Misplaced Pages

Revision as of 18:54, 20 December 2008

This is a simple survey to see if there is any support for, and what people think of, the ideas of enhancing the protection of BLP articles on Misplaced Pages, the articles about "living people". THIS IS NOT A POLL TO CHANGE ANYTHING TODAY. Rather, it is just a survey to get a general idea of what people think, and feel, of the ideas to address the outstanding problems with biographical informations on living people. The two ideas that seem to be floated most often are to use Misplaced Pages:Flagged revisions on these articles, or to use Semi-protection by default on all of these articles. This spawned from this conversation on User:Jimbo Wales's talk page.

Two example studies that were compiled of vandalism in general, and vandalism to high profile BLPs, are:

• General study of anon vandalism to WP articles
• Study of vandalism to articles about United States Senators

It may be worth reading some user essays on the subject: Category:User essays on BLP

Again: This survey is not to change anything today but to just see what if any options are popular, and to see if there is support to later work further on one, the other, both, or none.

What does Jimmy Wales think?

This is what Jimmy had to say about this:

"I would be thrilled with the implementation of flagged revs on all BLPs as a test. I don't think I should force it through, but I strongly support that we experiment with it. What I will do is this: I will gladly serve as a formal point of contact to ask the Foundation directly to implement whatever we decide on. What I recommend is a timed test, i.e. turn on flagged revs for all BLPs for 3 months, and have a poll in the last two weeks of that period to determine whether we want to keep it. I feel confident that we will.

I would also support us simply copying what the Germans have done with it. I know there are concerns about volume, but the Germans are able to deal with it just fine as I understand it. Yes, we have more edits, but we have more community members. So I reckon we can deal with it quite well. However, I'm *thrilled* about it for BLPs and merely *supportive* for all articles.--Jimbo Wales (talk) 23:24, 19 December 2008 (UTC)"

Implement Semi-protection for all BLPs

1. Have always supported this. I'd also support a proposal made by Greg Kohs (comment on the proposal not the proposer…) to experimentally semiprotect a subset of BLPs (for example, all those beginning with "A") and after a couple of months compare the vandalism & editwarring stats with the unprotected control group. – iridescent 19:28, 19 December 2008 (UTC)
2. I find no compelling reason to not use a tool we already have, semi-protection, to protect the reputations of living people on whom our project contains articles. If an IP wants to edit a BLP, create an account. One of the main reasons I first created an account back in early 2007 was that I was sick of being shut out of editing semi-protected articles. It's not a big deal to place semi-protection, and the reward (less vandalism and general junk on BLPs) far outweighs the risk (less IP editing generally). SDJ 19:29, 19 December 2008 (UTC)
3. 2nd, but I think Flagged is better. rootology (C)(T) 19:31, 19 December 2008 (UTC)
4. I agree. Tired of people coming in and trashing BLPs and I have to clean it up, especially when I don't look carefully enough for a few weeks and bad and good edits get mixed up. Frankly, I think you should give IPs maybe 50 to 100 "free" edits and after that they have to register. It's not like they are protecting anything being anonymous or giving anything away being registered. (Or maybe it is, but I think most of them don't know what it is any more than I do.) CarolMooreDC (talk) 19:53, 19 December 2008 (UTC)
5. The only risk is that article subjects who see defamation will not be able to fix it without registering. As long as we do this via a template or some such that also includes, in the sitenotice text, links to the OTRS email address and BLP noticeboard, there is no real reason not to at least test the idea. I know a lot of people make much of anons providing a lot of our content, but I think that is less true now than it was and should be balanced against the fact that they also provide rather a lot of vandalism. In the case of BLPs, vandalism is a more pressing problem than it is in, say, articles on Pokémon. Guy (Help!) 20:02, 19 December 2008 (UTC)
6. Agree per CarolMooreDC. Willking1979 (talk) 20:14, 19 December 2008 (UTC)
7. BLP's should be semi protected, I have always supported thus, but again, it won't happen, Misplaced Pages and "change" do not mix well. — Realist 20:15, 19 December 2008 (UTC)
8. Absolutely, but as a second choice. I would favor doing this if it can be implimented more quickly. It shouldn't be required once flagged revisions are implemented, however. Cool Hand Luke 20:31, 19 December 2008 (UTC)
9. I agree. I have frequently called for indefinite semi-protection of BLPs suffering from inordinate vandalism -- Ilan Pappé, Norman Finkelstein, Vladimir Lenin and several more. This would prevent not only anon ISP vandalism, but also the constant vandalism by one-off throw-away acocunts that has characterised these and other BLP articles. This peoposal would contribute significantly to reducing this problem. RolandR (talk) 20:55, 19 December 2008 (UTC)
10. First choice. shoy (reactions) 21:21, 19 December 2008 (UTC)
11. Second choice, or step on the way where we need to be (Semi and flagged for all BLP, flagged for everything in article space) ++Lar: t/c 21:39, 19 December 2008 (UTC)
12. Second choice. Actually choice 2.5, as the optimum would be stricter than just flagging. Collect (talk) 22:00, 19 December 2008 (UTC)
13. Second choice. FloNight♥♥♥ 22:09, 19 December 2008 (UTC)
14. . Thanks, SqueakBox 23:17, 19 December 2008 (UTC)Support this option. Thanks, SqueakBox 23:17, 19 December 2008 (UTC)
15. . Keep it simple. I prefer this marginally over introducing yet another layer (flagged revisions), but both have merit. Cheers, Casliber (talk · contribs) 23:44, 19 December 2008 (UTC)
16. Obviously I support something I have been saying for months should occur. MBisanz 03:47, 20 December 2008 (UTC)
17. I haven't seen any real convincing evidence to support a measure this drastic. Is there evidence of anonymous users or brand new users causing significantly more harm to BLPs than autoconfirmed users? And if so, does anyone have a link? --MZMcBride (talk) 16:46, 20 December 2008 (UTC)

    Gurch had some stats somewhere based on what was being reverted by Huggle, I believe – iridescent 18:52, 20 December 2008 (UTC)

18. This is the least radical of possible steps, and it should be tried first. It will certainly help the situation somewhat, and then we can judge what to do further. If we do try flagged revisions, we should try it for a very small subset first, as proposed by Cenarium and others in the section on General Comments. (I'd suggest US Presidents)DGG (talk) 17:10, 20 December 2008 (UTC)

Oppose Semi-Protection for all BLPs

1. Object. It's a cost-benefit thing. A fair number of people who want to exploit BLPs are sloppy and edit from static IP addresses. This would cut down on abuse somewhat, but it would also reduce accountability. Think how useful the Wikiscanner was at sniffing out trouble. Why would we want to make BLPs marginally less difficult to exploit, while shutting down Wikiscanner capabilities completely? Durova 19:45, 19 December 2008 (UTC)
2. Oppose strongly, this would be a clear break with the ability of anybody to edit. Flagged revisions are far superior for the reasons I have commented in the flagged revision section above below. Davewild (talk) 20:58, 19 December 2008 (UTC)
3. Oppose in the strongest possible terms. This would only
   1. Create a false "don't need to watch this page anymore" sense of security.
   2. Make determined BLP attackers more difficult to identify at a glance, because they have to build up a certain number of meaningless edits before doing deliberate damage.
   3. Funnel drive-by BLP violations to tangentially related non-BLP articles which are unprotected and probably unwatched.
   4. Leave good-faith newcomers unable to edit at all when their favorite baseball players (or whoever they're interested in) are s-protected and they don't know how to game the "autoconfirmed" criteria. Then if they figure out how to ask for help, we do what? Ask them to "just go edit asteroids or pokémon for a week"? Forget that! — CharlotteWebb 21:18, 19 December 2008 (UTC)
4. Oppose More BLP paranoia. We have a functioning and strict BLP policy. We just need to enforce it and edit accordingly. We don't need to invent these contrivances to "solve" this non-problem. And I oppose flagged revisions generally. Protonk (talk) 21:20, 19 December 2008 (UTC)
5. Not only "No" but... While we need to be responsible in our treatment of BLPs, a default position of semi-protect is antithetical to the basic premise of Misplaced Pages. It is obvious that a great deal of vandalism comes from IPs, but so do a tremendous number of constructive edits. In addition, I suspect that the overwhelming majority of us here first contributed as IPs before becoming sufficiently hooked to register an account. It is foolish to create an impediment to the contributions of those who would join this community. And consider, in many cases the way we find out that there's a problem on a BLP is when an IP (probably the subject or someone close to them) repeatedly tries to fix a problem and gets reverted. The need to protect the subjects of thisthese articles would be addressed in a far superior way through the use of flagged revisions on problematic articles.Xymmax _{So let it be done} 21:42, 19 December 2008 (UTC)
6. Oppose Could serve to lock in vandalism, such that the subject who found something untrue in the article can't just fix it, while the benefits from doing this are largely obsoleted by flagged revisions. -Steve Sanbeg (talk) 21:44, 19 December 2008 (UTC)
7. CharlotteWebb put it nicely. Let's not forget the point of Misplaced Pages. John Reaves 21:49, 19 December 2008 (UTC)

