Wikipedia talk:Flagged revisions

Also see the discussion at Wikipedia talk:Flagged revisions/Sighted versions.

For an example of the development version of the FlaggedRevs extension on a public wiki, see this. AmiDaniel (talk) 08:48, 3 June 2007 (UTC)[reply]

If you have come here from a blog post, please know that there are posts out there which have badly misunderstood this feature. Far from being a closing up of Wikipedia, this is an opening up. In cases where we now protect or semi-protect (limiting who can edit), we hope to simply flag (meaning that anyone can edit). It is my hope that in the future we will be able to open the front page of Wikipedia to public editing for the first time in more than 5 years. This is a step towards being more wiki-like, not less wiki-like. Anything else you hear is FUD. --Jimbo Wales 21:22, 29 August 2007 (UTC)[reply]

A list of blog posts on the issue. Sdedeo (tips) 22:37, 9 September 2007 (UTC)[reply]

The previous discussion is here. Sorry about the mess, I really believe this will be helpful. --Merzul 19:57, 31 August 2007 (UTC)[reply]

Splitting it into two proposals, very good decision.--Father Goose 20:54, 31 August 2007 (UTC)[reply]

Unfortunately, I have to leave my office now, or I will have to stay all night, so I'll leave this in a rather uncompleted state. There seems to be some support for this general idea, so I guess you guys will have to clean all the inconsistencies, and so on. --Merzul 21:23, 31 August 2007 (UTC)[reply]

Note that all logged-in users, even new users, are shown the most recent version irrespective of what version is tagged.

Is it not possible to set in "preferences" that flagged versions are shown instead? This has been requested several times. PaddyLeahy 09:25, 1 September 2007 (UTC)[reply]

Hmm, there are already quite a few preferences, we try to avoid adding any. Voice-of-All 14:42, 1 September 2007 (UTC)[reply]

That does not seem like a sensible reason for denying a sensible request.--Father Goose 17:53, 1 September 2007 (UTC)[reply]

Although it is not a preference I would have enabled, I still think it would be a very good idea to include it. GDonato (talk) 20:29, 1 September 2007 (UTC)[reply]

It's an idea, but there are technical issues of confusion/replication of errors/bogus bug reports and the issue of adding more clutter to a cluttered up page. Adding preferences is always a last resort. It also discourages article development.Voice-of-All 02:19, 2 September 2007 (UTC)[reply]

I doubt you'll be able to avoid adding this preference sooner or later -- a huge number of users are going to beg for it.--Father Goose 02:34, 2 September 2007 (UTC)[reply]

Voice of All (and others), the reason why I suggested such a feature (and I think others have also suggested it, maybe for different reasons) may be very useful (I will leave the exact way of implementing to the developers, another way of doing it would be to use 'sighted.wikipedia.org/wikipedia/en/'), is that it could enable to see the trusted part of the wikipedia only (as in 'the reliable encyclopedia'). Pages which do not have a sighted version should be turned into redlinks at that moment. Users which have the right to sight pages could then use that to go through subjects quickly to see which pages do exist, but do not have a sighted version, and convert whole subject areas into reliable subjects (the reliability of a page also depends on the reliability of the pages a page links to.

Another advantage of such a feature would be, that it is possible to just sit back, and read a reliable encyclopedia, without having to worry about reading vandalised pages, or pages which are not up to standards. I personally would use this 'sighted wikipedia' as my reference wikipedia, while I would continue editing in the normal wikipedia. Hope this explains. --Dirk Beetstra ^{T C} 10:35, 2 September 2007 (UTC)[reply]

I was referring to the simpler idea that logged-in users could choose to see the same version as external readers (but taking advantage of other custom settings available only when you are logged in). PaddyLeahy 12:39, 2 September 2007 (UTC)[reply]

I don't know, it might be one of those things that seems obvious, but you end up not needing it at all. I don't think any more pages than now will be vandalized. If that is the same rate, why wouldn't users want to help revert vandalism? Do you really think there will be a lot of wikipedians that actively don't want to see problems with articles and help, even when those problems are infrequent? If you turned this proposed feature on you would constantly have to re-load the page to see the current draft, I think that would become annoying for even the users that thought they wanted it, and many would discontinue using it. Better to wait and see I think, I could be wrong :) - cohesion 01:38, 23 September 2007 (UTC)[reply]

It is not 'not want to see problems with articles and help', it is about using the encyclopedia for what it is supposed to be, an encyclopedia. If I need some data, I don't want to worry about the ugly part, I want the data. And then it would just be nice to have a possibility to browse the trusted part of the database. For the rest I will just continue to do what I am doing now .. getting the rest up to scratch. --Dirk Beetstra ^{T C} 02:06, 23 September 2007 (UTC)[reply]

