Wikipedia talk:Flagged revisions/Sighted versions/Archive 3

From Wikipedia, the free encyclopedia

Archive This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
← Archive 2 |
Archive 3
| Archive 4 →


Contents

Security is an illusion. The danger is believing that illusion

This entire moves seems like a total PR-push, and will have very little effect on vandalism. All it will do is

  1. discourage new users from trying out wikipedia and thus reduce the number of potentially good editors we will get.
  2. Establish an unenforcable beurocracy, and thus create the illusion of a secure system while merely providing cover for vandals to do their work.
  3. It will not really reduce destructive vandalism, it will only cause vandals to get usernames and make 300 random edits before they are allowed to present official versions of their vandalism.
  4. encourage further wheel wars between editors as POV-pushing gets "sanctioned" by one editor who can "sight" an edit which, while technically not vandalism, is just POV crap.
  5. Creates the illusion of "we support this article as true" I KNOW THAT IS NOT WHAT THIS PROCESS IS SUPPOSED TO DO! However, that has no bearing on the general perception. It will still appear to the public that this will be an process whereby we are certifying the veracity of what an article says; which is the wrong thing to do.

Just my two cents. --Jayron32|talk|contribs 17:37, 23 August 2007 (UTC)

Re #3 -- if someone can make 300 random edits without vandalism, they're quite possibly improving the encyclopedia enough that it'd be worth dealing with a piece of "sighted" vandalism or two before that user is blocked. -- Myke Cuthbert (talk) 17:48, 23 August 2007 (UTC)
I agree with Jayron32 points, specially with #5. This process will create the illusion that some article versions are officially sanctioned instead of "incomplete"/"under construction" which is the real status of 90+% of Wikipedia articles. It'll also give more arguments for critics of this project as this "officially sanctioned" articles will get torn because of inaccuracies, POV, bad writing and other problems. Even the most experienced Wikipedia editors are not completely free of this common defects. --Victor12 17:57, 23 August 2007 (UTC)
What's wrong with saying "this article has been verified"? Wikipedia will never become a reliable source unless we establish a fact-checking system to verify articles. I'd like to produce a quality encyclopedia that is reliable and therefore useful. CloudNine 18:28, 23 August 2007 (UTC)
This assumes that established editors know more than IPs, thus they can better decide which versions are better. This is not the case as most people editing wiki are not experts. So while this will work for obvious vandalism it will not work for more difficult issues such as POV, the established also incur in POV, all the time! This will make it easier to impose their views instead of discussing them. --Victor12 19:24, 23 August 2007 (UTC)
If they make 300 good edits and then get blocked for review vandalism, that's much better for Wikipedia than three vandalistic edits and a block. CloudNine 18:24, 23 August 2007 (UTC)
What if they make 300 POV-pushing edits, none of which are flagged as vandalism, and then continue to POV-push once they are auto-approved? It's not simply vandalism that is a problem, but revert wars when savvy POV-pushers are involved. There are plenty of editors here that make hundreds or even thousands of edits to a single page (which they have staked out as their own), very few of which go unreverted by other editors, and it's not unheard of for someone in an edit war to have no formal warnings issued to them - any automated process of granting "surveyor" status to editors will have no way of telling if their 300 edits resulted in ZERO unreverted changes, and simply give greater power to the dedicated POV-pushers. Conversely, if all it takes is one warning to cost an editor their surveyor status, then all a POV-pusher has to do is post a warning on the page of any editor who has reverted one of their edits, and eliminate the competition. Again, vandalism is something I think we can handle - the deeper, more profound threat is the type of editor that deliberately manipulates WP to promote their POV agenda, and I expect that such editors will do whatever it takes to attain the highest "status" possible as reviewers/surveyors, in order to further their ends. If the new planned system is not designed to prevent such abuse, it may only make things worse. Dyanega 20:30, 23 August 2007 (UTC)
The illusion of safety is dangerous. 9th Edition of Encyclopedia Brittanica was a wonderful book, but with hindsight it was riddled with POVs. I hear that some editor supports this ludicrous theory that the world is not flat- and Arts are subjective. Now the articles on my watchlist are mainly unvandalised- though 3 get hit twice a day, this would stop that. A bigger problem is bots and administrators making ill-informed (well meant) blocks and changes- made in the belief that their interpretation of a 'fact' is the only acceptable interpretation of that 'fact'. Adding another layer of good intention rings 'alarm bells'. ClemRutter 19:07, 23 August 2007 (UTC)
I started out neutral and stayed that way reading much of this page. But Jayron convinces me this proposal is a bad idea. Especially persuasive are the discouraging effect this could have on new, possibly unregistered editors and the creation of new layers of authority and review. I don't think what we'd gain in "quality control" is worth what we'd lose in recruitment and retention of new editor. And I don't think adding new layers of authority over what is "correct" will help. It could well exacerbate tensions that already exist in the community. David in DC 18:58, 23 August 2007 (UTC)

The public already believes we support any article we put up as true. --AnonEMouse (squeak) 19:12, 23 August 2007 (UTC)

I agree with anonemouse. the public (such as my spouse) takes wikipedia to have a basic bit of truth to it, as with any encyclopedia. The benefit of actually checking the sneaking vanalism and pov before sending it out to the general public I think outweighs the potential loss of new editors. Maybe for articles flagged as controvertial, a third party surveyor should be employed? --Rocksanddirt 19:36, 23 August 2007 (UTC)
If this is done right, there shouldn't be much loss of new editors. Pyrospirit (talk · contribs) 19:41, 23 August 2007 (UTC)
I actually think there will be no net loss of editors. Some POV pushing SPA's might leave over it, and those who dislike working collaboratively, but that would be it.--Rocksanddirt 20:36, 23 August 2007 (UTC)
Agree completely with anonemouse -- my colleagues constantly laugh at errors in Wikipedia and think that WP's paid staff of writers are idiots. They don't realize it's a volunteer effort, and that there is no "Wikipedia corporate office" which is writing and approving articles. People who know that neither Wikimedia nor we writers endorse every thing on WP will also know that "sighted" versions are also not officially endorsed. This extension mainly makes it so that people who are under the main misconception are bitten by vandalized, libelous, or obviously incorrect articles less often. -- Myke Cuthbert (talk) 21:06, 23 August 2007 (UTC)

Fundamentally violates what wikipedia is.

I feel that this concept and more so its method of implementation fundamentally violates what wikipedia is.

Sighted revisions effectly deny the community the ability to edit an article when they see an error. If we hide errors/vandalism from viewers fewer will make changes to the wiki. But this is our main line of recruitment. Few of us came to be editors because we woke up one morning and thought "I know, i'll write an article for wikipedia today"... most of us saw mistakes and then changed them on the spot, this aroused our curiosity and before we knew it we'd been here over a year and had thousands of edits :) by disrupting this system we are isolating our current editors. This upsets me as I honestly think its one of the worst things we could do.

My second major gripe is with the 'surveyor', who seems to basically be a new class of editor, who has the power to decide what is and isn't right. But there are two problems with this:

  • 1.) Is being right (or rather not being wrong) all there is to making a good article?
  • 2.) Is an olympic sports article surveyor qualified to survey medical and biochemistry articles?... obviously not.

I'll explain the first because it may not be obvious, but not being wrong (i.e. the facts in teh article are correct) isn't all that make an article. Many of our FAs are good, but not perfect. A recent one talked at length about a russian princess, but for example didn't explain how she died (she was murdered). So although i'm sure all the facts were correct, they omitted a major part of that article. To endorse the article with what is effectively a mark of quality though seems counter productive as it discourages editors from making positive changes especially if they are quite sure, but not 100% sure if they are right.

In conclusion this is a nice idea, and well done for spending the time to come up with a serious and theoretically implementable proposal... but fundamentally this violates what a wiki is all about, its long term effects could be very bad and at the end of the day I can't see it achieving any of the goals set out. We will forever be under threat from libel, we will always be accused of inaccuracy by the Britianic/Encarta etc. old guard and media. It doesnt even really tell a user how good the article they are seeing is, its very qualitative. And there are so many articles, I cannot bear to think how long it will take to survey them, especially if we consider there are only 1300 admins and it sounds like this surveyor class needs to authorised by an admin. Who is going to do that work? admins are already stretched thin! If I have misunderstood any aspects of this please feel free to correct me, but I can say right now that fundamentally I cannot agree with this, it just violates what wikipedia is about. WikipedianProlific(Talk) 17:52, 23 August 2007 (UTC)