   Building a stable and reliable encyclopedia? Cool Hand Luke 22:02, 19 December 2008 (UTC)

   So why not semi-protect the whole mess? Protonk (talk) 22:35, 19 December 2008 (UTC)

   Why not? Still the site anyone can edit, but no longer the site anyone can drive-by vandalize. BLPs just happen to be the most potentially damaging class of articles. Cool Hand Luke 00:50, 20 December 2008 (UTC)

   Because it betrays one of the fundamental purposes of wikipedia: that editing be as open as possible. No one promised that these purposes were going to be concordant--the meat of good policy debates comes from situations where guiding principles conflict or are mute. WP:NFC is a compromise between our goal to be a 💕 and to be a comprehensive encyclopedia. It would be unacceptable to betray either pillar through inaction or immoderation. Likewise we have here two competing principles, we need to satisfy both partially. Semi-protecting BLP's preemptively and indefinitely doesn't do that. It's lazy, it won't stop determined vandals, it will not stop real threats to BLPs (no one really believes obvious vandalism on a page but they can believe misinformation inserted by dogged SPAs), and it turns away too many productive edits and prospective editors. Protonk (talk) 01:28, 20 December 2008 (UTC)

   The fundamental purpose of Misplaced Pages is building and maintaining an encyclopedia. If any supposed principle gets in the way of that, we should ignore it. Cool Hand Luke 01:34, 20 December 2008 (UTC)

   No, its not just an encyclopedia, its a 💕 that anyone can edit. We can't toss away "free" or "anyone can edit" because it gets in the way of "encyclopedia" in some cases. Any idiot can see that you would get a stable encyclopedia by restricting who can edit it. Paper and other electronic encyclopedias have been doing it for years by restricting editing to their staff members. "Anyone can edit" is why Misplaced Pages is successful, and why we have 10 million articles in 250 languages. If you disagree with that principle, I hear Citizendium is recruiting. Mr.Z-man 04:01, 20 December 2008 (UTC)

   Clearly I don't. You don't need to lecture me; I have a passing familiarity with the success of Misplaced Pages. Cool Hand Luke 05:30, 20 December 2008 (UTC)

   Sorry, I was confused by your comments suggesting we abandon the "anyone can edit" principle in favor of more stable content. Mr.Z-man 06:06, 20 December 2008 (UTC)

   I'm sure you do. I voted for you. I don't think he was intending to lecture you, just make clear the point that the foundation issue of a 💕 is in contest with the implementation of semi-protection for a class of articles. Further, since you alluded to IAR, he was probably hoping to show that this was an ends based argument, not some position taken merely out of advocacy or inertia. The real point is we can all go back and forth appropriating "what wikipedia is" for our end of the argument because the definition is too fluid and vague to serve as a decision rule here. I accept that wikipedia faces a tradeoff in anonymous editing. We may disagree on the magnitude of that tradeoff, but we agree that there is one. Protonk (talk) 06:12, 20 December 2008 (UTC)

   We absolutely agree. There's always a trade off. My first edit was to start a new article that was redlinked. We haven't allowed IPs to do that in a long time, and we've certainly lost some contributors who would have begun like I did. In this case, I just think we have particular editorial obligations to BLP subjects which trump volunteer recruiting.

   Incidentally, if anyone wants to move some of this thread to the currently-redlinked talk page, I wouldn't mind at all. Cool Hand Luke 06:39, 20 December 2008 (UTC)

8. Pointless when you can have flaggedrevs. Semiprotection can and should be used liberally but only when there has been a problem: pre-emptive protection of all BLPs is just overkill. Moreschi (talk) 22:18, 19 December 2008 (UTC)
9. No - flagged revisions are OK, even necessary, for BLPs, but stopping all anonymous or non-autoconfirmed users from editing them at all is far too extreme. Dendodge Talk 23:27, 19 December 2008 (UTC)
10. Oppose per the very good reasons outlined by CharlotteWebb and Xymmax. This is throwing the baby out with the bathwater. Mr.Z-man 00:16, 20 December 2008 (UTC)
11. Oppose; flagged revisions offer a much less intrusive way of protecting articles. — Coren  00:39, 20 December 2008 (UTC)
12. oppose per Charlotte, Xymmax and Coren. JoshuaZ (talk) 01:00, 20 December 2008 (UTC)
13. Strong oppose - This chucks the baby out with the bathwater. IPs do work too, you know. - NuclearWarfare _{My work} 01:19, 20 December 2008 (UTC)
14. Oppose Many useful contributors to BLPs are anons - protection should be on a case-by-case basis. --Philosopher  01:41, 20 December 2008 (UTC)
15. Support a compromise solution. Any BLP should be automatically semi-protected for a period of time on a request from any established user if an IP is making reverts in the article.Biophys (talk) 05:21, 20 December 2008 (UTC)
16. Cost-benefit. Most BLPs do not ever attract disruption. Those that do can already be made subject to some sort of protection.  Sandstein  07:50, 20 December 2008 (UTC)
17. Oppose; this essentially and eventually becomes disallowing anons to edit and thus a Foundation issue, not a community one. It's too slippery of a slope towards Citizendium. -Jéské Couriano 07:51, 20 December 2008 (UTC)
18. Oppose this is overkill. It would reduce the amount of vandalism on BLPs, but we have to weight this against the huge adverse effects. Hut 8.5 10:12, 20 December 2008 (UTC)
19. Oppose, it would not cost wikipedia just the anon edits, but also a number of contributors, who started out as anons and got hooked on Misplaced Pages. bogdan (talk) 11:30, 20 December 2008 (UTC)
20. Strong oppose another kick in the face for anonymous editing. Antithetical. Skomorokh 13:11, 20 December 2008 (UTC)
21. Strong oppose. Studies have shown that the anon edits make up a significant amount of vandal reversions. Making it harder for anons to fix errors in BLPs is not beneficial to the project. It also goes against the whole idea of being an encyclopedia anyone can edit when the majority of BLPs pose no problems. - Mgm| 14:24, 20 December 2008 (UTC)
22. Oppose Unnecessary. --Tango (talk) 14:28, 20 December 2008 (UTC)
23. I think a blanket semi-protection of BLPs will probably be more harmful than the vandalism it will prevent. Flagged revisions are a better solution. -- Jitse Niesen (talk) 16:12, 20 December 2008 (UTC)
24. Oppose—any systematic protection of pages is effectively a way of categorically denying people the right to edit that page. That seems to me to be in conflict with one of our core principles, that anyone should be able to edit. While abuse concerns might even justify the use of sighted-revision-by-default FlaggedRevs (which I also intensely dislike), systematic protection of any level is simply unacceptable. {{Nihiltres|talk|log}} 16:47, 20 December 2008 (UTC)
25. Oppose the autoconfirm barrier is way too high. -- lucasbfr 17:19, 20 December 2008 (UTC)