Beetstra is right. Many of us tend to forget that the billions of hits on Wikipedia are nearly all from people who want to find some (hopefully reliable) information quickly - i.e., users, not editors. Walkerma 03:07, 23 September 2007 (UTC)[reply]

Right, but they aren't logged in. We're talking about people who are logged in. Why would someone logged in want to see an old version of a page? I'm sure this feature will eventually get added b/c people like preference creep and it's hard to stop, but I don't think many people would use it. - cohesion 04:06, 25 September 2007 (UTC)[reply]

Maybe you only log in when you want to specifically edit, but I know I am pretty much always logged in, and I can't imagine I'm the only one. And I've read there's apparently a LOT of accounts used simply because of preference setting. With that in mind, since a lot of people would rather not read vandalized pages, why wouldn't say want to see "an old version"? Am I missing something with your statement, or do you ignorantly assume logged in = wants to edit? ♫ Melodia Chaconne ♫ 11:31, 25 September 2007 (UTC)[reply]

There is now a preference to do this. Voice-of-All 22:05, 6 April 2008 (UTC)[reply]

Okay, so "all logged-in users, even new users, are shown the most recent version", but what do people who are not logged in see; the current version or the last tagged one. I don't like it when we pay so much attention to what logged-in people see, we also do it with date stamps. The vast majority of people reading Wikipedia will not be reading it from an account. --Arctic Gnome (talk • contribs) 19:46, 8 February 2008 (UTC)[reply]

"Users who are not logged in will initially see the most recent sighted version, if there is one. If no version is sighted, they see the most recent one, as happens now. Users looking at a sighted version can still choose to view the most recent version and edit it."

The above is from the first paragraph of Wikipedia:Flagged revisions/Sighted versions, which contains the details of the proposal. (The project page here was split into two subpages, that one and Quality versions. -R. S. Shaw (talk) 06:48, 9 February 2008 (UTC)[reply]

Which we must not do, until de.wikipedia finds out what the consequences of that are. I have a feeling it'd really really hurt. --Kim Bruning (talk) 21:41, 3 May 2008 (UTC)[reply]

I don't want to offend your feelings, but do you also have some arguments in that respect? At the moment, it's the vandalism that really really hurts. --B. Wolterding (talk) 21:59, 3 May 2008 (UTC)[reply]

Like I've explained each time this came up, it does likely change the dynamics of how the wiki works. People won't see the latest version, so they'll have less of a tendency to reach out and fix problems, possibly. It also adds a new responsibility that needs managing. All in all, a pretty big change. It's a shiny new toy, for sure, but in this case it might be a good idea to hold back a bit, say several months, and learn from the mistakes of others first (ie, de.wikipedia). --Kim Bruning (talk) 00:39, 4 May 2008 (UTC) I hope Jimmys vision works over at dewiki. We'll see if that's how it pans out.[reply]

So are you saying that de.wikipedia is the testbed for en.wikipedia? -- Imperator3733 (talk) 20:49, 12 May 2008 (UTC)[reply]

We can learn and have learned things from the other Wikipedias. They steal from us every day; we are a little less diligent about stealing from them. - Dan Dank55 (talk)(mistakes) 20:54, 12 May 2008 (UTC)[reply]

Now that the appropriate pages exist on meta, etc... perhaps this page is redundant? I think I will trim it down to just the essentials. --Merzul 12:54, 3 October 2007 (UTC)[reply]

At Wikipedia:Village pump (proposals)#Stable versions is coming; what standards, guidelines, and processes need to be written? Thanks. -- Quiddity (talk) 01:33, 1 April 2008 (UTC)[reply]

This is a phenomenally bad idea that will (possibly) spell the end of Wikipedia as we know it. This is the beginning of a members-only club, everything we swore to hate. This basically means anonymous users can't make live changes to the publicly-displayed pages. Registration is now required. You can not edit this page right now. And so forth. This is it. The end. Equazcion •✗/C • 06:42, 3 Apr 2008 (UTC)

Sighted versions, maybe. But "quality versions", while itself a mixed proposition, will be a way to offer readers a "guaranteed good" version of a page. Provided the selection process is not deeply flawed, this is definitely something we want.--Father Goose (talk) 06:54, 3 April 2008 (UTC)[reply]

I might not understand the proposal completely. I should've gone to the meta page before I opened my big mouth. But the fact that this will only happen to certain selected articles should probably be mentioned here too. Equazcion •✗/C • 06:58, 3 Apr 2008 (UTC)