Don't forget, this proposal mentions two new classes of editor: surveyors, who can mark revisions as "sighted", and reviewers, who can mark revisions as "quality." Surveyors would consist of pretty much all experienced, helpful editors (300+ edits and had an account for 60+ days); thus, limited marking revisions as "sighted" to these editors would be similar to limiting moving pages and editing semi-protected pages to editors with accounts at least 4 days old. The more controversial class of editor is the reviewer, who can make a revision marked as "quality" and therefore appear as the default for non-logged in editors. Reviewer would indeed be about admin level. Pyrospirit (talk · contribs) 18:21, 23 August 2007 (UTC)
As the proposal says, admins can grant it out to anyone whether they have 300 edits and such yet, like on a Request Page, so it would include even more people. Voice-of-All 18:24, 23 August 2007 (UTC)
Indeed. Almost all major contributors would be surveyors. Pyrospirit (talk · contribs) 18:37, 23 August 2007 (UTC)
Yes I realise this, its part of the problem "Almost all major contributors would be surveyors" = who is going to regulate that? the 1300 administrators? who are already stretched thin on other topics? Not to mention this second higher class of quality surveyor --> what qualifies say a biographies quality surveyor to judge a science article? Its a terrible idea, at the moment we're up front about the fact that the contents of a page may be totally wrong, but this (even though its not meant to) effectively makes us look like we're endorsing the quality of the article in some way. Its just not workable. If you consider the average number of edits per article is only around 15.7 at the moment! We'd be better off waiting 5 years before bringing in something like this to consolidate material. WikipedianProlific(Talk) 18:46, 23 August 2007 (UTC)
I disagree. Look again at the requirements for an article to be sighted:
  • Is clear of vandalism.
  • Is clear of libel and unsourced statements about living persons.
  • Is clear of unencyclopaedic content.
  • Has been around for several days.
  • Has some depth (not a one line stub).
  • Has been checked for basic accuracy.
  • Has been spell checked.
  • Is readable (not cluttered-up pages with no wikilinks) and not tagged for cleanup.
I'm an editor of car articles, computer articles and general science articles. I'm sure I could look at a biography article and decide if it's up to the expected standard against those most basic 'sighted' criteria. Sure - I couldn't give it the GA or FA stamp of approval - but anyone can check for vandalism and that kind of obvious stuff. The hardest part would be 'checked for basic accuracy' - but I would imagine a quick 'google' would suffice to establish that the article isn't complete crap. SteveBaker 18:59, 23 August 2007 (UTC)
The thought of someone who edits car articles reviewing say an article on a medical topic actually makes me shiver with fear, if I said in an article that the cytochrome P450 oxidises a ubiquitin molecule tagging associated components for deletion by a peroxisome could you tell me if I was wrong? (its extremely wrong). This is a method of consolidating quality effectively, but the encylopedia isn't good enough to do that, at least not yet. It might cut vandalism but then its exchanging it for several deep layers of beauracracy, it won't reduce libel (Infact i think it will increase legal problems as people will see it as an endorsement of quality, even if that isn't what its meant to be) at least now we're up front that the pedia' isn't nessessarily 100% accurate. I think the analysis of what a Good Article and featured article is has been severely misunderstood in the proposal. FA doesn't by any means make an article perfect, it just means its quite good, enough so to go on the front page.WikipedianProlific(Talk) 21:18, 23 August 2007 (UTC)
Edits to the stable version (in other words, current=stable) by surveyors can be autosighted to maintain the stable version. So all of these edits (often spelling corrections/tweaks/tags/categories) made by anyone who has been here for a good while would be maintained without extra effort. Voice-of-All 20:19, 23 August 2007 (UTC)

With regards to the number of surveyors, there shouldn't be a problem there. Because becoming a surveyor would require a certain level of experience, surveyors would be very unlikely to decide to vandalize. Anyway, if they did abuse the sighting revisions power, an admin could remove their surveyor flag. Pyrospirit (talk · contribs) 18:50, 23 August 2007 (UTC)

But vandalism isn't really a problem? Its not out of control. Almost all vandalism is removed within minutes of being left. This just seems like a stealth policy to sneak in a set of proposals which fundamentally close doors to new editors. Recruitment will suffer from this. And yeah an admin could remove the flag - but they have plenty of work already, this just creates more without a real need - it will create so much beuracracy. No matter what angle this gets thrown at me from I still don't think its a good idea, it really changes accessibility of the pedia. WikipedianProlific(Talk) 18:58, 23 August 2007 (UTC)
If you've ever tried to call up WP in the middle of a classroom lecture and found "BEETHOVEN SUCKS HORSE PENIS!!!!!" instead of the section you wanted to discuss, then the fact that the vandalism will be removed within a few minutes is pretty small consolation. Vandalism is a major problem to WP's perception. When people come to expect that there's a chance of vandalism on any page, when comics make WP vandalism part of their daily routine, and when news anchors vandalize WP on the air, it all adds up to a major perception problem for WP. -- Myke Cuthbert (talk) 21:11, 23 August 2007 (UTC)
I think recruitment will improve because of this. See Wikipedia_talk:Flagged_revisions#Why_this_will_encourage.2C_not_discourage.2C_new_editors below.--ragesoss 21:16, 23 August 2007 (UTC)

Tell me if I got this right...

Okay I'm not quite sure if I understand everything, so I'll give a try at summarizing this...

If this was to be accepted, we would basically be reviewing a version of an article. If the article is "presentable" under the different criteria stated in the page, it will gain the status of "sighted revision", so we can see when it was last seen without vandalism, etc.. Same thing for "quality version", except the criteria are for FA and frequently visited articles.

I'm not sure if I understand this fully, can someone please tell me (in a quick and simple way) what I didn't get right, and what I didn't understand at all?

Thanks in advance!

Zouavman Le Zouave 17:53, 23 August 2007 (UTC)

That's pretty much correct. One thing I'd like to mention is that quality revisions are not only presentable, but the single best revision of a featured article. It may also apply to Good Articles and A-rated articles, depending on how this works out. Pyrospirit (talk · contribs) 18:14, 23 August 2007 (UTC)

Meaning of "sighted"

I'm a little confused by what is meant by "sighted". Is this simply a misspelling of "cited", or is there some other meaning? If "cited" is the word that was meant, I'm still a little unclear on how it's being used here.

Its sighted in the sense that it was last viewed in a presentable condition (whatever thats supposed to be). Or at least thats what I thought, it would have been nice if they'd used slighly simpler to the point terms, like I don't know "Last unvandalised revision" maybe.WikipedianProlific(Talk) 18:00, 23 August 2007 (UTC)
It used to be called "non-vandalism", but the authors of the software disagreed with this term. Given localisation, we can use whatever term we like.
James F. (talk) 18:05, 23 August 2007 (UTC)
It's a rather confusing term, and I think we should replace it with a term such as inspected. That's a lot easier to understand, and while it doesn't imply thorough review like quality does, it still shows that it's at least not vandalized and fairly well-written. Pyrospirit (talk · contribs) 18:07, 23 August 2007 (UTC)
I'd second this, or at least would if I supported the proposal. "Sighted" seems unclear to me. Plus it could easily be confused by the inexperienced with the very different "oversighted" though that name never made sense to me, I propose we change that to "baleeted" :-). It's not like we don't have enough confusing terms (e.g. MediaWiki/Wikimedia/Wikipedia) the wub "?!" 20:33, 23 August 2007 (UTC)
I thought the point of "sighted" was exactly that it is unclear what it means. This pushes people to look up the definition instead of jumping to conclusions. NB: random viewers will not encounter the term: All the will see is a wikipedia page with a box at the top saying "Click here to view or edit the latest version" or something similar. Only if they get involved enough to edit will they notice that they had been looking at something called a "sighted version". 82.6.79.225 21:18, 23 August 2007 (UTC)
Sorry. I searched for "etymology" and overlooked this section here. Instead starting Etymology (Derivation) of sighted?. I don't think making people look things up is a reason to use a well-known word non-naturally. Make up a more completely new word if that's what you want to do.--SportWagon 22:33, 23 August 2007 (UTC)

The first sighted version is going to be the problem

Personally, I rarely read Wikipedia articles top-to-bottom, except perhaps when under some kind of review (e.g. FA; GA). Obviously, when a revision is marked as "sighted," there's some level of trust that we're going to have to put in one another that we really can trust the cited version. So If I see an article history that goes something like:

  • (cur) (last) 12:06, 23 August 2007 User A (Talk | contribs) (7,992 bytes) (fixing spelling) (undo)
  • (cur) (last) 12:05, 23 August 2007 Surveyor Z (Talk | contribs) (7,990 bytes) (sighted) (undo)
  • (cur) (last) 12:04, 23 August 2007 User B (Talk | contribs) (7,998 bytes) (Capitalization) (undo)
  • 500 older edits in the history

etc, it would seem perfectly logical to compare the diff of 12:05 to the diff of 12:06, see the user fixed a spelling mistake, and mark the new version as cited. What I worry about is that there's going to be very few people who will want to read an entire article, including external links, every time they want to make a revision. This means that the first person to mark a revision as "sighted" is going to be very important, because most likely, most subsequent "sighted" tags are going to be applied mainly by checking diffs. "No problem," you say, "we'll just organize patrols of all sighted revisions. Yeah right. There are 2,411,528 articles on the English Wikipedia! It will be completely unfeasible to read through bajillions of kilobytes of text on any reliable basis. The sheer manpower this would require may not be insurmountable, but it would detract enormously from all other areas of improvement. "That's not much different from the present system" you say. But the key is that under this new system, there's going to be an illusion of increased accuracy, and will bring with it an obligation on our editors to stand up for that accuracy, but doing so will mean sacrificing any growth in our encyclopedia. --YbborTalk 18:17, 23 August 2007 (UTC)