Implement Flagged Revisions for all BLPs

1. My choice. rootology (C)(T) 19:21, 19 December 2008 (UTC)
2. Choice. But, of course, there are a number of highly divergent ways that flagged revisions could be implemented. CIreland (talk) 19:37, 19 December 2008 (UTC)
3. I have a dream that one day Misplaced Pages will be a respectable source of quality information, and not the punchline of hack comedians. WilyD 19:40, 19 December 2008 (UTC)
4. Makes sense. Durova 19:41, 19 December 2008 (UTC)
5. First choice. Flagged revisions make semiprotection largely unnecessary. Eventually I'd like to see everything in article space using flagged revisions but this is good for now. Also has the advantage that IPs can actually make substantial edits and changes to BLPs they just need to get approved. Also removes the issue of someone starting an article that they can then not edit which occurs if we semiprotect. JoshuaZ (talk) 19:43, 19 December 2008 (UTC)
6. Most anonymous edits to BLPs are good and we don't want to lose those. Flagged revisions is a reasonable way to protect biographies from the bad anonymous edits. -- Ed (Edgar181) 20:02, 19 December 2008 (UTC)
7. We should really be pushing towards this for all articles, actually. The question is whether it will scale to the size of enWP, or even just enWP's biographies, but the deWP experience seems to indicate that it should. Guy (Help!) 20:03, 19 December 2008 (UTC)
8. Yes, yes, yes. If it can be activated on articles selectively, I would personally spread it to all the BLPs I could, and I'm certain that others would help pitch in. Cool Hand Luke 20:31, 19 December 2008 (UTC)
9. Definitely. Vastly preferable to semiprotection - David Gerard (talk) 20:45, 19 December 2008 (UTC)
10. Absolutely, I have long supported this, far better than semi protection (which I would strongly oppose) for the strong reason that it preserves the ability of anybody to edit, while preventing the casual reader from seeing vandalism/libel. Semi protection would either lead to BLPs being hopelessly out of date, a impossible surge in requests for edits and the moving of vandalism to talk pages and other articles. Flagged revisions enables the reader to update/make corrections as necessary which is the great strength of wikipedia and I am confident (and would help ensure) we could keep BLPs flagged quickly. Davewild (talk) 20:49, 19 December 2008 (UTC)
11. Flagged Revisions would be a big step forward. Semi-protecting a substantial portion of our articles is just too antithetical to Misplaced Pages. --Alecmconroy (talk) 20:50, 19 December 2008 (UTC)
12. Second choice. Much preferred to protection. NonvocalScream (talk) 21:16, 19 December 2008 (UTC)
13. Second choice. shoy (reactions) 21:21, 19 December 2008 (UTC)
14. Second choice, or step on the way to where we need to get. ++Lar: t/c 21:37, 19 December 2008 (UTC)
15. Prepared to endorse this. Reasonable balance between dealing with BLPs in a responsible way and encouraging free anonymous participation. Would prefer some selectivity in which BLPs are flagged, but not a deal breaker. Xymmax _{So let it be done} 21:46, 19 December 2008 (UTC)

    This is just to gauge general feelings, the nitty-gritty details could include partial or selective activation of flagged revs on BLPs. (If this isn't yet technically possible, we could buy Brion something nice for Christmas). WilyD 21:53, 19 December 2008 (UTC)

    OK, but the section header said "implement on ALL BLPs" ... Xymmax _{So let it be done} 22:20, 19 December 2008 (UTC)

    Basically, we needed to finally have THIS discussion, rather than pointless little "What if this? What if that?" chats under FLAGGED and the protection policy and 100 other places, to see who actually supports what in general--is all-semi-BLP a good/dumb idea? Is all-flagged-BLP a good dumb/idea? And based on this, the appropriate next steps can then be looked at, and shaped up. The specifics--"Protect all via a transclusion of a widget under the positron matrix"; "Flag rev all via the MacGuffin's inverse matrix", and so on, can be hashed out later and doesn't require 1,000 people to say Yay! or Nay! This is just to see if the ideas even have legs. rootology (C)(T) 22:34, 19 December 2008 (UTC)

16. I don't see how this is technically possible without some sort of BLP namespace, but, yes, do it. John Reaves 21:51, 19 December 2008 (UTC)
17. Probably the best bet. We need to do something about the fact that BLPs of relatively obscure people (which in many cases will be their first Google hit) can be targeted with vandalism that stays for long periods of time. The Siegenthaler incident is probably the most notorious case, but such incidents can and will happen again if nothing is done. At the same time, cutting off all anonymous BLP edits seems to go too far. *** Crotalus *** 21:54, 19 December 2008 (UTC)
18. I was leaning towards semiprotection as the best approach, but Crotalus' response made me think it over a bit. Our biggest problem for BLPs is when a little-watched article is dinged, often by someone with a grudge against said person; these are the ones that come back to bite us in the collective ass most often. Flagged revisions would provide a level of oversight to the process, and ensure that we don't have an issue where the first anyone here hears about a problem is when it's being frothed about in the media. Tony Fox (arf!) 21:59, 19 December 2008 (UTC)
19. Support -- actually I would like it stricter than the current German implementation. Currently "affiliated editors" seem to stake out notable BLPs or groups thereof, and we would need to make sure that the person accepting the flagged revisions was aware of this. Collect (talk) 22:06, 19 December 2008 (UTC)
20. Support. Even short term vandalism or poorly sourced edits can be harmful. We can do better and should. FloNight♥♥♥ 22:12, 19 December 2008 (UTC)
21. I like this option too. — Realist 23:01, 19 December 2008 (UTC)
22. Support, first choice. We try it first on BLP's, then if it helps stabilize content and reduce the amber waves of crap that some of our high-profile BLP articles are hit with throughout their time of notoriety, we expand the flagging to ALL articles.GJC 23:10, 19 December 2008 (UTC)
23. Semiprotecting them all is a bit extreme, but this seems an ideal compromise. IPs aren't stopped from editing, their edits just have to be processed by, say, a rollbacker. I presume we can tweak the settings on a per article basis so it applies to, say, all non-admins for some articles, or so only admins can 'sight' the revisions or something? Dendodge Talk 23:24, 19 December 2008 (UTC)
24. It's worth a try. This may be better than semi-protection because of the essential problem with banning editing by IPs: barriers to editing tends to favor editors with stronger POVs. Flagged revisions could change the POV balance less than semi-protection while still reducing vandalism. ·:· Will Beback ·:· 23:29, 19 December 2008 (UTC)
25. Weak support I am concerned this will be unwieldy in terms of time used up implementing it, but certainly more owrthwhile trialling with BLPs than with FAs or GAs. Cheers, Casliber (talk · contribs) 23:46, 19 December 2008 (UTC)
26. Second choice, prefer flaggedrevs on all pages. Mr.Z-man 00:17, 20 December 2008 (UTC)
27. Support evolutionary step as a reliable encyclopedia. Docku: What up? 00:30, 20 December 2008 (UTC)
28. Support in the alternative; I don't think singling out BLP is that useful, but if we have a choice between flagged revisions on all pages and only on BLPs then I'd rather have at least the BLPs protected. — Coren  00:31, 20 December 2008 (UTC)
29. Agree with Coren above. I'd prefer indefinite semiprotection of BLPs and flagged revisions of all articles, but better this than nothing. – iridescent 00:40, 20 December 2008 (UTC)
30. I think the common wisdom about BLPs is wrong: BLPs are not subject to greatly increased risk of total harm —rather, BLPs are subject to greatly increased risk of complaint about harm. In other words: BLPs are more sensitive indicators of our overall performance than other types of articles, but by themselves are probably only somewhat more risky than average. In a great many non-BLP articles there is a risk of vandalism causing extreme harm: consider a statistics article vandalized in an unfortunate way at an unfortunate time which results in an engineer miscalculating the safety of his design. In that case there is no singular obviously wronged person to notice and complain that his article is incorrect—so we may not feel the pain of the vandalism but a great many people may be harmed. Moreover, the impact of significant harm caused by problems in BLP articles is generally limited to a few people (the subject and his close associates), while errors in other areas can impact more people. I've witnessed firsthand an error from Misplaced Pages making its way into a software design document. While not itself a safety-impacting error, it was enough to make me sure that safety-impacting errors do occur. Even to those of us who care only or mostly about BLPs, flagging only BLP articles is clearly insufficient to protect the subjects: if a vandal changes Bomis to call it a child porn site, the actual harm to Jimmy would be just about equal to the harm caused by the same claim placed directly in the article on Jimmy himself. I think it's unfortunate that we continue to special-case BLPs, since when we do that we're killing our best indicator of potential harm throughout the project while only making a fairly limited improvement. All that said, if people wanted to flag only BLPs as a testing point, I wouldn't oppose it, since we do need to start learning about the impacts of flagging. I do oppose flagging+semi at this time, but only because I think we should make one change at a time in order to measure the results. --Gmaxwell (talk) 01:01, 20 December 2008 (UTC)