The selection process is wholly unspecified as yet; for the time being, flagged versions is pretty much just a technical implementation. What articles end up having a "quality" version may follow something like the Good and Featured article processes, and the path of least resistance would be to simply have those processes designate a "quality" version in the course of a review.--Father Goose (talk) 08:25, 3 April 2008 (UTC)[reply]

There are a few similar comments in the Village Pump thread, such as "appears to negate the whole concept of a wiki" and "wait a year". Wikipedia is an evolutionary pool of ideas, and evolution requires two things to work: change and stability. Flagged Revisions is simply a tweak, and a very important one, to the processes that promote the stability necessary for Wikipedia articles to evolve. The Germans (I know) and the Russians (I hear) are going to adopt it sooner rather than later, and it might come to Wikibooks in the near future. Let's see how it works in those contexts before we kill it off, okay? - Dan Dank55 (talk) 16:13, 3 April 2008 (UTC)[reply]

I'm still on the fence here. Whether or not this works will depend largely on how it gets implemented. If we're not very careful then yeah this will be a bad thing. I think this is largely the result of people, especially the foundation/Jimbo, being sick of Wikipedia's reliability being the butt of a collective internet joke. Prioritizing the search for a solution to that is going to have some major downsides. "Openly-editable" and "reliable" are fundamentally two conflicting terms. But as it's already been set in motion, I suppose we'll have to wait and see. Equazcion •✗/C • 18:14, 3 Apr 2008 (UTC)

Every one of those statements is perfectly reasonable, unlike the daft things I have seen outside of Wikispace. - Dan Dank55 (talk) 18:47, 3 April 2008 (UTC)[reply]

Why is there a push for "quality versions" when "Sighted versions" is the solution? Zginder ^{(talk) (Contrib)} 00:30, 4 April 2008 (UTC)[reply]

Sighted versions isn't an actual validation process, just a sight-check for vandalism. Most of our articles at any given moment are vandalism-free, but can still contain outright falsehoods. Chasing out the falsehoods takes real legwork and fact-checking, not just a "not vandalism" rubber stamp.--Father Goose (talk) 09:16, 7 April 2008 (UTC)[reply]

Wow! The sky is falling! I have been suggesting for some time that we need "verified" versions; I had not thought that the software would actually make these the default for the "outside" to see, I was thinking that it was going to take kludges, such as simply tagging all unverified articles as such. This is better, then, than what I'd been suggesting.

It is crucial. For some reason which remains obscure to me, a great deal of effort has gone into scrubbing non-notable (but verifiable) articles from the encyclopedia, when these don't harm the reputation of Wikipedia at all. Who sees them? People who are looking for them Or maybe random browsers, but they can expect to find obscure stuff. What harms the reputation is articles on important topics that contain unverified information, or that are POV imbalanced. Now, verified versions (and that is an application of flagged revisions) will, indeed, create new problems. But these are standard editorial problems for an encyclopedia! I do know how we could pull it off, efficiently, but ... until people are willing to try to understand new proposals before rejecting them ... it's not going to happen. I speak of WP:PRX, of course. Which was not at all about voting, in spite of the constantly repeated refrain at the MfD, nor did it propose to create any bureaucracy (though it could create a spontaneous hierarchy that might replace the functions of a bureaucracy without the inefficiency if the community chose to do so). I'm making this comment here, but I'm in pull mode. Want to know about this? Ask me. And the discussion will probably take place, for the most part, off-wiki. Too much poison in the air here at this point.

It was said, "Openly-editable" and "reliable" are fundamentally two conflicting terms. No. Not at all. Thinking that way is essentially a failure of imagination. Consider the possibility before us. I won't get into the nit-picking distinction between being logged-in or not, and, really, all that might need to happen is that an IP user presses a button and, presto, "automatic session log-in as IP." (Warning: your IP address will be shown for edits you make. If you wish to avoid this, register an account and log in.) Piddling detail. So what we have is an encyclopedia that any one can edit, with a layer over it, seen when not "logged in," which consists only of a subset of articles. Presumably those which are (1) notable and (2) verified. Reliable. (The set of users who can tag as verified would be those who could handle the responsibility, and follow consensus.) This, I think, might make some deletionists extraordinarily unhappy, they will have to find something more useful to do, or at least equally as satisfying as crushing other's work, for with the verified layer, the arguments for actual deletion based on non-notability largely disappear. (Though there is still the work of reviewing notability tags.) Not "anyone" can verify or otherwise flag an article, this would be a privilege; it might indeed be automatically assigned, perhaps so many edits without problems, or perhaps by admins, but could be removed for failure to follow guidelines that the community would establish. Frankly, if a debate was over a notability tag rather than life or death for an article, I'd care much less about the result, and, as far as I'm concerned, notability standards could become stricter. It will be useful if the flags for versions could be not just "verified," but also more than one "notable" tag. The very top layer is verified *and* clearly notable, but below this there could be the complete encyclopedia, where the standard for notability is that a single user took the trouble to create an article and another to tag it as "noticed and therefore notable," having that right. And below that would be submitted articles. Verified tags would be independent, and verified articles could exist at all the notability layers. True deletion remains for hoaxes, copyvio, libel, etc. (But some hoaxes might better simply be tagged as hoaxes! I have one in mind, what was really a humorous essay -- but, unfortunately, placed in article space.) And nearly all true deletions would be routine, speedies.