Theres a thought, it could create a whole new type of vandalism - marking an article as sighted when in fact it hasn't been read through. At least at the moment we're quite up front about the fact some articles may be greivously wrong - but to put our quality seal on it just leaves us wide open to critisism in my opinion. WikipedianProlific(Talk) 18:20, 23 August 2007 (UTC)
Not with a decent tag and disclaimer. People often don't even know they can edit, and assume it's like Brittanica. Voice-of-All 18:26, 23 August 2007 (UTC)
You think putting a tag up on top of every page saying "someone said this article was OK, but it could very well be a piece of crap" would be an improvement? That's already the case, the only difference is now we don't need a banner at the top of the page to tell people that. --YbborTalk 18:29, 23 August 2007 (UTC)
I doubt there will be much vandalism by sighting bad revisions. Remember, editors would have to have the account for 60+ days and make at least 300 mainspace edits to get the right to mark revisions as sighted. Editors with that many edits almost never vandalize. Anyway, just granting surveyors the ability to unflag a revision as sighted as well as flagging it will solve the problem. As for reviewers, it'd be an admin-level position, if not higher, so vandalism from them would be about as common as vandalism from admins. Pyrospirit (talk · contribs) 18:32, 23 August 2007 (UTC)
Oh no, I'm not considered about people surveying as a form of vandlaism, I'm conerned about users who survey in good fiath, but just happen to miss something. I think things would stick around much longer if we thought we could "trust" a certain version any more than we can trust a certain version today. --YbborTalk 18:36, 23 August 2007 (UTC)
Don't forget, if someone saw the misspelling in the article and fixed it, it would still show up in the diff even if they weren't a surveyor. I don't think this would make misspellings last any longer because of checking only the diff; on articles I frequently edit, I usually just check the diff between my last edit and the current revision anyway. Misspellings would still be eliminated through the standard wiki process. Pyrospirit (talk · contribs) 18:54, 23 August 2007 (UTC)
I think you misunderstand my point. I think this system will work fine for misspellings. My concern is bolstered by your very reply, "I usually just check the diff between my last edit and the current revision anyway." And that's exactly what's going to happen with sighted versions. People are going to compare the current version to the last sighted version, the problem is the 500 edits before the last sighted version. If something slips through the first surveyor, the odds that it's going to stay are very high, even higher than a comparable change would be today. --YbborTalk 19:00, 23 August 2007 (UTC)
We can't expect it to be a perfect system, but if the large majority of sighted versions are checked all the way through the first time only, that's still a big step up from nothing. Some things will slip through, but on the whole we'll have a framework in place for sifting wheat from chaff, a system that can be built upon later for more fine-grained sifting and redundancy.--ragesoss 19:05, 23 August 2007 (UTC)
Are you confident "the large majority of sighted versions [will be] checked all the way through the first time"? Because I'm not. --YbborTalk 19:07, 23 August 2007 (UTC)
Majority? Yes, as long as versions aren't automatically checked as reviewed when edited by someone with sighting rights. We'll obviously want a separate page that details the procedure of reviewing (all the things that need to be checked for), but I'm confident that half-plus-one will be conscientious reviewers. Whether it would happen 51% of the time, 75%, 90%, or 98%, I'm not sure. I expect that a good reviewing interface with hard-to-ignore instructions would help push compliance to the high end.--ragesoss 19:19, 23 August 2007 (UTC)
What I'm trying to say is that frequent editors of a page already often consider their last edit a "sighted" revision in a sense, so not much might be changed there. The problem is currently solved by new editors coming in and noticing the problem; they wouldn't simply check the diff from the last sighted revision if they notice a problem. Pyrospirit (talk · contribs) 19:17, 23 August 2007 (UTC)
Fortunately, that mechanism (correction by anons) still works in this case, since the only errors they'll see are ones that slipped through the sighting process.--ragesoss 19:21, 23 August 2007 (UTC)

Quality...is dubious. Sighted is OK.

I have a couple of FA's that I've been the majority editor on...so I have some knowledge of what this means.

I like the 'sighted' proposal and would certainly support it.

The 'quality' proposal is much more difficult. It assumes that a 'featured' article is 100% finished. But what happens if the world changes? If I have an article about (say) a kind of car - today, it's 100% finished so I push it (at enormous effort - trust me on this!) to FA status and it's flagged as 'quality' and now appears by default when someone seeks information about that car. What happens if a new model of that car comes out? I add half a dozen paragraphs and a new photo - but those will always be hidden by the 'quality' version. In order to get a simple addition to this article (and an addition that a lot of people will be coming here to read about) - I now have to get the article re-stamped through the FA process? Geez - that's an awful lot of work for everyone. The number of re-FAC's from articles that just need a tiny change will be significant. Will we have to go through all of the bloody annoying nit-picking all over again? You just know that things that passed the first time are going to get complained about the second (third, fourth) time...it's in the nature of the process.

Lots of times, changes in an FA are not even due to the world changing. Frequently people add new translations of the article into new languages (this happens a lot over the months immediately after passing FAC because non-English authors like to translate FA's into their native languages because it's more productive to translate a great article than a mediocre one). What happens when the Category system changes (eg My Mini article went from being an 'English automobile' to being a 'British automobile' then to being a 'Defunct British automobile'). Suppose a template or an image or a linked article is renamed? Even FA's with absolutely perfect, up to date content get updated quite frequently because of events that are out of their control. We can't have them gradually rotting because the infrastructure around them is changing - and we can't have editors having to go back to the Reviewer team for every stupid little cat: change.

And meanwhile - for however long this process could take - the article may well be flat out WRONG as far as casual readers are concerned! We'd be in a situation where these 'quality' articles are in much worse shape than typical un-flagged articles - and the much better version can't ever be easily improved!

I believe that if this process is going to work, we should nominate one or two keepers for each 'quality' article. Keepers being (perhaps) the people who nominated the article and pushed it through FAC in the first place - who would have the same rights as Reviewers - but only for those specific articles that they "keep". This would allow (in practice) the original majority author of a truly great article to have the authority to re-stamp it as 'quality' with whatever changes are needed without a great deal of red tape and associated delay. The assignment of 'keeper' status would be done by the FAC process - so as a part of FAC acceptance, they would look at the edit history and pick one or two people who could clearly be trusted to 'keep' the article maintained.

If the 'keepers' proposal is not possible - then we would need a DRAMATICALLY slimmed down re-FAC/re-quality-flagging process - something that could be repeated (if necessarily) as often as once a week for a month or more - and which would under no circumstance take more than 24 hours to approve the new version of the article. The sheer administrative overhead of this seems horrible to me...but it's better than nothing.

Failing that - I strongly oppose the idea of the 'quality' version of the article being the one casual visitors see when they come to Wikipedia. At most, I'd agree with having a button somewhere on FA's banner that says something like "Click here to see a quality reviewed version of this article - which may be a little out of date" ...or something like that. It would be a waste of time because nobody would click on it - but what the heck - it's better than the crappy old, outdated, redlink/missing-image-ridden 'quality' version showing up all the time.

SteveBaker 18:19, 23 August 2007 (UTC)

If it was changing so rapidly, you would unsight the quality version and review it later when it was stable. The problem is Wikipedia will never been taken as a reliable source if this is not implemented eventually. CloudNine 18:22, 23 August 2007 (UTC)
The majority of changes after passing FAC are:
  • Addition of more translations of the article (requiring no change to the substance of the article - but adding a template).
  • Refinement of categories - somewhat due to there being no easy way to find all of the categories an article should belong to - somewhat due to the continual 'churn' in the architecture of categories (again - no substantive change to the parts of the article that matter - but requiring of an edit).
  • The advertisement of newly minted FA's on the front page results in MANY more eyes on the article. The improvements in my article after being on the front page were vastly greater than the ones leading up to and during the FAC process. Having half a million people read your article really improves it!
Having to unsight the quality label in order to make even the most trivial change means that no article will retain the quality label long enough for it to be useful to anyone - just about every FA would be 'unsighted' within a week for some really good reason - and it's unlikely anyone would want to go through that entire horribly painful FAC process again - and if they did, it would bring the FAC review process to its knees! If the proposal is to be able to easily remove quality status but to NOT make it just as easy to reinstate it - then it's useless.
SteveBaker 18:40, 23 August 2007 (UTC)
This is indeed a major issue and one pretty much deferred until later so far (note that the proposal suggests implementation only for OTRS issues). I pretty much agree with SteveBaker about the way to handle it. The simple way to implement this is for "keepers" to have full reviewer status, with the understanding that it shall only be used on that article. Use on other articles would be grounds for having the status removed. Discussion of whether the nominated keeper is appropriate would, unfortunately, be a necessary part of the review to make this work, but it should be uncontroversial in many cases...authors of feature articles are unlikely vandals. WP:OWN issues are likely and users who feel there are problems with specific keepers should have a place to complain (e.g. bureaucrat's noticeboard) if they can't work things out on the talk page. The keeper(s) should be mentioned at the top of the talk page so people could bug them to review a new revision. (There might be problems with harassment here...better idea, anyone?). If a keeper goes on wikibreak (or quits in disgust) he or she should notify the bureaucrat noticeboard, & maybe nominate a successor. Should the keeper lose his/her status if the article fails a FARC? Guess not, they are trusted, but they should obviously refrain from using it until the article regains status. Just suggestions for comment. PaddyLeahy 00:23, 24 August 2007 (UTC)


apply the flags only to certain articles?