    I think you're wrong about potential harm. When people want to know about Jimmy Wales (or any individual), they do not type "Bomis" into a search engine. They type the person's name. If there's something harmful in their biography, it is much, much more likely to reach the eyes of people curious about that person, thereby causing harm. That said, thank you very much for not being opposed to BLP flagged revisions as an experiment. Cool Hand Luke 01:20, 20 December 2008 (UTC)

31. This seems ideal to me, if technically feasible. Deli nk (talk) 03:07, 20 December 2008 (UTC)
32. I could support some 3-month or 6-month trial of this to see what happens. I would prefer this to semi-protection for all BLPs. As of right now, Category:Living people has 324,956 articles in it. About 1 in 8 articles on the English Misplaced Pages is a BLP. Looking at the recent changes to BLPs, it looks like a BLP is edited about 12 times per minute, or once every 5 seconds. I would prefer we try out FlaggedRevs on BLPs first, before we even think of enabling it on all articles — to see how out of date the 'sighted' BLPs get and what kind of backlog results. But I think there needs to be a clear plan in place (how someone gets/loses 'reviewer' status, how many 'reviewers', when and when not to 'sight'/'unsight' a page) before it goes into effect. Vandals could still game the system, but I guess if you make it a real chore for them you could dissuade most of them. But once FlaggedRevs is enabled on all BLPs, will all BLPs be 'sighted' as is, or will 'reviewers' have to review them all one by one? If they need to first be reviewed one by one (which I think is the case), how long will it take to do the first sweep ('sight' some revision of all 324,956 BLPs), and who's going to do it? --Pixelface (talk) 03:43, 20 December 2008 (UTC)
33. Second choice, some improvement, but a little clunky. MBisanz 03:49, 20 December 2008 (UTC)
34. My first choice. As mentioned above, semi-protection may inadvertently encourage editors with a stronger POV (i.e.: the barrier to edit is higher); and affect the neutrality of our articles. I hope that we move to try this solution before semi-protection. Lazulilasher (talk) 04:17, 20 December 2008 (UTC)
35. Seems to be the best course of action, keeping the "anybody can edit!" philosophy intact, while also affording more protection to the subjects of BLP articles. Definitely worth a trial, if nothing else. Lankiveil 05:00, 20 December 2008 (UTC).
36. This is the best alternative over two extremely unsatisfactory situations: semi-protection on all blps that would be an extraordinary break in our open-editing wiki spirit, and a situation where blps are largely unmonitored and prone to vandalism and wp:blp violations. I think it's manageable, but we'll still probably face huge backlogs. Cenarium (Talk) 14:12, 20 December 2008 (UTC)
37. Support. This is the perfect middle ground. It allows editing, but at the same time builds in a safeguard against vandalism. The only possible problem is scaling, but we can't dismiss it before a trial has been done. - Mgm| 14:26, 20 December 2008 (UTC)
38. Support There is a lot to discuss regarding the details, but some form of FR for BLPs seems like the best option to me. --Tango (talk) 14:30, 20 December 2008 (UTC)
39. Second choice.Peacock (talk) 15:22, 20 December 2008 (UTC)
40. Flagged revisions are very promising, but it is a big change so we need to go slowly and carefully. There are real concerns about backlogs which need to be addressed and trials are probably the only way to find out. And of course there are many details to be filled in. -- Jitse Niesen (talk) 16:16, 20 December 2008 (UTC)

    It's known that the details need to be filled in. What we want to find out is whether it's worth trying to fill in those details, by seeing how people feel about flagged revisions, whether it's likely the community will give a mandate to implement them once the details are worked out. WilyD 16:41, 20 December 2008 (UTC)

41. Conditional Support -- I think flagged revisions are a wonderful idea, but should be strictly kept to featured articles (helps us look good) and biographies of living people (legal reasons). I am strongly opposed to them going farther than that, I think it would break the spirit of wikipedia and turn it into a cool kids club of people who can commit revisions. --ScWizard (talk) 08:10, 20 December 2008 (UTC)
42. Support I completely agree with ScWizard above. -- lucasbfr 17:19, 20 December 2008 (UTC)
43. Support; appears to be an excellent idea, and if it turns out to be unworkable we can turn it off. It's worth a try. I also think we should only try one thing at a time: this would be a good preliminary test for flagged revisions in general on en:. Antandrus (talk) 17:25, 20 December 2008 (UTC)
44. Weak and qualified support—while on the one hand I strongly support the general use of FlaggedRevs, it is my impression that its use here implies a sighted-revisions-visible-by-default implementation of FlaggedRevs, which I strongly oppose. FlaggedRevs would be helpful in combatting vandalism, but I can't shake the feeling that a sighted-revisions-visible-by-default implementation constitutes effectively denying editing to those not allowed to sight edits (in other words, most people). I'd accept sighted-by-default FlaggedRevs to any more extreme protecting-BLPs proposal, however, as it's got many advantages over other schemes. I must note that BLPs are one of the few areas where sighted-by-default FlaggedRevs would be acceptable to me: I don't want any implementation on BLPs to be used as an excuse to enable sighted-by-default FlaggedRevs elsewhere. My preferred implementation of FlaggedRevs would make the "sighted" or "unsighted" status visible as a means to aid more basic BLP-monitoring efforts. Sighted-by-default FlaggedRevs would be preferable to semi-protection, in any event. {{Nihiltres|talk|log}} 18:10, 20 December 2008 (UTC)

Oppose Flagged Revisions for all BLPs

1. There is no way for the use of stable/flagged/cabal-approved revisions to be limited to one category, and it wouldn't make any sense to do so as anyone can add or remove a category. This would be a social restriction rather than a technical one, and I see no reason to honor it really. What purpose would be served by writing a policy that says "this feature should only be used for living people, other articles are unworthy"? Oppose in favor of explicitly allowing edit-flagging on any and all content pages. — CharlotteWebb 23:03, 19 December 2008 (UTC)

   Do we know for a fact this technical limitation can't be changed? rootology (C)(T) 23:07, 19 December 2008 (UTC)

   I just found out we can indeed do this from a technical standpoint. People can certainly remove a BLP category, but thats sort of a red-flag action in and of itself that would also be limited by the flagging process. rootology (C)(T) 23:10, 19 December 2008 (UTC)

   It would be a technical restriction if the 'surveyors' agreed to only enable FlaggedRevs on articles in Category:Living people, or if you had a bot with 'surveyor' status do it. --Pixelface (talk) 03:48, 20 December 2008 (UTC)