Knowledge without hierarchy is far less useful than knowledge which is arranged in layers of increasing detail an decreasing overall importance, with the ability to drill down through the layers. That was the vision of hypertext that had us so excited in the 1980s. I considered the development of wikis very hopeful.--Abd (talk) 01:36, 4 April 2008 (UTC)[reply]

We already have a members-only club: the set of autoconfirmed users, who are able to mark pages as "patrolled". You may think of that as "the end" if you choose.

I think this is a great idea! The main thing I want to see is the very bottom-level flagging, used to flag pretty well all changes as "good" except vandalism.

On many pages, if you glance at the page history it's obvious that there's vandalism from time to time and between it, people keep reverting back to the same version. I'd like there to be a tab labelled "stabler version" such that if you click it, you get that page that keeps getting reverted to. I had proposed that this be done automatically: e.g. if three editors reverted back to the same identical version, it would be automatically marked as the stable version. (I had a more complex algorithm than that.) But here we can easily do essentially the same thing just by having ordinary users flag things as checked, in a similar way that we mark pages as patrolled.

I look forward to implementation! --Coppertwig (talk) 17:28, 4 April 2008 (UTC)[reply]

Look at it this way: It won't change anything, except that it will provide one more feature as a convenience. As it is now, anyone can study the page history (and talk page) and figure out that a certain version seems to be accepted by several editors and likely free of recent vandalism. This takes time. The new feature will simply allow people to find such a stabler/less-vandalised version at the click of a button without having to study the details of the page history themselves. If what you want to do is look up the atomic weight of an element (for some purpose that's not important enough to require true reliability, but you still want a high probability of getting the right value) this sort of thing could be highly convenient. The most recent version of a page could easily have the value arbitrarily changed; the "stabler version" is far less likely to, although it can also suffer from the same problem. What's wrong with letting people find that stabler version at the click of a button rather than having to spend time studying the page history? --Coppertwig (talk) 22:57, 6 April 2008 (UTC)[reply]

My main reservation about the "sighted versions" portion of the proposal is that new or unregistered users will no longer have direct interaction with the encyclopedia, but instead will post changes, then have to wait for someone to sight them. Not all the details are in place, so that may or may not be how it will actually work. But if it does end up functioning that way, it will make it so that people are no longer editing the encyclopedia, but instead "suggesting" edits which may or may not be approved. There's a definite psychological difference -- for better or worse, I can't say, but it might scare some editors off. It would probably also make "afterthought" or "fixup" edits impossible, where a person writes something, checks it after writing, notices something is wrong. Yes, you should preview before saving, but even I sometimes rewrite stuff after posting it even though I preview religiously.

An additional problem is that I wonder if people will sight good-faith, non-vandalism edits that are just not very good. I'd rather people be actively reverted than ignored, and given no explanation as to why their edits never appear in the encyclopedia. Depending on how it is implemented, Sighted versions may end up being a complete abandonment of how wiki editing works, where every edit must be approved by a bureaucracy first.

I'm totally for Quality versions, as a means of offering a "known good" version to readers, but Sighted versions flirts with "control freak" issues that I am very wary of.--Father Goose (talk) 02:15, 7 April 2008 (UTC)[reply]

It might be best to start out by having the stable version be the default page content only for trolled/semi-protected pages, like Evolution. Voice-of-All 03:28, 7 April 2008 (UTC)[reply]

That might not be bad.--Father Goose (talk) 05:45, 7 April 2008 (UTC)[reply]

If they're semi-protected then presumably they'd already be free of vandalism, generally. Or did you mean in place of the semi-protection... Equazcion •✗/C • 05:50, 7 Apr 2008 (UTC)