In order to prevent vandalism, it would be enough to flag just articles which are prone to be vandalised. (For example, I only work on mathematics articles, which are hardly every vandalised. Such a flagging procedure appears rather hindering than helping in such cases). Somewhere it was mentioned that 6% of the edits are vandalism. How many pages are vandalised on a regular basis?

Another suggestion: vandalism can be handled better if enough people watch the articles. Is it possible to display something like: This page is currently watched by 493 users, vandalism will be in approx. 31 hours (or 5 minutes or whatever, depending on the page). If there are too little people watching a certain page, a request could be generated to editors who are belonging to an appropriate Wikiproject.

Jakob.scholbach 18:20, 23 August 2007 (UTC)

The sighted flag will be a useful tool not just against vandalism, but also for improving newer articles. I think the sighted flag should be used on all pages. However, I'm fairly sure the quality flag will only be used on very high-quality articles such as FA, GA, and A-rated articles. Pyrospirit (talk · contribs) 18:24, 23 August 2007 (UTC)
Most GAs are abominable, although I am told the system has been improving. FA judges pretty pictures, minutiae of article presentation, and the presence - not the quality - of footnotes. Let us not pretend that we have a functioning evaluation system, especially in order to make such articles harder to fix. Septentrionalis PMAnderson 18:30, 23 August 2007 (UTC)
This might have the unfortunate side effect of moving vandalism towards those articles, however. Ben Hocking (talk|contribs) 18:26, 23 August 2007 (UTC)
Could you give an example of one of these "abominable" GA's? And from the FA's I've read, they're all excellent articles. Pyrospirit (talk · contribs) 18:40, 23 August 2007 (UTC)
Was this question directed at my comment? I wasn't talking about GA's or FA's. I was talking about my conjecture that if you create a tool that helps prevent vandalism and only apply it to those articles that are vandalized the most, then the vandalism will move to those articles currently (mostly) vandalism-free. Ben Hocking (talk|contribs) 20:04, 23 August 2007 (UTC)
No, it was to me; for FA, try Daniel Webster, sourced, as it still is, out of Profiles in Courage and Henry Cabot Lodge's century-old treatise on the magic of eloquence. Septentrionalis PMAnderson 21:32, 23 August 2007 (UTC)

(unindent) Is it known how many constructive edits are made by editors which are not logged in? I guess we will lose these ones to a large extent. Jakob.scholbach 21:58, 23 August 2007 (UTC)

In the articles I watch, it's about 3 good edits for every 2 vandalisms for those not logged in. As this is totally from the gut, there could very well be a selective perception factor involved, either way. Ben Hocking (talk|contribs) 22:22, 23 August 2007 (UTC)

Another question to Pyrospirit: how does the sighted flag improve new articles (at the page of the proposal, no positive effect except linked to vandals and "the public perception of WP" is mentioned)? I can only see the following disadvantage: anonymous editors will be pushed back from editing because they don't see the most current version where an error etc. might already have been fixed. This may decrease their voluntary help. Jakob.scholbach 22:05, 23 August 2007 (UTC)

Brief proposal regarding this talk page

This talk page is getting rather crowded and long; anyone think we should divide it into 2 parts? One for discussion of the surveyor position and sighted flag, and the other for the reviewer position and quality flag. This might make the whole discussion a little easier to sort through. Pyrospirit (talk · contribs) 18:26, 23 August 2007 (UTC)

Perhaps this should be split into two separate proposed policy pages? CloudNine 18:30, 23 August 2007 (UTC)
I was thinking more along the lines of a couple level-1 headers on this page for better organization. Pyrospirit (talk · contribs) 18:35, 23 August 2007 (UTC)

This solves the wrong problem: need black-, gray-, and white-listings

Flagging good versions is not the correct path. Flagging the bad versions is the correct choice. Just like how black-listing email spammers is much less effective than white-listing people you trust.

How?

Every user starts out on the grey-list and every edit has to be vetted by someone on the white-list. As a user attains accepted edits then they are white-listed. If a user vandalizes then they are penalized severely and put on the grey-list and then eventually on the black-list. So, instead of assuming a user is a good user you assume what is most natural: "unsure" == grey-listed. It's heavily based in merit. You must earn "respect" as an editor and you lose it when you abuse it.

A point system would be easy. (Exact point values are semi-arbitrarily chosen to demonstrate my point.) Start at zero. If an edit is vetted then you get a point. If you get marked as vandalism then you get penalized 5 points. Hit 500 points and you are white-listed and can now vet others' edits. Hit -50 points and you're black-listed. An admin can immediately black list you (aka block) if necessary per normal procedures. If you are found to have vetted a vandalistic edit then you are immediately grey-listed as your assessment is questioned. A white-listed user abusing this latter point (say, for tactical reasons in an argument) is punished much more severly.

This puts responsibility on the vetter and the vetted with checks and balances on both. No group can beat the system because you are penalized for participating negatively. My main problem with the proposed version flagging system is that it groups vandals and non-vandals together into "unsighted versions". As a surveyor, you have to "sight" every non-sighted version which means you are flagging good edits, not the bad ones.

For a similar concept, see the karma system for slashdot.

Ultimately, I think there is too high volume of edits to effectively, and explicitly, flag good versions as proposed. A more automated way (i.e., points), IMO, will be the only effective way to keep the volume manageable now and in the future as volume continues to increase. Cburnett 19:08, 23 August 2007 (UTC)

Ugh, no. This would be horribly bureaucratic, would create a massive cabal of old-time editors, and would completely discourage newcomers. Wikipedia is supposed to be an open, welcoming community, not a bunch of distrustful, suspicious encyclopedists who don't like anyone new. Pyrospirit (talk · contribs) 19:32, 23 August 2007 (UTC)
Agreed. Karma tends toward becoming an elitist system, and can be gamed too easily. Sxeptomaniac 19:43, 23 August 2007 (UTC)
Yeah and granting surveyorship based on edits and time here can't be gamed nor elitist? Gimme a break. If you want to convey privileges based on anything then you open yourself up to elitism and gaming. Name a viable system that doesn't allow either of these. And if it's your interpretation, Pyrospirit, to take meritocracy as closed and anti-new-editor then that's your choice. If you want to assign credit to users who demonstrate it then you by definition are separating the trusted from the possibly-trusted from the untrusted. If being trusted goes to your head then you get elitism. You pose your argument against my proposal like it's a unique problem. Being a surveyor can entirely cause elitism and be gamed and yield POV pushing. As I said, it's inherent to any system you put in place and, as I've also said, I think it's more effective to mark bad edits than good edits. Cburnett 20:05, 23 August 2007 (UTC)
Also, points given per edit? Consider this scenario: A new user just registered their account and hears about the "karma" system. They are frustrated with the slowness of the system, so they then make a bunch of very minor edits to boost their edit count up to 500. They then become "whitelisted" without any of the experience needed to be knowledgeable about Wikipedia and use their "whitelisting" powers to POV-push by only accepting edit they like, causing all sorts of chaos and disruption. Pyrospirit (talk · contribs) 19:49, 23 August 2007 (UTC)
You assume a minor edit and a non-minor edit are the same. Throw in a time requirement if that makes you happy about being impatient. Throw in an option for a vetter to assign points (5 through 1). If it's a strong contribution then give them five. If it's average then give them three. If it's a typo correction then give them one. Raise the white-listing bar to strike a good balance.
Not accepting an edit means someone else can accept it. Marking an edit as vandalism when it is not is "punishable" by someone else. I'm not sure how you can get away with POV pushing since it would require each and every white-listed editor to agree to POV push otherwise the pushers get their just rewards by someone who isn't pushing. Cburnett 20:05, 23 August 2007 (UTC)
I think you can fix most of these problems with a karma system by taking out the descrete levels. Suppose we rate all editors on a smooth scale from zero to infinity - we give our most prolific/accepted editor (Jimbo maybe) an initial score of 10 - and the most evil vandal gets a zero. When you join Wikipedia, you get a 1.0 - when you make an edit and someone disapproves of it, your score gets reduced by (let's say) 1/100th the score of the person who disapproves - and if it is accepted and 'praised' by someone, your score goes up by 1/1000th of their score. This doesn't create an 'us and them' because there is no specific number above which you are "in the cabal". What's more, the good guys with higher scores can use their expertise to rapidly promote people who are obviously OK - but the only moderately experienced wikipedians have less power - and so on down. I don't think I want such a system - but it can be done. You can use the same scheme to rate articles - by averaging together the scores of the people who made each edit. Those that have lots of edits by experienced editors will get pushed to the top of the 'reliability' pile - and old article that's been vandalised into oblivion will get pushed towards the bottom of the reliability pile. Reverting vandalism (presumed to be edits by low-scoring individuals) will restore the article's status. SteveBaker 23:34, 23 August 2007 (UTC)

terrible idea

If I understand correctly, this flagged revision system will mean that if I edit a page, that edit will not be immediately shown to all users who visit (unless I am blessed with some kind of special status from an admin?) This, to me, breaks wikipedia. I understand the desire to fight vandalism, improve Wikipedia's image, etc., but this is the wrong solution. Anything that does not allow "edit this page" to function as expected must be avoided. Sdedeo (tips) 19:12, 23 August 2007 (UTC)