2. Strong oppose as someone who is continuously vigilant of a conservative reading of our blp policy, as both going against our "anyone can edit policy" and making it harder to actually deal with random blp vios. I often spot blp vios reading an article on somebody I have just read about elsewhere and the idea that I and others could not do this kind of blp removal IMMEDIATELY just seems completely counter-productive. Plus its a huge time waster. Thanks, SqueakBox 23:24, 19 December 2008 (UTC)
3. Strong oppose, but I will not be opposed to an X month trial so we can see how bad backlogs really get. I suspect this will go the same way as the New Page Patrolling backlog (ie, dies). - NuclearWarfare _{My work} 01:21, 20 December 2008 (UTC)
4. Oppose - unless an exception was granted for high-edit and/or articles tagged with a "recent event/death/whatever" tag. --Philosopher  01:46, 20 December 2008 (UTC)
5. Oppose; I do not see what flagging revisions will do except for irritate vandals and provoke them to become more egregious in their vandalism. This is especially of concern with BLPs, and flagged revisions more like than not will not affect vandal pagemoves (a common JA/G tactic). We need to be able to see and remove vandalism on sight, registered account or otherwise, and this would firmly plant anons in a second-class citizen situation. Flagged revisions at present are too double-edged - good anons cannot reverse the actions of their crueler counterparts unless they access and read the history (which a lot of anons likely do not know how to do), and good anons updating an article will constantly repeat their edit or become a vandal when they find that the good edit they've just made doesn't show up, even after server-side delay is factored in. -Jéské Couriano 03:42, 20 December 2008 (UTC)
6. Oppose I'm irritated by the invocation of the Wales at the top of this page, but that isn't the reason. Flagged revisions, when they come to wikipedia, should be trialed in a much smaller and more controlled area than BLPs. Further, if we don't have the willingness or editor resources to revert vandalism to these articles (ostensibly the reason we are pursuing this), how the hell are we going to flag all of these revisions as sighted? I'm not going to. If I see an article edit on the RC feed I won't give it a second thought unless it is obviously vandalism, then I will revert it. That's basically how Huggle stops most vandalism and it does so very quickly. The stuff that doesn't get caught is plausible info that upon deeper reflection, is vandalism. No technical process will make those edits more scrutable. What we will find ourselves with will be hundreds to thousands of articles with revisions in the sighting queue and no motivation to review them. Those edits will never make it into mainspace without being sighted so the motivation to pay some cursory look at them for the sake of keeping obvious vandalism out is gone as well. Aside from the effectiveness issue, I still don't understand how this notion of sighted revisions helps us gain new adherents or live up to our guiding principles. We can bullshit ourselves into thinking "sighted revisions means we can unprotect George W. Bush", but the same number of anonymous edits to that page will stick with protection as without. And in return we will have blocked off anonymous access to every MLB, NFL and NHL player page (or, placed that access behind the gatekeeper of a revision sighter). Great. Protonk (talk) 05:30, 20 December 2008 (UTC)
7. Oppose again this is overkill. We are the encyclopedia anyone can edit. Hut 8.5 10:18, 20 December 2008 (UTC)
8. Oppose Not yet. We should try semi-protection first. And then perhaps we wil still thik we need ths--and if we do, it should first be implemented on a vulnerable subset so we can see how practical it is.DGG (talk) 18:49, 20 December 2008 (UTC)

Implement Flagged Revisions for all articles / content pages

1. This would include templates, images, categories, and anything else that might somehow be part of the dependency tree of any article or portal page, etc. — CharlotteWebb 20:33, 19 December 2008 (UTC)
2. An important move to make in making this an encyclopedia. Has worked well on de:wp - David Gerard (talk) 20:46, 19 December 2008 (UTC)
3. I would support but doubt consensus would be reached without first seeing it implemented in some articles such as BLPs. Davewild (talk) 20:50, 19 December 2008 (UTC)
4. If it were possibly to do this and get it through, it would be a good thing, but not on Misplaced Pages space, please (at least at first). rootology (C)(T) 20:56, 19 December 2008 (UTC)
5. This is where we need to get. Maybe this isn't the first step, but this is the destination, in my view. ++Lar: t/c 21:21, 19 December 2008 (UTC)
6. I'm prepared to be persuaded, but I'm going to have to see how this functions in the wild first. Let's start with problematic BLPs and see how it goes. Xymmax _{So let it be done} 21:50, 19 December 2008 (UTC)
7. Third. I do not know how readily it could be implemented with the large number of minor edits on non-controversial topics. Perhaps the limit should not be on the "topic" but on the activity therein -- if an article gets more than (say) ten edits a day that it goes into "flagged" status? Collect (talk) 22:09, 19 December 2008 (UTC)
8. To move the emphasis to the quality of all content. FloNight♥♥♥ 22:16, 19 December 2008 (UTC)
9. Moreschi (talk) 22:17, 19 December 2008 (UTC)
10. bibliomaniac15 22:54, 19 December 2008 (UTC)
11. Prefer this, though I support implementing flaggedrevs in pretty much any way. Mr.Z-man 00:18, 20 December 2008 (UTC)
12. Support; this will allow stability without sacrificing the ability to edit as semi protection schemes necessarily must; it will also greatly reduce the number of times we have to protect at all. — Coren  00:33, 20 December 2008 (UTC)
13. Yes. It has worked well on .de and on the English wikinews as well as a variety of other projects. JoshuaZ (talk) 03:24, 20 December 2008 (UTC)
14. Third choice, weakly, we don't have the manpower to implement this idea, but if it will help the BLP issue, we must adapt. MBisanz 03:50, 20 December 2008 (UTC)

    Essentially to get rid of vandalism, libel, most blp issues, etc. Not too soon though, I am afraid that we might rush things and ultimately screw up. But only with an expiration system to prevent inevitable backlogs. Cenarium (Talk) 04:00, 20 December 2008 (UTC)

    Indent, as explained, I prefer the option to #Implement Flagged Revisions for blps and Flagged revisions with expiration for all non-blp articles / content pages. Cenarium (Talk) 14:16, 20 December 2008 (UTC)

15. Support Possibly with a different implementation for BLP vs non-BLP articles, but we can discuss that later. --Tango (talk) 14:31, 20 December 2008 (UTC)
16. First choice. Peacock (talk) 15:23, 20 December 2008 (UTC)

17. Support iff the implementation is not sighted-revision-viewed-by-default, for all users and all pages. While I think that flagging revisions is probably a good idea, making sighted revisions what's visible by default effectively denies editing to those without the ability to sight edits. Even if we could handle the backlog of unsighted revisions (our ability to do so is dubious), the disincentive to editing for anonymous users would definitely be a bad thing. It's been shown by academic studies that anonymous and infrequent users contribute a significant amount of our content, and I think that sighted-by-default FlaggedRevs project-wide would disrupt that positive influence. For all that it would effectively destroy vandalism, we could do that any number of ways which would be obviously unacceptable by our foundation principles, and I think that this is one of those ways.

