all can contribute to the development of articles.

Q:Isn't this just creep towards implementation across the whole wiki?

A:In the first instance, this is a trial, not an implementation. It took years to get us to this discussion, it may take months before this is even approved/rejected. If this proposal were fully implemented, then in itself, it would not lead to wider use because it specifically applies only to BLPs and articles that should be protected under our existing policies. This should also reduce fears of creep beyond scope, since this proposal is no more liberal about Flagged protection than our existing state of affairs. To implement this wiki-wide would, in my opinion, be impossible to achieve because so many oppose it, including the author of this proposal.

Q:Why trial both Flagged protection and BLP simultaneously?

A:Principally, if we're to treat the two as having different requirements, we need two systems. And if we have two systems, we need to make sure that we can cope with them both at the same time. Plus, it is a more valuable test of the backlog concern and a more efficient process to run multiple tests at the same time.

Q:What if one half of the trial works and the other half doesn't?

A:If the community, in discussion following a trial, decides that only one of Flagged protection or BLP protection is working out, I would hope we'd see that in the consensus formed. If it happens, we could simply undo the part of the system that doesn't work and retain the part that does. That's why it will be so important during these discussions that people do more than just write support or oppose, since it is the reasons that matter.

Discussion

Trial conditions 1.1 Pre trial

Regarding this: "The top 2000 BLPs by quantity of vandalism over the preceding 3 months shall be selected, and from these, half shall be selected at random for flagging. The remaining unflagged articles will act as a control group for any comparative metrics required during subsequent discussions.". Why are you only picking highly vandalised pages? The proposal is to apply to all BLPs, so a representative sample of all BLPs need to be in the trial, both highly vandalised and low visibility pages. MickMacNee (talk) 16:28, 28 January 2009 (UTC)[reply]

Yeah, I think I was thinking about how to choose the sample, and thought that as FR is meant to help with the vandalism it would be a reasonable metric. I see your point though; how about picking 1000 articles at random from the BLPs, and 1000 extra at random as the control? My only concern with that is guaranteeing sufficient activity to see if FR had any effect.... Fritzpoll (talk) 16:32, 28 January 2009 (UTC)[reply]

Well, plenty of people see FR as being usefull in preventing harmfull additions to low traffic articles, where the damage occurs because nobody is watching the article. It seems sensible to include them. MickMacNee (talk) 16:35, 28 January 2009 (UTC)[reply]

What if we excluded "top importance" articles from the trial? These get sufficient watchlist coverage anyway. Fritzpoll (talk) 16:42, 28 January 2009 (UTC)[reply]

If its going to be a true all BLP articles trial, for simplicity, it should simply select from all types of BLP artices. Defining what a top importance article is is just going to open up a whole other can of worms for a later date. MickMacNee (talk) 17:03, 28 January 2009 (UTC)[reply]

Ok. I'll alter it in about 30seconds - if you can check the new wording, that'd be appreciated Fritzpoll (talk) 17:06, 28 January 2009 (UTC)[reply]

The text "1000 articles from Category:Living people will be selected at random to be flagged, with a further 1000 selected at random to act as the control group to provide the basis for any comparative metrics required in subsequent discussions." in this version seems fine, but has been changed since. (PS, I haven't realy though about whether 2000 as a figure is enough, my original issue was the selection of highly vandalised articles).

BLP Reviewer status to be granted to Users in "good standing"

I hate this expression. Is there any way it can be better defined? What is to happen if you are denied by one admin, will it be considered 'admin shopping' to seek for a second/third/fourth opinion? I have never wanted Rollback so have no clue what the accepted norms are. MickMacNee (talk) 16:35, 28 January 2009 (UTC)[reply]

Well, rollback is basically given to people who have a demonstrable history of reverting vandalism appropriately, but how that is applied varies slightly from admin to admin. There are often discussions on close calls at WP:PERM when admins aren't sure, so the second/third opinions are there. I think "good standing" is used because it is deliberately vague, but conveys the sense of "don't give this to a recidivist vandal" Fritzpoll (talk) 16:41, 28 January 2009 (UTC)[reply]