Why? This will allow users previously blocked by semi-protection the chance to edit any article, while denying vandals the visibility they crave. To me it brings us closer to the "anyone can edit" philisophy. (I don't see where it says "and have it appear immediately" in the slogan :)). CloudNine 19:15, 23 August 2007 (UTC)
What's the relevance of "anyone can edit" if it does not appear immediately? --Victor12 19:20, 23 August 2007 (UTC)
If the purpose were to only apply flagging to semi-protected pages, that would be fantastic. However, flagging is to be applied to *every* page. -- Ken g6 19:23, 23 August 2007 (UTC)
This discussion is getting massive, so I'm not sure if anyone's still reading, but I am happy to see (reading other comments) that the essential breakage here is considered harmful by many. Personally, it seems cooked up by some bureaucrats (not the technical definition) and is just a variant of "we will review contributions in a reasonable amount of time, and incorporate them if we like." If you want that, go to citizendum. This is not to dismiss the worries that prompted the invention, but we need to think harder, and smarter. Sdedeo (tips) 19:30, 23 August 2007 (UTC)
The "special status" is intended to be handed out pretty liberally, perhaps even automatically. You're a user that has been around a while and has hundreds of edits and has never been blocked , you would almost certainly get the status. As far as I can tell, any editor that isn't either brand new or troublesome will be given the permission. --Tango 20:06, 23 August 2007 (UTC)
That was the promise for adminship. Did it happen? Septentrionalis PMAnderson 20:13, 23 August 2007 (UTC)
No, because delete/protect can cause seriously issues (as a dev I will get rid of nasty history merges using delete that it takes a sysadmin to revert in MediaWiki 1.12). Revealing deleted information can cause complaints (oversights helps out there a little now). Editing the main page is no joke. It has too many serious abuse potentials, where is the "surveyor" right doesn't. So the "it will be just like adminship" idea is a non-sequitur. And if you though only admins could review, you clearly didn't read the proposal. Voice-of-All 20:23, 23 August 2007 (UTC)
No, but that was promised for ideological reasons without taking into account the practical nature of the site once it grew beyond a certain point. This proposal will only work if it is given out very freely. If only a few people have it, they won't be able to keep up with all the edits. Restricting adminship works, we don't have any major admin backlogs at the moment (to my knowledge). Restricting surveying wouldn't, which is why it won't happen. --Tango 20:29, 23 August 2007 (UTC)

I am not interested in applying for special status from an admin, and I think I'm in good company. At best -- if the status is granted automatically by the software as in semi-protection -- it would amount to semi-protecting the entire wiki. Sdedeo (tips) 20:26, 23 August 2007 (UTC)

It wouldn't be semi-protecting the entire wiki, since it would allow anon and new users to continue to edit pages, their edits would just have to be vetted by someone more experienced before they became visible to the general public. --Tango 20:29, 23 August 2007 (UTC)
No, because anons still edit the current revision, which is how new edits are made by anyone, which then get reviewed. Semi-pro just does not allow edits. Voice-of-All 20:30, 23 August 2007 (UTC)
Fair enough. Call it hemi-protection, then; it still breaks "edit this page." Unless you want to change the label to "suggest an edit to this page" for those users who don't make the cut. Question: is the status really going to be granted automatically by the software? I was under the impression that this was not the case. Sdedeo (tips) 20:34, 23 August 2007 (UTC)

So its essentially a protection with an option to continue editing out of sight?

This is a much needed feature. Protection hinders development of the article and does not guarantee a stable and reliable material. In cases of content disputes it is really helpful to have an outsider come and read the article before the material becomes available.--Alexia Death the Grey 19:14, 23 August 2007 (UTC)

It's not exactly like protection. More like making a single version the most visible until a better version can be developed. Pyrospirit (talk · contribs) 19:20, 23 August 2007 (UTC)

My two cents

I agree with WikipedianProlific on this one. This seems like an unnecessarily bureaucratic addition that would create a new special class of editor. This is in direct opposition to efforts at preventing hierarchies. What absolutely essential need for this is there? It seems just like a cool new toy that we don't direly require to continue to function and improve. Complete protection is not used that commonly or for so long as to create a large problem. There are very good reasons for halting all edits to an article, and a loophole to this seems improper. VanTucky (talk) 19:14, 23 August 2007 (UTC)

Agreed. I support some form of "reviewed, stable versions", but only via a pure wiki process, like what FA uses. Creating a privileged class of users with special rights is simply unnecessary. It's a technical and exclusionary solution where a social and inclusionary one should be used.
FA involves discussion. We can't have a discussion about whether or not each edit is vandalism. For this to work, it needs to be unilateral (with the ability to undo mistakes, of course). If we allow everyone to make that unilateral action, then it would be pointless, since vandals would just sight their own edits. Therefore, it needs to be a restricted ability. The current proposal is to have it handed out fairly liberally. --Tango 20:01, 23 August 2007 (UTC)
A "review process", yes. A "Reviewer" user class, no.--Father Goose 19:33, 23 August 2007 (UTC)

Point-by-point criticism

Oppose is my vote for now. I am amazed at the depth of the Requirements for Sighted pages on the proposal page. Here, I would like to make a point-by-point criticism of those requirements. They are roughly arranged in order from best to worst.

  • Is clear of vandalism.

Fine. But I'd rather see this done with bots, for backlog reasons stated by many. When an anonymous user posts, the post gets flagged for bot review. The user sees his post, but with a warning at the top that it has been flagged and isn't yet visible to all users.

If the bot finds something odd, *then* the post is flagged for human review. "odd" is to be defined (and should be standardized), but is more lax than normal anti-vandal bot rules: more things should be flagged, and a few valid edits being flagged isn't a problem. For instance, external links to any source not "approved" could be flagged. However, blocks like *.gov and *.mil should be "approved".

Anonymous users viewing the page should see a tab labeled "flagged versions" or something, where they can see their contributions.

  • Has been around for several days.

I don't like this. "Wiki-wiki" means quick. This is not quick at all. The bot review process shouldn't take more than an hour or two IMHO.

  • Has been spell checked.

I would rather present questionable spelling instances, in the new edit *only*, to the user originally submitting the edit, rather than waiting for a reviewer. I wouldn't even mind having Wikipedia show them one at a time, via Javascript or something:

Wikipedia thinks you may have made a spelling error here:
...This was teh last book he ever wrote....
Would you like to replace "teh" with "the"? Yes No Always Never
  • Is readable (not cluttered-up pages with no wikilinks) and not tagged for cleanup.

"not cluttered-up pages with no wikilinks" is good. I wonder if a bot could be made to auto-wikilink such things? But it seems to me that the user's contributions should be accepted, and the bot's tagged for human review.

"not tagged for cleanup" is a mess. Do you realize how many pages are already tagged for cleanup, and not being cleaned up? There is no way you'll get enough editors to fix up everything that need "cleanup".

  • Has some depth (not a one line stub).

I'm not sure I like this. If I'm curious about something, I'd rather see that Wikipedia at least knows about it. And, of course, articles converted to one-line stubs would be reverted by an anti-vandal bot.

This is really hard. If I insert "ex-" in front of "wife" in an article, and the reviewer knows nothing about that person, how is the reviewer to check whether or not that person got divorced? But conversely, if that person did get divorced, and "ex-" wasn't added, that might be just as bad!

  • Has been checked for basic accuracy.

Similar deal. Basically, it seems to me that to do this with any efficiency, we would need another encyclopedia to check the encyclopedia!

This is even harder. I have a hard time even understanding what "unencyclopaedic" means.

All in all, I do not support the proposal as currently stated. I would not support it unless the work for reviewers were much reduced.

-- Ken g6 19:20, 23 August 2007 (UTC)

Agree with most of the criticism, but think it is still worth a try. --AnonEMouse (squeak) 19:33, 23 August 2007 (UTC)
However, with concerns about the reviewers catching more subtle mistakes and vandalism, I would point out that, from what I understand, reviewer status would be handed out to anyone who's been around very long (without getting into serious trouble), so many editors who work on the page will also be reviewing it. They will most likely be familiar enough with the subject to catch most mistakes (particularly the WP:BLP concerns). I think that the workload would not really increase, as most editors are doing that sort of reviewing of edits to articles they watch now, in order to revert vandalism or copyedit. Sxeptomaniac 19:57, 23 August 2007 (UTC)
Bots only catch blanking, "PENIS", "SHIT FUCK" and crap like that. Thats only a portion of the vandalism. Voice-of-All 19:59, 23 August 2007 (UTC)
That's because bots currently have to be very careful not to revert a valid edit. Flagging a valid edit is sometimes acceptable. Thus the bots could use other methods, such as flagging any unknown external link. They might also be able to use spam filtering methods like bayesian filters. -- 71.214.53.137 21:19, 23 August 2007 (UTC)
(That was me, forgot to log in) -- Ken g6 21:21, 23 August 2007 (UTC)

And the spell check reminds me: this is a formula for Anglo-American disputes, and the unhelpful form of nationalism which insists that WP must use the local spelling, no matter how rare or incomprehensible in English. Septentrionalis PMAnderson 20:09, 23 August 2007 (UTC)

"Has been around for several days" refers to the article, not the latest edit being sighted. Several of these are just disincentives to apply sighting to very minor articles: remember, pages without sighting will behave exactly as now, and that will be the large majority of our 1.9M articles, at least at first. I agree that the "basic accuracy" point is a problem: it makes sighting a quality stamp willy-nilly, instead of just a check that the article has not been embarassingly vandalised. For that reason I'd like to see this removed from the proposal. Similarly, the BLP bullet should be heavily caveated, although editors should obviously be more than usually careful about articles where BLP is relevant. PaddyLeahy 21:35, 23 August 2007 (UTC)

Can "quality" or "sight" articles simply be offered instead of imposed?