    That being said, I do think that FlaggedRevs has potential, for BLPs and elsewhere. Showing, to some extent, what's been reviewed and what could be vandalism—that is, the status of being sighted or not—would be a good step in the direction of visible quality without sacrificing the ability of most people to effectively edit. I find that being able to selectively, but not systematically enable sighted-by-default for particular pages liable to much vandalism or libel might be worthwhile, as a different form of protection which could be applied to articles, such as BLPs, for example, as an alternative to semi-protection. {{Nihiltres|talk|log}} 18:39, 20 December 2008 (UTC)

Oppose Flagged Revisions for all articles / content pages

1. Strong oppose, ludicrous idea that would be unworkable and make wikipedia the website not to edit. Thanks, SqueakBox 23:25, 19 December 2008 (UTC)

Flagged revisions do not restrict the ability to edit the page. - Mgm| 13:35, 20 December 2008 (UTC)

1. Oppose: It would make us more reliable, at the cost of being less informative. We'd get massive backlogs, which no-one will want to tackle. Because of these massive backlogs, newbies will complain that their edits aren't being processed and then leave us. Then we have even less editors to handle the growing backlog. Rinse, lather and repeat. Dendodge Talk 00:04, 20 December 2008 (UTC)

   Thanks for that, a cogent explanation of the problem this would cause. Thanks, SqueakBox 00:25, 20 December 2008 (UTC)

2. Crazy strong oppose Yeah, lets increase the existing backlog by a factor of 12. I'm not sure how many articles we miss at NPP, but I'd estimate 150/day, at least, translating to 4500/month. At least those articles are still allowed through. I shudder to think of the backlogs that have to be gone through before edits from, say two months ago are let through. - NuclearWarfare _{My work} 01:26, 20 December 2008 (UTC)
3. Oppose without some kind of trial first. This could either be a small-number-of-pages trial, or a "flag but show most recent version" type test as I proposed elsewhere on this page. davidwr/(talk)/(contribs)/(e-mail) 01:34, 20 December 2008 (UTC)

   ...and does absolutely nothing to help the BLP issue. Many readers don't even notice the edit tabs, why would they notice a "prior flagged revision available" note? Why would anyone even care to flag such articles? What is the purpose, exactly? Cool Hand Luke 01:43, 20 December 2008 (UTC)

4. Templates Only, see above for BLP thoughts--Philosopher  01:47, 20 December 2008 (UTC)
5. Oppose for now, see how it works on BLPs first. About 1 in 8 articles is a BLP, and a BLP is edited about once every 5 seconds. 156,430 users have made at least one edit in the last 30 days. Even if you made all of them 'reviewers', is that enough people to 'sight' all 2,665,657 articles? --Pixelface (talk) 03:19, 20 December 2008 (UTC)
6. Lor no; see my opposition to flagging BLP's above. Flagged revisions are too much stick and too little carrot. -Jéské Couriano 03:43, 20 December 2008 (UTC)
7. ABSOLUTELY NOT. Yeah I used caps. I really do think this would be a very bad idea, for many reasons, most of which have been said again and again. Wizardman 04:27, 20 December 2008 (UTC)
8. Strong Oppose - the backlogs would probably explode... VX! 05:04, 20 December 2008 (UTC)
9. Uhhh How many permutations are there going to be? I guess I'll note here that I'm opposed to implementing flagged revisions elsewhere but my rationale is at one of the various "Whatever it is, I'm against it" sections above. This is confusing. Protonk (talk) 05:32, 20 December 2008 (UTC)
10. Oppose Disaster waiting to happen. Let's not go there. PeterSymonds (talk) 09:50, 20 December 2008 (UTC)
11. Oppose this will prevent huge numbers of constructive editors from contributing and is directly opposed to the spirit of Misplaced Pages. Hut 8.5 10:22, 20 December 2008 (UTC)
12. oppposeWe simply don't have the rescources.Geni 13:10, 20 December 2008 (UTC)
13. Oppose until a trial has been run and there's evidence we actually have the resources to handle it. - Mgm| 14:28, 20 December 2008 (UTC)
14. Strong Oppose - Would the site still be wikipedia after such a monumental change? I mean last I checked the front page said: "Welcome to Misplaced Pages, the 💕 that anyone can edit." --ScWizard (talk) 17:06, 20 December 2008 (UTC)

Implement Flagged Revisions for blps and Flagged revisions with expiration for all non-blp articles / content pages

1. This is the best course of action: it will avoid harmful backlogs by automatically making visible to IPs revisions that are old enough on non-blps. While blps require more scrutiny, most articles comparatively don't and applying flagged revs without expiration on all would certainly create huge backlogs and prevent a mass of edits from going live. This would still allow to get rid of the quasi-totality of vandalism and other disruption, since most is reverted within a few hours. With the understanding that this is very flexible: it can be deactivated on certain pages, be prevented by the abuse filter, etc. Cenarium (Talk) 13:49, 20 December 2008 (UTC)
2. Best option, since the "backlog" issue becomes moot. rootology (C)(T) 14:11, 20 December 2008 (UTC)
3. Agreed (rather like what I supported before). Obviates all concerns about manpower. Collect (talk) 14:25, 20 December 2008 (UTC)
4. This is my new favourite, as long as the expiry is no longer than 24 hours. This would allow us to prevent many of the bad edits from even happening, without creating massive backlogs. The one problem is whether or not it is possible. Dendodge Talk 17:20, 20 December 2008 (UTC)

Comments

This section is silly. This is basically just a "preliminary" discussion. Throwing in more options similar to existing options is just going to split people up even more, make consensus look even more fuzzy, and decrease the chance we get anything decided here. These are the kind of details that should be worked out if/when we decide we want flaggedrevs at all. AFAIK, "expiration" isn't even an option yet in flaggedrevs. Mr.Z-man 17:06, 20 December 2008 (UTC)

Implement both Semi-Protection and Flagged Revisions for all BLPs

• Comment. Is this not redundant? As I understand them, Flagged Revs are only placed for IPs, so if it's semi-protected, what's the need for FRs? SDJ 19:31, 19 December 2008 (UTC)

1. See my objection to semiprotection. Durova 19:46, 19 December 2008 (UTC)
2. Isn't the above a comment, Durova? WHO sees flagged revisions first is an option, it could be set to "everyone" so it is possible to have and need both SP and FR. ... suppor this. ++Lar: t/c 21:23, 19 December 2008 (UTC)
3. Lar makes a very good point. The two are not mutually exclusive; one is presentation, the other control. Support (which I didn't think I would, even though I added this section header myself!). rootology (C)(T) 21:33, 19 December 2008 (UTC)

   I consider myself progressive on this issue, but doesn't FR make semi-protection redundant (assuming that the flagged version is the reader default)? Cool Hand Luke 21:46, 19 December 2008 (UTC)

   I took it to mean that... BLPs are semi'd--the drive-by vandalism by IPs is a thing of the past. That still leaves sneaky vandals on usernames that are past the semi-limit, but wouldn't have the "Flagger" ability or whatever we'll end up calling it. So, it would be a double layer of protection, like an Oreo, with the BLP as the creamy center. The benefit of this is casual readers (our real userbase, not us) will only ever hopefully see a "good" version, without the chance of any crap snuck in by untrusted users OR anons. If I've butchered the technical aspects of this, someone please correct me. rootology (C)(T) 21:51, 19 December 2008 (UTC)

   I think it is fairly redundant, unless it's taken to mean "Semi today, Flagged when we work out the details". WilyD 21:54, 19 December 2008 (UTC)

   I'm with WilyD here. I support semi as a stop-gap to work out the details of flagging. I suspect that most sneaky vandalism actually comes from savvy throw-away accounts (flaggers would probably be more comfortable approving edits from a blue-link user), such that using both protections is redundant. Cool Hand Luke 22:07, 19 December 2008 (UTC)

   Lemme clarify that I'm not exactly endorsing that position, just explaining it. I'm conflicted about semi-prot. On the one hand, we clearly need to do a lot more than we're doing. But the barrier to entry is high. It's a tricky cost-benefit. Flagged is just soo much cleaner. Semi-protting is worth exploring, but I'm tepid. WilyD 22:18, 19 December 2008 (UTC)

I don't like any of these options but something has to be done for BLPs

Leave your alternate idea here

1. For all articles, show the current version but have a link at the top to the last-flagged version. This would alert users immediately that the version they are looking at has not been reviewed and give them an opportunity to go to the last reviewed version. davidwr/(talk)/(contribs)/(e-mail) 20:35, 19 December 2008 (UTC)
2. Spread the word that Special:RecentChangesLinked&target=Category%3ALiving_people and other patrols exist. If that doesn't work to reduce the time vandalism stays on biography articles, then look at more severe actions like making BLP articles flagged-revision or semi-protect by default. "Use a light touch first, use more pressure only if necessary." davidwr/(talk)/(contribs)/(e-mail) 20:35, 19 December 2008 (UTC)
3. I don't like the idea of semi protecting anything until such time as is demonstrably required. I don't like the idea of not allowing anonymous editors in, automatically. As a suggestion, perhaps categorize all BLP as a BLP and then place a noticebot in a public IRC channel. That would be a suggestion, but as for protection, I don't think I like the idea. NonvocalScream (talk) 21:15, 19 December 2008 (UTC)
4. We just need to enforce a conservative interpretation of blp, its only cos loads of editors don't want to that we see this problem. Thanks, SqueakBox 23:26, 19 December 2008 (UTC)