Flaggable condition

This should be done for all 2000 articles, and then the thousand for FR selected randomly. This may be the major benefit of flagging, and we should control for it. I will emend the proposal, but do revert and explain if you disagree. Septentrionalis PMAnderson 17:19, 28 January 2009 (UTC)[reply]

Yes, that sounds more statistically useful Fritzpoll (talk) 18:16, 28 January 2009 (UTC)[reply]

More Famous People

It's been suggested to me that some high profile people should be deliberately included in this trial - the 'Hilary Clintons' of this world, whose entries are frequently interfered with by vandals. Personally I feel that high profile politicians lives are less likely to be damaged by malicious edits than less notable people, but including this category in the trial may help it garner more support from people who are scared of offending powerful politicians and litigious businessfolk. Maybe we could include 500 'famous' BLPs whose entries are currently protected or semi-protected. BTW this work to produce a 7 day consensus is well worthwhile and very laudable. Riversider (talk) 20:40, 28 January 2009 (UTC)[reply]

They would be covered by the flagged protection portion of the trial - I've thought this through :) Fritzpoll (talk) 23:04, 28 January 2009 (UTC)[reply]

Minor suggestion

Would you consider Wikipedia_talk:Flagged_protection#Key_tweak per the reasons VoA states? I think this would reduce the amount of incorrect flagging with few restrictions. GDonato (talk) 23:02, 28 January 2009 (UTC)[reply]

WIll have a look in the morning - feel free to add yourself Fritzpoll (talk) 23:04, 28 January 2009 (UTC)[reply]

I think it would basically consist of removing the fourth point for autoconfirmed in the table. Having autconfirmed able to edit but not review semi-protected articles seems more sensible to me (prevents Grawp having too much fun) GDonato (talk) 23:09, 28 January 2009 (UTC)[reply]

I think the limits will always be arbitrary - even the autoconfirmed limit is a bit arbitrary. I'll have a think and a look. Fritzpoll (talk) 07:59, 29 January 2009 (UTC)[reply]

I will await further comments on this - I am very open to revisions that move us towards consensus. Fritzpoll (talk) 14:39, 29 January 2009 (UTC)[reply]

Tools

Can you make a tools list Wikipedia talk:Flagged protection and BLPs/Tools for BLPs, sort of tools list similar as the tools list for semi protection Wikipedia:Flagged_protection#Tools ? Thanks Mion (talk) 14:27, 6 February 2009 (UTC)[reply]

Only the BLP part.Mion (talk) 14:27, 6 February 2009 (UTC)[reply]

Trial Process Thoughts

A proposed trial with a control group! Hallelujah. Too many of the "trial" designs that have been tossed around on this site and on this issue avoided this step, which is necessary for meaningful evaluation of results.
Unfortunately, it will probably still be easy to tell whether the article actually is flagged or not, limiting the use of the control as it applies to vandals and editors that know about the trial. I've never been able to think of a way around this.
What appears to really be missing is a clear separation of the analysis phase from the consensus discussion. Those should be two separate discussions. And the analysis discussion should come first. GRBerry 01:32, 19 February 2009 (UTC)[reply]

I am very sad

That Wikipedia has not implemented this yet, especially for WP:BLP articles. Not a single fucking day goes by that I don't see some horrendous form of libel on the very small list of articles that are on my watchlist. Not a SINGLE day. JBsupreme (talk) 08:02, 10 March 2009 (UTC)[reply]

Well, I obviously agree with you there. Problem is, I'm not getting many people here talking about this proposal, and Jimbo appears to be wandering off doing his own thing anyway. Fritzpoll (talk) 10:14, 10 March 2009 (UTC)[reply]

Still an active proposal?

Just trying to clean up Category:Wikipedia proposals so wondering if this is still an active proposal or if it can be tagged otherwise? Hiding T 09:53, 25 August 2009 (UTC)[reply]