I believe in place of protection. I remember reading somewhere that Jimbo was promoting sighted versions as an alternative to semi-protection.--Father Goose (talk) 07:22, 7 April 2008 (UTC)[reply]

Yes, the idea is that we don't have to shut out people from editing pages by putting them under semi-protection; instead of semi-protecting we can set the default view to the last "sighted" version. Anons can still make changes, the edits just won't show up to unlogged-in viewers until somebody with a few edits under their belt marks the latest version "sighted". It's kind of like seconding a motion. -R. S. Shaw (talk) 07:38, 7 April 2008 (UTC)[reply]

That means we'd still have to deal with any IP vandalism. It just wouldn't show publicly in the interim. I dunno. Equazcion •✗/C • 07:42, 7 Apr 2008 (UTC)

I gather it wouldn't show for the vandals either. That would probably discourage them from continuing after an experiment or two. But while "sighting" might prove to be a good option in certain extreme cases, it may be hard to resist the temptation to eventually apply it to all articles. Just as there was no "study" performed to see what we lost with prohibiting new users from creating articles (despite one promised), I doubt there will be one with sighted versions, and we won't know if we lost good potential contributors because of it. I'd personally rather do 10 vandalism reverts than lose one good contribution, or worse still, scare away a good contributor. I can't say if sighted versions will end up being a good idea or not, but I wish we were more willing to concretely measure the drawbacks than look only at the benefits and dust off our hands.--Father Goose (talk) 09:09, 7 April 2008 (UTC)[reply]

I agree, we should do a pro/con analysis. I think it would be difficult right now since the implementation is still up in the air. Otherwise I'd start a page now that lists pros and cons of semi protection vs. sighted version. Equazcion •✗/C • 09:16, 7 Apr 2008 (UTC)

(unindent) More than a pro/con, I'd want to see an actual study of contribution patterns before and after implementation, to see if it produces a dropoff in new (good) users, for instance.--Father Goose (talk) 11:15, 7 April 2008 (UTC)[reply]

I'm not all that curious there. We're talking about protected articles that switched over to sighted, so # of new editors could only increase, I'd gather. Are you more likely to create an account if you weren't able to edit the article you wanted to at all? The point in keeping anons able to edit is to encourage them to join, so I'd imagine the more they're able to do, the more likely they are to join. Why would you assume the opposite? Equazcion •✗/C • 11:31, 7 Apr 2008 (UTC)

(ec)I agree about preferring to do 10 vandalism reverts than lose one good contributor. However, many of the terrible consequences discussed above would only occur if we took away from some users the ability to see the most recent version -- and I don't think anyone's proposing that.

The proposal is to supply more information, not less. People will still be free to read the most recent version, and to look through the page history and read any version they wish. But they will have additional information: that certain version(s) has/have been marked as being preferred or approved in some way.

Actually, there could be a problem if non-Wikipedians look at the approved or "sighted" version as the default: they will see less vandalism, and therefore be less likely to begin editing.

On the other hand, we might bring in even more new users, if it's made easy to compare the most recent version with the sighted version, perhaps via a tab. Then vandalism would be more obvious and more people might start editing.

Whether the most recent or the sighted version is presented as the default view should be based on what the reader prefers, in my opinion: what the average reader prefers for the default default, and with each user able to set the default in their preferences. Anything else would be overly controlling, IMO. --Coppertwig (talk) 11:43, 7 April 2008 (UTC)[reply]

(←) How can one find out what the average reader would prefer? The discussion archives are full of speculations about what would be more attractive for new contributors. Some say a person is happier if her changes show up immediately, others argue that a person is more pleased when his edits gets reviewed. It would be interesting to poll actual readers of Wikipedia rather than just reading speculations. Personally, I would be more impressed to see that my edits are being reviewed. I'm a Ph.D. student, so I'm used to having my fabulous papers reviewed and rejected; I consider that an integral part of doing science. Seeing a review process in place here would greatly increase my respect for this project. But like most people, I'm not average. :) --Vesal (talk) 14:57, 7 April 2008 (UTC)[reply]

Hi. As I noted at the VPump, this page (Wikipedia:Flagged revisions) would really benefit from an update, by someone who knows what is what, and the historical development of this whole concept. Specific things that might be useful include:

• a slightly more detailed summary of how it works (2 new UALs, 4 selectable ratings, (I think))
• perhaps with a link to the screenshots at http://www.mediawiki.org/wiki/Extension:FlaggedRevs#Use
• pointers to any current/ongoing discussions
• pointers to important historical discussions (mailing list threads, local/meta/de talkpage threads, etc). All I can find in the Signpost archives are 2006-07-10/More stable versions and 2006-08-07/Wikimania tech
• A very brief sense of how it developed; and how it differs from, or incorporates, meta:Article validation and Wikipedia:Stable versions
• anything else that might be helpful in bringing a newcomer to the idea, up to speed.