I think the biggest problem I have with this proposal is the fact that articles (given the over 1.9m) could get stuck in some limbo of not getting frequent enough sights. However I also like the idea that a previous FA/GA version of an article might be available for easy perusal. Is there a way to simply add an option at the top of pages that indicates something to the effect of "_(#)_ version(s) of this page have been designated as Feature Articles in terms of quality (etc)... To see that (those) versions of the article, click here." We could integrate parts of this current proposal to allow experiences/designated users to update those constant FA/GA links to include useful edits (including small sp/grammar/etc edits). It just seems like it would allow the best versions to remain on the public face (which is what this proposal is all about) while also inviting users new and old to participate in the current version as it evolves (which I think is being neglected in the current proposal). --Bobak 19:23, 23 August 2007 (UTC)

This is a wonderful idea, and one that I fully support. I think it would make the proposal far more workable than it is now. Wrldwzrd89talk 19:25, 23 August 2007 (UTC)
Yes, this looks like an excellent idea. If this was put in place, it would resolve all my concerns about quality revisions. Pyrospirit (talk · contribs) 19:27, 23 August 2007 (UTC)
Perhaps simply make the default be the current version, but have a tab for the last sighted version, as opposed to the reverse of that currently suggested? Pyrospirit (talk · contribs) 19:43, 23 August 2007 (UTC)
That (having the current version be the default with a "stable version" tab) was proposed here earlier, with the consensus against it at the time. I too like the idea of a previous FA/GA/A version of an article being linked at the top of the page, but only the version that went through the formal review process (i.e. I don't support the bit about experienced/designated users updating the constant FA/GA links to include useful edits (including small sp/grammar/etc edits)). ~ Danelo 19:59, 23 August 2007 (UTC)
I'm surprised it didn't get consensus as it was pitched there, I really think (B) in those choices assumes that people are vandalizing everything on Wikipedia. It's actually quite the opposite, 90% of the 1.9m articles probably aren't ever touched --which is actually why it's problematic to have to wait for the process and unimaginably large backlog that will likely appear if this gets implemented as is. I'd love to see more articles go through a QA process, but I don't like having potentially problematic "sight" or "quality" picks that could be up for arguments between educated people (esp. when you delve into hot topics related to nationalism, religions, etc). I wish I had known about that earlier discussion. --Bobak 22:16, 23 August 2007 (UTC)

I think people have the wrong idea, perhaps

Maybe I've read a lot of the opposition wrong, and the general idea wrong, but it seems a lot of people aren't reading what would happen correctly. Let's see if I got this right:

  • A select number of articles, mostly high profile, FAs, GAs, and otherwise commonly vandalised articles would get switched to this new mode -- the MAJORITY of articles won't be affected, at least not initially.
  • Said articles will have two states, the 'current' one which includes all edits, and the 'safe' one which does not include any unflagged edits
  • Anyone who has permission (whatever it takes to get that) may flag an edit that they did not make as safe
  • As people now revert vandalism ANYWAY, the time spent doing it this way instead of the normal way actually won't change TOO much. Rather than reverting, people would flag yes or no. I assume for the sake of the GDFL as well as WP:AGF, all old edits will still be visible and in fact even later flagable as safe.
  • It will be very clear to anyone reading the page, at least from the top, that this system is there. And again, it's in a small number of articles, reletively (though obviously, the most visited on the flip side).

Am I missing something? In my mind, this is a very VERY good idea on a way to reduce vandalism and, as mentioned elsewhere, keep quality articles up to quality. I've seen some pages that just keep going in CIRCLES. I can easily see this helping the circle turn into an upward line. ♫ Melodia Chaconne ♫ 19:45, 23 August 2007 (UTC)

The problem is that "sighting" revisions, a power that may be given to large numbers of editors, has the same effect of making it the visible revision by default to anons. That is what is causing all these concerns, though I personally think it'd do more good than harm with some modification. Pyrospirit (talk · contribs) 19:52, 23 August 2007 (UTC)
I don't see anything in the proposal saying that this would only be turned on for certain articles (beyond initial testing, anyway). Where did you get that? --Tango 19:57, 23 August 2007 (UTC)
one of the suggested ideas (way above) was to just let the new program grow orgainically rather than blank the whole site and have to initially survey everything. A page only goes onto the system once it gets "surveyed" the first time. My thinking is that it will spread pretty fast in areas that have active editors. --Rocksanddirt 20:10, 23 August 2007 (UTC)
I don't think anyone's suggested not showing anything to non-logged in users unless there is a sighted version. If there is no sighted version, they'll see the latest one, obviously. That's quite different to the system only being used on particular articles. --Tango 20:22, 23 August 2007 (UTC)

Huh?

Could someone explain what this debate is about to me in simple terms? Smartyshoe 19:52, 23 August 2007 (UTC)

On certain high-profile Wikipedia articles, a new software application would be created that would make edits made by new editors have to be reviewed by editors who have been given a special "reviewing" privilege before those changes appear on that article. If you read through this discussion, you'll learn more detail about it. Badagnani 19:57, 23 August 2007 (UTC)
Or to summarize, this debate is about introduction of a new level of metabeaurocracy. A right question to ask is who will control the controllers?(Igny 20:00, 23 August 2007 (UTC))
interesting observation. as the "surveyor/reviewer" is going to be (at least proposed to be) much broader ability than admin's. it will be the regular user/dispute resolution system that over see's this. administrator's noticeboard for article review/surveyors.... --Rocksanddirt 20:05, 23 August 2007 (UTC)
If some editors are given a special reviewing privilege, then can someone or some effort be made to explain why this proposal is not similar to the case covered in Wikipedia:Perennial proposals#Hierarchical structures, and therefore gets swiftly disposed of?  DDStretch  (talk) 21:14, 23 August 2007 (UTC)

Fact checking

There is a big difference between spotting vandalism and checking the facts of an article. "Sighted" should just mean "clear of obvious vandalism". There can be another flag "fact checked" if people want to be able to mark articles as fact checked. Even if sources are cited, actually getting hold of that source and checking it can be a lot of work. It's worth doing that work before featuring an article, but having to do it before every edit is visible to non-logged in users would slow the system down too much. --Tango 19:56, 23 August 2007 (UTC)

Why this will encourage, not discourage, new editors

Many discussants opposing this proposal assume that having sighted versions that appear by default to anonymous readers and editors will turn off new editrs, since their edits will not appear immediately. There will no doubt be some new editors who feel this way, but there are also many would-be editors who don't edit because the next person's edits automatically overwrite theirs, without any sort of intermediate determination of who was adding valuable material and who was adding nonsense. The affirmation of having one's (anonymous of new-user) edits validated by another editor is also a powerful psychological motivation to continue contributing; as it is, many new users go for a long time without even knowing whether anyone has seen their contributions.

Because only article with at least one sighted version will be affected by this and un-developed and under-developed articles will be the most likely to lack sighted versions, most situations where substantial worthwhile content is added by new anonymous users will be unaffected. Their edits will still show up, because there is no sighted version of what they are editing. In developed articles, the signal-to-noise ratio is much lower, and new users should appreciate Wikipedia's editing processes more if they see that some sort of vetting goes on with changes to major articles.

It's hard to measure the potential editors who don't participate because of the lack of something like this, but I strongly suspect that they are more numerous (and more capable) than those who would only edit without such measures in place.--ragesoss 20:02, 23 August 2007 (UTC)

Part of the reason polling now is very idiotic is that many major points of this proposal are still stated as conditionals. Will we start off with the stable being the default? Probably not (per Eloquence), we can switch later. At least a poll should be on these points first, then on the whole thing, rather than the total "yes/no" for the whole proposal, which is not even fully worked out. Voice-of-All 20:28, 23 August 2007 (UTC)
You're preaching to the choir. (Although I am in favor having the stable versions as default.)--ragesoss 20:41, 23 August 2007 (UTC)
Indeed -- I'm shocked that this went to a site-wide poll without getting things hammered out with a test wiki, etc. I'm happy to have site-wide interest in the proposal, but a flat out yes/no on all aspects of this is way premature. -- Myke Cuthbert (talk) 21:25, 23 August 2007 (UTC)

UNsighting

I have one minor suggestion, primarily to reduce sight-warring: once a Surveyor "unsights" a revision, that revision should be permanently unsighted. If Surveyors are to be trusted (and this whole process depends on it) that means there is some valid concern to the revision when a Surveyor unsights it. Thus, any sight-war is restricted either to (1) edit-warring between Surveyors as new revisions are created, and Wikipedia already has measures to discourage that; or (2) moving the sightable revision farther back in history -- which I believe is a positive solution, as Surveyors may come to an agreement in temporarily presenting an older, smaller, and less controversial version of an article. As a side note, I am not taking a position as to whether Admins would be able to reverse an unsight. Jpers36 20:31, 23 August 2007 (UTC)

Hmm, valid concern. How about if once a revision is unsighted, the editor who originally sighted it can't re-sight it. Surely you can understand the need to unsight a controversial edit, take it to the talk page, discuss it there, and then after a consensus emerges, re-sight it? That would provide some safeguard against sight-warring, while not emptying the options box. --YbborTalk 22:00, 23 August 2007 (UTC)
How about requiring an admin to re-sight it? Then it turns sight-wars into wheel-wars, and we already know how to handle them. --Tango 22:30, 23 August 2007 (UTC)

Creating baselines for wikiprojects

Having only discovered this discussion today (thanks to the topnote on my watchlist) I will keep my comments short. I really like the sighted idea. I watch labour related articles, and would be very interested in being able to work through them to set a sighted status. I would see this as a baseline effort which could then be maintained though an ongoing process. I would additionally be interested in being able to maintain a list (presumably at the WP:UNION page) of labour related articles and their sightedness… something like the assessment system, with varying levels of decay. i.e.