   Keep in mind that the problem of vandalism is completely distinct from the problem of what material should actually be in an article when that material is well-sourced. Let's not confuse them. JoshuaZ (talk) 02:06, 20 December 2008 (UTC)

5. Endorse Davidwr's idea. --Philosopher  01:48, 20 December 2008 (UTC)
6. A while ago I did a survey of how long it took to revert some vandal edits (see User:Hut 8.5/vandals). The vast majority of edits were reverted extremely quickly. However a handful of edits were missed by RC patrol and survived for much longer, and presumably it is these which pose the biggest risk since there is a much greater chance that someone will see them. In order to catch these edits we need some way of marking an edit as "patrolled", which I understand is already done on some wikis. Alternatively we could use flagged revisions while still giving all users the most recent version to read. Hut 8.5 10:30, 20 December 2008 (UTC)
7. Alternate idea - If you have the capability to edit semi protected articles, your edit goes straight though. If you do not then your edit gets flagged. Kind of a cross between "flagged revisions for BLPs" and "semiprotect BLPs." --ScWizard (talk) 17:10, 20 December 2008 (UTC)

Leave BLPs exactly as they are

1. Since this is the closest thing to a "both options are bad" section, let me just say that (absent context), both options (semiprotection and flagged revisions) are bad. Semiprotection would prevent anonymous users from just casually removing obvious BLP violations whenever they see them, which would make such violations harder to fix, and keep them on the article longer. Flagged revisions would mean that when an IP acted to remove BLP-violating information, we would nonetheless continue to display the bad content until some approved user allowed the article to be fixed. Either one implies that you must be authorized to remove libel - that's Bad Bad Bad Bad. I realize that this badness isn't the poller's intention; it's only what will actually happen. We shouldn't stop IPs from fixing things. — Gavia immer (talk) 19:50, 19 December 2008 (UTC)

   How do you propose we stop IPs from breaking things? IPs are more likely to add violations than remove them. Cool Hand Luke 20:34, 19 December 2008 (UTC)

   The same way we do it now, which mostly works (ANI/AIV reports are not a representative sample of all IP edits). My experience is that most IP edits are generally trival (neither especially good nor especially bad), with an admittedly visible volume of test edits. My further experience is that most bad IP edits are either spam/SEO crud or experienced users attempting to avoid scrutiny. Restricting casual editing of BLPs because of a (very visible) minority of edits is the wrong way to do things. I do think there are addressable deficiencies in our monitoring BLP-sensitive articles; I just don't think these proposals are the way to go. — Gavia immer (talk) 20:55, 19 December 2008 (UTC)

   From my experience, with the exception of highly visible pages (Governors, cabinet members) where vandalism is reverted quickly, IP editors are actually more constructive than destructive, hence (of course) the idea that Misplaced Pages is the encyclopedia that anyone can edit. --Philosopher  01:51, 20 December 2008 (UTC)

   I hadn't thought of that, but flagged revisions would mean that if an IP saw BLP-violating information, then a 'reviewer' made a bad 'sight' (which is bound to happen I suppose). --Pixelface (talk) 05:13, 20 December 2008 (UTC)

   Right - there would be mistakes, oversights, approvals made too speedily, etc. I myself have accidentally rollbacked an article to a heavily vandalized version because I didn't notice another editor squeezing in a rollback before me. There would also be the problem that today's puppetmaster with carefully autoconfirmed accounts can be tomorrow's puppetmaster with carefully aged accounts and approval to make revisions sighted - if someone wants to get this, they'll get it, however slowly. — Gavia immer (talk) 15:41, 20 December 2008 (UTC)

2. Gavia has said it really well. - NuclearWarfare _{My work} 01:33, 20 December 2008 (UTC)
3. I agree with Gavia's concerns. This is a problem I hadn't considered, and I originally came here with the intention of endorsing semi-protection, but this argument has changed my mind. JulesH (talk) 10:46, 20 December 2008 (UTC)

Polls are evil

Keeping in mind that this isn't a poll, it's a survey to see which ideas are most popular, and worth looking more at later, and keeping in mind there isn't a more practical way to get a wide cross-section of users to simply sound off

1. Dogmatically: "All polls are evil", but some are more evil than others. This is among the least evil ever :) ++Lar: t/c 21:24, 19 December 2008 (UTC)

   What makes this poll especially evil? We've discussed this matter for over a year, and a poll seems like a good way of assessing community support of moving forward with implementation of one of these proposals. How should we decide? More inconclusive threads at the Village Pump? A fiat from Jimbo? If there's a better way please suggest it. ·:· Will Beback ·:· 23:51, 19 December 2008 (UTC)

   It's traditional to have "All polls are evil".... somebody had to vote for it. :) (Note I've voted in several of the other categories too) Personally I think (as I said before) it's about as unevil of a poll as you can get. Because, maybe this time it would be the start of actually implementing some very needed changes. Hope that helps. ++Lar: t/c 00:53, 20 December 2008 (UTC) (PS I don't actually think all polls are evil, myself)

2. All polls are , but some polls are more than others. This one is just fine. :) - NuclearWarfare _{My work} 01:34, 20 December 2008 (UTC)
3. Why no to have an open discussion without the dreadful oppose/support? The result of these so called "polls" is that they die sooner or later and nothing happens. Why? Because they polarize views. ≈ jossi ≈ (talk) 18:50, 20 December 2008 (UTC)

General comments

Comment Just to break the monotony of the structure of this poll/survey/whatever that do not allow discussion or the arguments/counter arguments cycle. As you may be aware, there is a proposed trial of flagged revs on featured articles. The class of BLPs is too huge for a trial and it would certainly be impossible to handle without a first experience and a phase of analysis and resolution of problems. So indeed, sighted revs on all articles, and especially blps, is to be considered and planned, but before we need a trial and a substantive reflexion on flagged revisions. The primary problem with flagged revs on the English Misplaced Pages is that due to its size and the very high number of editors, backlogs will become excessively large and so edits, good ones, won't be flagged for extensive periods. Allowing someone to edit and then not displaying the revision in a timely manner on a huge number of articles is actually worse than not allowing to edit on a limited number of articles. Very likely, it will discourage IPs and new users from editing Misplaced Pages, and prevent addition of good material. Now, there is a solution: automatically flagging revisions after a certain period of time, for example 18 hours, with a possibility to adjust the delay depending on the backlog, deactivate it on certain pages, for example on blps that have proved problematic, make the delay longer for blps, create a special page listing only blps, allowing the abuse filter to prevent automatic sighting for certain edits that are very likely to be vandalism, etc. We could also create a higher level of flagged revisions (like semi protection and full protection), with a smaller usergroup able to flag with this level on articles where it is enabled, for articles with serious problems. A comment on newly created blps too: those can't be handled by sighted revs before being sighted for the first time, if ever. So we still need to filter at Special:Newpages. When a blp is identified as such, for example through categories, that can be detected by the software, and has been sighted, it can then be monitored by flagged revisions in a specific manner. Cenarium (Talk) 23:44, 19 December 2008 (UTC)

For the purposes of a test, it could always be done first just on one letter--flag all the "T" BLPs, or the "As", for a week or a month, to see how it goes. rootology (C)(T) 00:50, 20 December 2008 (UTC)

No. The most damaging sort of edits are the ones that survive RC patrol, so I don't think any kind of automatic flagging is a good idea. I'm unclear why everyone claims that backlogs will be huge (an untested assumption, honestly). Isn't flagging a user right that can be distributed like rollback? Cool Hand Luke 00:55, 20 December 2008 (UTC)