Thanks :) -- Quiddity (talk) 03:42, 7 April 2008 (UTC)[reply]

Seconded. Listing Port (talk) 17:27, 7 April 2008 (UTC)[reply]

Someone has made a start of this at Wikipedia:FlaggedRevs fact sheet. If anyone with knowledge about the unaddressed points could improve that page, we could link it from, or merge it to, this page. -- Quiddity (talk) 19:12, 30 April 2008 (UTC)[reply]

With this feature, is there the danger of a new kind of vandalism, where a malicious user marks a vandalized version as the "flagged version" shown to the outside world, but editors remain unaware because they see the recent version where the vandalism has been deleted? --Marcinjeske (talk) 02:44, 20 April 2008 (UTC)[reply]

Not everyone will have the power to tag articles as assessed. We aren't sure how liberal we will be with handing out this power, but you can bet that anyone caught knowingly marking vandalism as quality will get in trouble. --Arctic Gnome (talk • contribs) 06:47, 20 April 2008 (UTC)[reply]

It is very unlikely for Editors to vandalize. Only as likely as say, users that go rogue and "go out in a bang". Also, if it was deleted, there is a good chance that the deletion was auto-reviewed. Last, users still know if the stable version is different than the current, no matter what. Voice-of-All 19:59, 20 April 2008 (UTC)[reply]

Pardon me for disagreeing, but editors (that is to say, with accounts) vandalize all the time. If you want to be malicious, it is not that hard to get an account. Yes, usually these are short-lived accounts or sock puppets, but I would be concerned because of the two aspects: 1) that editors don't see the flagged version, but the current, so less chance for noticing it, and 2) the vandalism is indirect.. the editor would not be making the bad edit, they would simply be referring to the bad edit already made. I think limiting the handing out of this power as above may address my concern... but I still fear that for controversial articles (where this feature would be most helpful), it may also be the most open for abuse... imagine edit-warring over the flagged version by point-pushers. But I guess time will tell - it probably is worth it to try.--Marcinjeske (talk) 23:00, 3 May 2008 (UTC)[reply]

In this case "editor" is not the same thing as a user with an account. "Editor" is a user right (similar to the current rollback right). Mr.Z-man 23:05, 3 May 2008 (UTC)[reply]

I get that the implementation is in the form of a new rights group to join "rollbackers, administrators, bureaucrats, bots, checkusers, and oversights". I hope there is no suggestion of naming that right "editor" and leading to endless confusion? My original concern is mostly alleviated if the right is treated like WP:oversight is now. The big issue with vandalism (both obvious vandalism "wiki sucks" and sophisticated vandalism of changing meaning - "The company's founder was a well-known communist.") is the time it takes to correct it. Edits (in decreasing order of reaction time) are patrolled by bots and people at Special:RecentChanges, on logged in user watchlists, and by people noticing something while reading an article and checking the page history. I hope that these existing mechanisms can be used to monitor change to flagged revisions. Otherwise, the people in the best position and with the knowledge to fix it (the editors) won't notice because they see the current version.

For an example of subtle kind, until my series of edits this week, the Monarch article had a section on Monarchs in Africa that was solely composed of a description of a modern pan-African Imperial Empress... with no third-party source and no mention of well-documented historical monarchies. Much more than the "this is stoopid" form, wildly misleading information is what gives Wikipedia the bad rep. --Marcinjeske (talk) 03:57, 4 May 2008 (UTC)[reply]

Editors under this system will routinely look at diffs between current and sighted version. Either the bum info is in the current version too, in which case we are no worse than at present, or it has been removed from the current version by a clueful person without capital-E Editor rights, in which case such a significant edit should attract the attention of an Editor with the page on their watch list, and be rapidly (within days if not hours) sighted. The most popular idea is hand out "Editor" rights very widely, so it would be the most common status for registered users with a significant editing history. That way, it is possible for Editors to sight a large fraction of wikipedia and keep it under active observation on their watchlists. (Think of this as a replacement for Recent Changes patrol, which would require much less effort under sighted versions). While this would mean that occasional "sighting vandalism" would surely happen, it would also allow it to be quickly reverted. The trick is to make the qualification onerous enough to deter most vandals and slow down the rest of them, while light enough to hand out to a large fraction of active editors. There has been a lot of productive debate (see archives) about how to do this, and I think all suggested solutions are available in the current implementation. PaddyLeahy (talk) 14:39, 4 May 2008 (UTC)[reply]

Flagged revisions (both sighted and quality) went live on the German Wikipedia on 6th May. This gives an updated summary of the current sighting statistics.