  • articles with 10+ unsighted revisions
  • articles with 500+ unsighted character revisions

or some such criteria. I think this would help distribute the workload, even for articles that are not watched by everyone interested at the particular project.

As for the editing/version/change is bad/elitism/other difficulties - I'll leave that argument to others, but the idea seems worth using, and would undoubtedly get polished along the way.--Bookandcoffee 20:35, 23 August 2007 (UTC)

My opinion

I think that this is a good idea as long as: 1) Surveyor rights are given away very liberally to registered users. 2) Reviewer rights are given away somewhat liberally to registered users in a process similar to RfA. Actually, I think that Reviewer rights should possibly be simply another one of the tools granted to administrators. That way, users who become admins through the RfA process will also be able to mark an article as quality. --דניאל - Danielrocks123 20:47, 23 August 2007 (UTC)

Speaking for myself, when this rolls out I planned to have a personal system whereby people could simply list themselves on my talk page for surveyor rights, and provided they weren't trolls I'd give the rights out to anyone who asks. Moreschi Talk 20:50, 23 August 2007 (UTC)
I think the concept exceeds the human resources we actually have available. I mean we're talking about changing the user class system to:
  • IP Users (effectively more trouble than they're worth under the new system)
  • Newly Registered User
  • Surveyor User
  • Admin/Sysop User
  • Reviewer User
  • Beaurecrat User

But there are only 1300 admins at the moment, and we've already identified that in order to be able to class an article as quality reviewed, the reviwer will (for practical and technical reasons) probably need to be either already an admin or of a reasonable admin like ability. So if there are only 1300 admins right now, how many reviewers are we actually going to have? I expect maybe 25 to begin with, raising to 50-100 maybe in due course. So out of the nearly 2 million articles, we're talking about less than several hundred reviewers, and every article needs to be read. Its just not practical at all. It creates so much new work that it will slow other admin heavy tasks like AfD, IfD and so forth. And this isn't even touching on the issue of how can an expert in automobiles possibly comment on the quality of a medicine or biochemistry article? Then we have to consider the slippery slope of asking people about their qualifications, bam zip before you know it, its citizendium. WikipedianProlific(Talk) 21:10, 23 August 2007 (UTC)

At the moment, we effectively have one reviewer - the "Featured Article Director" (Raul654). Increasing that to 25 sounds great to me. --Tango 22:38, 23 August 2007 (UTC)

The problem is that IP or registered status is not a good indication of editor vs. reader

In a perfect world, the Wikipedia servers would look into a crystal ball and somehow magically know whether each person accessing Wikipedia was a frequent editor and which are strictly viewers-only. In that perfect world, strict readers would always see Sighted versions, and editors would always see Most Recent versions. If that perfect world existed, I would advocate rolling out flagged revisions to all pages, not just some quality articles and as an alternative to semi-protection.

Lacking a crystal ball, this proposal aims to use IP vs. registered status as an indicator... and I just don't buy it. Although, granted, I am assuming this based only on gut feeling. It would be interesting if there were a way to measure over a given time period how many IP addresses were view-only (no edits), and how likely a registered user was to make an edit in a given day. Maybe the correlation is stronger than I think.

In any case, even without this "crystal ball" to determine editor vs. reader, this at a minimum seems like a good alternative to semi-protection. But I don't see it being generally useful beyond that. --Jaysweet 20:48, 23 August 2007 (UTC)

I can't imagine that the number of editors (including anon IPs) is anything like responsible for wikipedia being the 8th-most-viewed site on the net. Surely it's obvious that the vast majority of anon IPs are just readers (and incidentally, the purpose of wikipedia is primarily to cater to them, not to editors). PaddyLeahy 21:52, 23 August 2007 (UTC)

Categories and quality

Broadly support in principle, amazed that we haven't attempted this before. There's been several comments on BBC radio programmes about the ability to screw up articles, so we need to smarten up before we become discredited. It's parallel, in software terms, to development and release versions. Comments

  1. How will categories and their contents be affected, ie, does a new category have to be reviewed and sighted before being universally visible?
  2. What happens to material pre-existing this procedural change? Does it all have to be reviewed?
  3. Could quality ratings (above a grade which has to be widely reviewed) be shown on the articles themselves?
  4. Stubs can be useful and might prompt a new contributor.
  5. Some vandalism is subtle, not the delete all or school boy obscenity varieties, ie corrupting info, etc. Reviewing could capture this before "release".
  6. The devil's in the detail.

Folks at 137 21:03, 23 August 2007 (UTC)

To address that last bit in the first paragraph, I don't think that we should be giving too much credit to what people think of us when deciding whether or not we should add a feature. We're here to build an encyclopedia, not please everyone (obviously we don't want to be held in as high regard as Uncyclopedia, but we shouldn't be fixated on our perception, either; a fine balance is what is needed). EVula // talk // ☯ // 21:17, 23 August 2007 (UTC)
Hmmm, if the work isn't well regarded as reliable then what's the point? If this new feature (or a version of it) strengthens Wiki's quality, then go for it. But can someone address my points? Please. Folks at 137 21:57, 23 August 2007 (UTC)
1,2,4 Pages without a sighted version behave exactly as at present. So sighting will spread gradually as editors take the time to do the initial checking on various articles, starting with popular and/or high-quality ones. 3: I think the software does this...see test wiki at top of page. 5 So can sighting, since you always get to check the diff against the last sighted version. 6: That's why we need to introduce it to see how it really works. It's turn-offable. PaddyLeahy 23:47, 23 August 2007 (UTC)
This message appears: "This is the latest quality revision, approved on 19 August 2007. The current revision can be modified; 6 changes await review. (+/-)". I'm interested in advertising when an article is GA or above. It's also possible to limit the quality reviews to these articles, at least while we're in "catch-up" mode. Folks at 137 06:38, 24 August 2007 (UTC)

When?

What would be the general edit count/time of serve to get this? It should be automatically given fairly early. It should be given early enough that the user probably would not know that they have this ability right after it is given because they would still be too new like me and the move ability.Zginder 21:05, 23 August 2007 (UTC)

I believe the number being thrown around at the moment is 300 edits and 30 days on the wikipedia. I dislike though, I had uploaded 2 featured pictures before I'd got 300 edits. Edit counts aren't indicitive at all of a users contributions to wikipedia. However, these sorts of details are quite dynamic and likely to change closer to implementation if it ever happens. WikipedianProlific(Talk) 21:22, 23 August 2007 (UTC)
One admin (I forget who) earlier said they would give Surveyor status to "anyone who asked as long as they weren't an obvious troll or vandal." That sounds about right to me. Out of the countless vandals and bad faith editors I've seen, there have been maybe two or three cases where it wasn't obvious to me in the first five minutes whether a given editor was good faith or bad faith. --Jaysweet 21:26, 23 August 2007 (UTC)
However, some of us long-time editors, who are more egalitarian in our outlook, and who do not want any part of such a program, will be shut out. Badagnani 21:27, 23 August 2007 (UTC)
Main space or all edits?Zginder 21:29, 23 August 2007 (UTC)
I agree that edit counts are not a good way to judge quality of contribution. However it's hard to see how someone reasonably gets to 300 or 500 edits (even minor ones) without having a high, high likelihood of being a positive contributor (and even if they're a vandal, the worst that happens is that their vandalism gets reverted the way it does now). On the other side, it's pretty easy for someone to have 10 or 20 edits and have also proved themselves to be capable editors; they will also be given the sighting tag. It seems like a win-win situation to me.
Regarding the possibility that someone would rack up 500 edits, wait 30 days, and then go crazy vandalizing: do people sit up nights now worrying that someone will rack up 1000 edits, interact nicely with everyone, contribute to various projects, pass an RfA, and then go crazy with the administrative tools? I don't. Yes, admins out of control has happened, but it's not something that happens with enough frequency that we worry about it. Nor will vandalizing surveyors. -- Myke Cuthbert (talk) 21:42, 23 August 2007 (UTC)
It should take a lot more 30 days to get 300 edits. It took me more than a year to get 500 tootle edits and I still do not have 500 main space edits after almost 18 mouths.Zginder 23:50, 23 August 2007 (UTC)
It depends on what kind of edits you make. I built up my first few hundred edits very slowly, however if you spend an evening on RC patrol, you could probably make 100 reverts. Do that for 3 days, and you've got your 300 edits. Or you could do some stub sorting, or disambiging, or categorising. All useful jobs which involve making lots of very quick edits. --Tango 00:45, 24 August 2007 (UTC)

Section Glitch?