German Wikipedian backlogs are huge, I've heard. As of a few months ago, I believe they were at a 100,000 articles that still needed to be checked though. I like the system put forth here though. If the most recent version is showed, but with an icon noting that it hasn't been checked yet, that might be far more useful. - NuclearWarfare _{My work} 01:41, 20 December 2008 (UTC)

This page seems to say that 98.91% of all sighted articles are on currently reviewed. It appears that some articles have never been sighted, but for over 775,000 articles they are more-or-less up-to-date. Cool Hand Luke 01:57, 20 December 2008 (UTC)

The first sighting of each article takes special care, as it explains here. This page shows that all of the articles are getting sighted. Once sighted, they go on a worklist, and they appear to be kept up-to-date. The column on the far right shows articles previously sighted that are not up to date. Note that this isn't increasing, even as all the pages are getting sighted; it never goes above 8000 or so. Anyhow, the German experience shows diminishing backlogs, not increasing ones. Cool Hand Luke 02:20, 20 December 2008 (UTC)

Even so, 8000 old-reviewed articles, with unreviewed edits, most of them being days-old and some weeks-old, preventing revisions from going live, is far from negligible. The English Misplaced Pages has a much quicker editing rate, especially from IPs, and not a lot of 'regulars' in comparison, so it's likely the backlog will grow much higher. The usergroup can be distributed, but it's also a reason we need, not only a trial, but a progressive implementation: at the beginning, we'll only have rollbackers and admins to flag edits, largely insufficient for all blps. If you are referring to sort of edits that led to the Seigenthaler incident, then having an 'expiration' (automatic flagging) system or not won't change anything: new unreviewed articles will be listed at Special: Newpages when filtering out reviewed pages. While blps are more sensitive and require more oversight, most articles don't, so if we intend to use flagged revisions on all articles, an expiration system is a necessity. For blps, it may be delayed longer, or deactivated if necessary. Expired revisions would also be dealt with differently than usual sighted versions:with distinguishable signs, a specific special page to list them, etc. We could also make a special page specific to blps. And Huggle could have an option to sight a revision when the previous one is sighted, but it would still be insufficient for all blps. Cenarium (Talk) 03:30, 20 December 2008 (UTC)

There are 400,000 BLPs, perhaps (probably less). German Misplaced Pages is already successfully sighting 775,000 articles with less users than we have. I don't think there's a bottleneck here; it's at least worth a trial run. Cool Hand Luke 03:33, 20 December 2008 (UTC)

However, you are missing the fact that articles on en are edited much more often. We can't only consider numbers anyway, we need to implement to see how it will turn out, and prudence demands a progressive implementation, otherwise we will be flooded, and a trial before that. I have already worked intensively on flagged revisions and I am not opposed to it, but try to find ways to adapt it to Misplaced Pages, in a pragmatic manner and respecting our wiki nature. Cenarium (Talk) 04:07, 20 December 2008 (UTC)

We can also just try for one letter to see how it goes. A went smooth? Lets add B? Still good? C and D, etc. rootology (C)(T) 03:37, 20 December 2008 (UTC)

Yes, the implementation needs to be progressive to give the time to grant reviewer rights and adapt, the alphabetical order looks acceptable, with the occasional exception. Cenarium (Talk) 04:07, 20 December 2008 (UTC)

Implementation would be inherently progressive as articles wouldn't be flagged to begin with. That is, for articles where there is no flagged revision yet, the most recent one would be shown, and edits would take effect immediately. As BLPs would begin to be flagged, those BLPs that are would show the most recent flagged version first. That's how the German Misplaced Pages managed to scale the system.

In addition, depending on the mechanism being used to implement BLP protection, it could be applied incrementally. FlaggedRevs already supports per-page configuration of the default viewing level. Essentially the "show most recent flagged version first" setting can be seen as an alternative to semi-protection, and can be used as such on a per-page basis.

I'm writing some notes on the German experience at m:FlaggedRevs Report December 2008.--Eloquence* 04:16, 20 December 2008 (UTC)

Excellent; I look forward to your thoughts. Yes, I like the German system of incrementally reviewing new articles into flagged revisions. Cool Hand Luke 05:19, 20 December 2008 (UTC)

This is interesting, thanks. Some useful and significant data has been gathered, but it seems still too soon to see the impact on user contributions. I do not believe that 10000 articles is a reasonable number though.

Indeed the implementation would be progressive in that sense. At the very beginning of the implementation, admins should be allowed to enable flagged revs on blps, essentially to be used on high-visibility or problematic ones. Then in a second step, we could enable Flaggedrevs by default, either massively or incrementally. The advantage of an incremental implementation, for example following the alphabetical order, is that, since most blps are likely to be sighted randomly, it'll relatively keep things 'under control', avoid dispersion, and focus the attention of users for feedback purposes (in anticipation to an extension to all articles, in particular). Of course, a trial prior to the implementation would provide sample analysis.

The backlog is the primary problem, especially in the event of an extension to all articles. I believe that a system of expired revisions could precisely resolve it: if the backlog is huge, we reduce the delay to expiration (for non-blps), if it becomes less important, we can augment the delay. Expired revisions can still be flagged like any other, the only effect is that it becomes the stable version (the version showed to IPs). Cenarium (Talk) 05:56, 20 December 2008 (UTC)

No, revisions should not be sighted randomly. In the German Misplaced Pages, articles are not sighted at all until someone flags them for the first time (after an extra-thorough review). By this method, 775,000 articles have been flagged. That's how we should tackle BLP flagging. I strongly oppose any expiration for flagged revisions. On the German Misplaced Pages, 98%+ of flagged articles are up-to-date. Cool Hand Luke 06:50, 20 December 2008 (UTC)

Revisions will be sighted randomly. Articles are seen by a user randomly, either through recent changes, viewing of an article, etc, and then the user decides to flag a revision of not. This is completely random, with a tendency for most viewed and most edited pages. You'd have to create a process to organize the flagging of blps and it won't be up and working in the beginning, and the random factor will still be there. Of course an article is not sighted until someone does (or automatically from the beginning if the creator is a reviewer), this is always true, it's the way flaggedrevs work, not only on de. Your numbers are not meaningful at all, we can't compare the situation with de, due to our size, our ratio regulars/all editors that is much weaker, etc. Have you considered the opposition to flagged revisions above ? Wouldn't you prefer flagged revs with expiration for non-blps rather than nothing ? You'll never get a consensus for flaggedrevs on all articles without an expiration for non-blps and for other content pages (templates, images, etc). Cenarium (Talk) 13:29, 20 December 2008 (UTC)

I disagree with any automatic flagging of revisions. I think automatic flagging brings the worst of both the current situation and flagged revisions. Not only does the person who makes the edit have to wait for a period of time for his edit to be flagged but we do not have the benefit of flagging - the stopping of vandalism/libel from entering the flagged article. Imagine the situation where news organisations pick up on articles where vandalism/libel has been flagged automatically as good edits. If we cannot cope with the amount of edits to flagged articles then we should scale down the number of articles to be flagged, equally if we are coping well then we can scale up the number of articles flagged. I believe we will be able to cope with flagged revisions on BLPs but I see no benefit to introducing flagged revisions that are automatically flagged on other articles. Davewild (talk) 15:56, 20 December 2008 (UTC)

I'm not sure we should be giving special technical (e.g., restricting editing through the software) treatment to BLPs. While we can reduce edits to these pages (by implementing semi-protection) or monitor additions more carefully (and we really should be doing that better), I think removing unsourced statements that currently sit in BLPs for months or years is perhaps more important than anything else. So I'm spamming Misplaced Pages:Database reports/Biographies of living persons containing unsourced statements. That list contains 500 out of about 17,000 similar pages. --MZMcBride (talk) 16:54, 20 December 2008 (UTC)