As of 12th May, 14.5% of the 750,000 pages had been sighted, and just over 2000 users had sighting rights, compared to 298 admins and 550,000 user accounts. 0n English Wikipedia, about 90% of user accounts have zero edits; if the same applies on DE:WP it looks like the rights have gone to a substantial fraction of the actively editing community, as hoped for EN:WP sighted versions as well.

Two users have each sighted over 4000 pages each. That's not so good IMHO: unless they are very active, that is too many for one person to monitor for useful updates (of course these probably are hyper-active users!). On average it looks like each sighter will have a few hundred pages to watch; not too awful since many will be very obscure.

Comparing the numbers for English Wikipedia, we have 3 times as many articles but 14 times as many user accounts, and 5 times as many admins. Presumably our hard-to-count "active user community" is bigger than the German one by something between 5 and 14 then, so we probably have an easier job of keeping EN:WP sighted and up-to-date than they do over there. PaddyLeahy (talk) 20:34, 12 May 2008 (UTC)[reply]

Very interesting, thanks for giving the overview. I guess it makes sense to wait and see how it works out over there, before we try it out here; but I hope we won't wait for too long ... Merzul (talk) 20:42, 12 May 2008 (UTC)[reply]

It would be interesting to know what their criteria for sighting articles is. IIRC, de.wiki has much stricter content policies than we do, I wonder if that is reflected in their flagged revisions policy as well. Mr.Z-man 20:50, 12 May 2008 (UTC)[reply]

The criteria that will be chosen for this is still beeing discussed. Right now you have to request the rights on a special page. One of the ideas for who gets the rights is: 30 edits during 30 days.FreddyE (talk) 08:29, 13 May 2008 (UTC)[reply]

I´m wondering if it wouldn´t be most efficient to sight per category as part of an "inital sighting process"...for example..a group of sighters could form a "sighting group" for category A, and sight all unsighted articles of that category, then move to the next? FreddyE (talk) 08:33, 13 May 2008 (UTC)[reply]

You can do this, if you like. You can order Special:UnreviewedPages by category. --Avatar (talk) 14:39, 14 May 2008 (UTC)[reply]

Unlike "guarding" articles on your watchlist, sighters do not "adopt" pages. So it's perfectly OK to do an initial sighting on 20,000 articles and not monitoring them further in the future. If they are changed by a person withough sighting rights (we plan do have at least ~10,000 sighters in de), the unsighted versions are listed on the special page Special:OldReviewedPages. Every sighter (and mainly the disburdened vandal fighter) can check this page and sight the new versions, which is very easy because you only have to take a look at the diff. --Avatar (talk) 14:39, 14 May 2008 (UTC)[reply]

Thanks for the update, Avatar; anything you want to share would be appreciated. (Avatar is an admin on de.wp, guys.) I just found out on the wikiquality-l list today that a good page to monitor progress is http://tools.wikimedia.de/~aka/cgi-bin/reviewcnt.cgi (in German). Currently it says (my translation):

Out of 748726 articles currently on de.WP, at least 122890 currently have a "sighted" version (16.41%). At this rate, every article will have been sighted at least once by June 27. In addition, 122291 articles are currently sighted in their newest version (16.33%). During the last hour, 798 new "sighted version" symbols were placed [and it also gives daily and monthly totals]. 2188 users have the right to mark articles as "sighted".

- Dan Dank55 (talk)(mistakes) 15:24, 14 May 2008 (UTC)[reply]

Hi Dank55, you can click on the US flag above the text and than you don't have to translate yourself :-) --Avatar (talk) 04:39, 15 May 2008 (UTC)[reply]

Yes, thank you, this is interesting, but many people here would like to know how this influences the quantity and quality of anon contributions to sighted pages. Is anything like that measured? Merzul (talk) 16:44, 14 May 2008 (UTC)[reply]

P.Birken ist searching for good metrics in the quality-l mailinglist. --Avatar (talk) 04:39, 15 May 2008 (UTC)[reply]

Hi Tim. Yes, but there were some words whose meaning was not obvious to English speakers (reviewed, set, marks, module), so I translated it myself. I steal, um, borrow many things from the German Wikipedia :) - Dan Dank55 (talk)(mistakes) 13:48, 15 May 2008 (UTC)[reply]

The results that you mention from de.WP sound good and useful. One question: since what I'm reading above indicates that people don't want to see too many editors able to flag/sight revisions, and since I don't see any specific criteria listed for who will be able to flag/sight revisions: what about admins? Any admin has been approved by a significant majority of the community who voted and commented in the case, and a bureaucrat has agreed that the editor should have admin tools; it seems reasonable that an admin should automatically have flag/sight tools as well. Nyttend (talk) 13:59, 15 May 2008 (UTC)[reply]

Agreed, although when considering giving admins significant new authority, it would be best to be clear what kinds of misuse would result in losing admin rights. There has always been resistance on en.wp to anything perceived as "bureaucracy", and if only admins are allowed to mark articles as "examined" (a better translation of "gesichtet"), I would be very likely to support an RFA for someone who would do a good job of this, even if they don't know how to use every admin tool. - Dan Dank55 (talk)(mistakes) 14:22, 15 May 2008 (UTC) It looks like no one is considering this, so I don't need to worry about it. - Dan Dank55 (talk)(mistakes) 19:25, 16 May 2008 (UTC)[reply]

According to the sighted page, We already have a proposed guideline on the requirements. They are:

The right will also be given automatically by the software to editors who meet the following conditions:

   * Has an account for 30 days
   * Has 150 edits
         o 30 edits to article namespace pages
         o 10 article namespace pages edited
         o 15 days of edits
   * Has confirmed an e-mail account

I for one think that spreading it out widely among editors is a good idea. --Falcorian ^(talk) 17:00, 15 May 2008 (UTC)[reply]

I think that we should be careful to spread out flag/sight tools, perhaps requiring a few months and several hundred edits. I do not think that it should be restricted only to admins; my point was that I think that all admins should have it. Nyttend (talk) 01:40, 16 May 2008 (UTC)[reply]

I agree -- all admins get these tools by default. For everyone else, have requirements similar to what's needed for AWB (maybe 400-500 mainspace edits), including making people apply for access. That way, it would really only be people who want to use the tools that will get access to them. -- Imperator3733 (talk) 04:05, 16 May 2008 (UTC)[reply]

The criteria at WP:Flagged revisions/Sighted versions emerged as the consensus after a very long discussion now archived at WT:Flagged revisions/Sighted versions. The consensus was that sighting rights really are not a big deal (because improper use has very mild effects and is trivially reversed), and hence should be much more widely spread than adminship. Most new admins have 1000-2000 mainspace edits so requiring a good fraction of this for a mere "sighter" is overkill. Sighting will not work unless there are enough sighters to promptly re-sight pages where useful edits have been made. If the aim is that most pages are eventually sighted, as planned for DE:WP and as many of us hope will be the case for English WP, then you need the sort of numbers Avatar is talking about above, ie several thousand for DE:WP, a few times more for here. PaddyLeahy (talk) 19:02, 16 May 2008 (UTC)[reply]

I think the 150 edits as on the page at the moment seems to be a reasonable balance between having enough people to sight versions and reducing the abuse potential. I do not see the need for the requirement to 'have confirmed an e-mail account'. Having e-mail does not seem related to the ability to flag versions - I can see the case for admins but not for just sighting versions. Davewild (talk) 19:18, 16 May 2008 (UTC)[reply]

Yeah, now that I think about it, something around 150 edits would be best. Would sighting rights be able to be revoked, or would abuse just result in a block? -- Imperator3733 (talk) 23:25, 16 May 2008 (UTC)[reply]

I seem to remember it being said that it's simply a new class of user, so that would suggest that it would be possible to "demote" people if necessary. --Falcorian ^(talk) 03:49, 17 May 2008 (UTC)[reply]

Please note that the German Wikipedia does a test run since May 6th. And they handle the opposite of "Note that all logged-in users, even new users, are shown the most recent version irrespective of what version is tagged." --Subfader (talk) 12:41, 13 May 2008 (UTC)[reply]

There may be many differences between the way German Wikipedia uses Flagged Revisions and the way English Wikipedia uses it. For English Wikipedia, the provisional plan at this point is to show all logged-in users the most recent version.--ragesoss (talk) 00:23, 17 May 2008 (UTC)[reply]

This is correct. In contrary German Wikipedia shows the latest sighted version. This is a long debate. Most users (of the German Wikipedia) don't see a point in having sighted versions which aren't shown as default. So the question is if the sighting of new edits scales, so no new version is unsighted for more then x minutes/hours/days. If it won't scale, most users will see the stable version approach as failed. But at the moment, we're still optimistic. --Avatar (talk) 13:15, 17 May 2008 (UTC)[reply]