I tried editing the section titled "A few points" and when I hit the edit button the entire text does not show up. It stops at a comment that says: "Hmmm..." by VoiceofAll. What is this? Brusegadi 21:40, 23 August 2007 (UTC)

New Vandalism?

The good thing is that this will discourage the easy to spot vandalism. Yet, what if vandalizing becomes hard enough that people really try. Non-factual edits that are hard to spot can become the next big thing for the lonely highschool boys all over the world. So all those guys that write things like "WHOAAAAAA!!!!!" will become more sophisticated and begin to change the birthdays of not well known characters, geographic locations of cities, etc... They would not even need the power to 'approve' this edits. I can imagine well intentioned "approvers" being convinced by the arguments provided by the IPs... What I am trying to say is that; are we trading one evil for a worst one? Brusegadi 21:49, 23 August 2007 (UTC)

It's possible, but I just don't see a group of high school boys slapping each other high fives and calling their friend's cell phones to say, "Hey, check out the Wikipedia article on Joyce's Ulysses. Yeah, do you see it! It now says it was written in 1913, before WWI! Isn't that a hoot!" -- Myke Cuthbert (talk) 21:53, 23 August 2007 (UTC)
This kind of vandalism happens already. But it shows up in the diffs: if someone changes a date or similar you should always be suspicious. PaddyLeahy 21:59, 23 August 2007 (UTC)
Yeah, I just worry that the registered volume of such vandalism will increase. I admit, my view is rather pessimistic. Brusegadi 22:03, 23 August 2007 (UTC)
For articles that matter (unlike Ulysses), there should be enough knowledgeable people watching to quickly revert such changes. (I'm only partly kidding.) Ben Hocking (talk|contribs) 22:18, 23 August 2007 (UTC)
I don't think that the intelligent vandals are as big an issue to wiki credibility. If not being able to make idiot edits makes vandals smarter then we are doing the entire world a service. It doesn't make vandalism any harder to spot, and it should discourage all the bored High schoolers out there. Adam McCormick 00:29, 24 August 2007 (UTC)
Subtle just isn't what most vandals are going for. It takes a lot more work than inserting random profanity or blanking a page. Sxeptomaniac 21:57, 24 August 2007 (UTC)

Editor removes notice of this discussion, this time after 4 hours (previously just 1 hour)

Edit summary for this removal: "*sigh* ... we went through this debate months ago. We do not need to advertise this to everyone -- at present time it's about solidifying a proposal rather than bringing in the useless banter with ads" Badagnani 01:42, 24 August 2007 (UTC)

Where can someone see the watchlist notice changes (and discussion about it if there's any)? ♫ Melodia Chaconne ♫ 02:39, 24 August 2007 (UTC)
Per the above comment, is the user saying that the proposal's general principle has already been deemed to be approved by a majority consensus? Or are they referring to the claim that it was previously decided not to advertise the proposal? VanTucky (talk) 02:43, 24 August 2007 (UTC)

Dear Melodia Chaconne (that's a very nice username!): you can see the removal of the notice here; there was no discussion, as before. Badagnani 02:54, 24 August 2007 (UTC)

New edit summary from same user: "Archiving lots of stuff including the asanine straw poll .. the last thing we want now is to turn this into a bloodbath of voting by people who don't even understand the concept)". Badagnani 02:52, 24 August 2007 (UTC)

Who is this? VanTucky (talk) 02:58, 24 August 2007 (UTC)
A future Surveyor, of course! Sdedeo (tips) 03:11, 24 August 2007 (UTC)
Not to be confrontational here, but if you are going to obliterate a huge segment of the discussion, you might first want to seek some opinions before doing so, you could've at least saved it on a separate sub-page, even if it was "asanine" in your opinion, or even "as a ten". dr.ef.tymac 03:45, 24 August 2007 (UTC)
I'm afraid I just didn't see the need to have to discuss it before hand -- the discussion page was too long (well over 350k), and the discussion I archived was either static or not helpful. If anything got archived that oughtn't have been, it is indeed quite easy to restore it to the current discussion page; thus why I didn't see it as being a big deal. Generally, the wiki way is to be bold and discuss only when we really need to. AmiDaniel (talk) 17:41, 24 August 2007 (UTC)

It was "archived" here. Badagnani 03:56, 24 August 2007 (UTC)

Oh, good lord. The watchlist notice was removed for exactly this reason -- such a public advertisement brings nothing but hot-heads and nonsensical straw polls to the discussion. At present time we're simply trying to formulate a decent proposal; once said proposal is formulated, then it will be brought to a "vote". On the same train of thought, that straw poll gives people the illusion that we're voting on the exact text of this proposal and discourages rationale discussion about its faults. I do find it quite strange that you should refrain from chastising someone for making such blatant changes to the interface without discussing (especially in a case known to be contentious), but rather should chastise someone who returns to it its initial state. In any case, if some active discussion was archived, please do restore what is needed -- I thought it was quite clear however that this straw poll was doing nothing but leading to more confusion, and indeed the talk page was quite ridiculously long (if I might add, filled with the nonsensical redundant jabber that a watchlist message such as this brings). AmiDaniel (talk) 04:16, 24 August 2007 (UTC)

Well, I for one don't think a straw poll would do any good at this point. Even without a watchlist notice, they do attract those who don't understand what such a thing is really about. But as to the "nothing but hotheads..." comment, I find it to be impertinent, to put it lightly. I myself found this discussion via the notice, and I no one has accused me of being an ignorant crazy yet. Feel free to do it now though. VanTucky (talk) 04:29, 24 August 2007 (UTC)
I apologize -- that was phrased rather poorly. The implication was not that everyone who comes to the discussion through such a notice is a "hothead," but rather that it would seem most of those who follow such an ad lack understanding of (and do not take time to research) what the proposal is about, etc., and typically think they need to voice a pro or contra opinion, thus leading to them to partake in little straw polls rather than providing constructive criticism for improvement. The first time the advertisement went up on the watchdetails about this it did indeed bring in a lot of hotheads, ranting about how it was some grand plot being led by Sanger, etc., and we certainly don't need such nonsense, which is more what I was hinting toward--not yourself personally :) AmiDaniel (talk) 17:30, 24 August 2007 (UTC)

That's so funny that the editor who makes rash edits, removing crucial text from this discussion page, summarily deletes notices of this discussion for the wider WP community s/he seems to distrust, leaves consistently rude (and poorly spelled) edit summaries, insults other editors by implying that any editor who disagrees with this proposal is a "hothead," etc. is the one calling other editors "hotheads." It's the very definition of "irony." Badagnani 04:36, 24 August 2007 (UTC)

Please try to keep comments on the topic of the proposal, not on other contributors. Mr.Z-mantalk¢ 12:36, 24 August 2007 (UTC)
The poll wasn't especially useful, but the advertisement did bring a lot of rational discussion, in a frenetic burst, to this project. When it "goes live", you're going to get exactly the same kind of response -- so better to hear it before you marry yourself to an approach that will never be embraced.
Prior to today's changes, the proposal suffered from too much groupthink from a limited number of participants. It seems clear today's influx of comments helped wrench it from that. It's clear you felt assaulted by the large number of naysayers who suddenly showed up, but the more attention you pay to them, the more likely this project will actually succeed.--Father Goose 06:21, 24 August 2007 (UTC)
I would agree that we do need a wider range of opinions here. That way we can generate a proposal that is much more widely accepted. In the short time that the watchlist notice was up we removed "Quality" revisions from the main part of the proposal. However, a straw poll, where people do not even have to give reasons for their opposition (reasons that we might be able to address) is not helpful at this point. Mr.Z-mantalk¢ 12:36, 24 August 2007 (UTC)

Some good/changes have come out of the discussion influx, so I certainly welcome it. I'd add the notice back myself, but I'm afraid someone would add a "straw poll" section again. Voice-of-All 12:38, 24 August 2007 (UTC)

I certainly have no problem encouraging constructive contribution to and discussion about the proposal; however, I would strongly suggest advertising it through another forum, such as the Village Pump, etc. AmiDaniel (talk) 17:34, 24 August 2007 (UTC)

This proposal is certainly not ready to be put to a straw poll. There are still major changes being made. Some kind of democratic final decision will probably be required for something this big (there is no way we'll ever achieve a consensus), but we're not there yet. A watchlist notice should only be put up once we're ready for a poll. People interested in working out the details of the proposal can find their way here by other means (village pump, etc). --Tango 15:27, 24 August 2007 (UTC)

I made the straw poll. Oops. Eyu100(t|fr|Version 1.0 Editorial Team) 19:26, 24 August 2007 (UTC)
It's okay. I was against having one, but it let us know some things about where people stand (for instance, now we know that we were not likely to have a real consensus either way for the drafted proposal as it stood then). No real harm done. -- Myke Cuthbert (talk) 21:36, 24 August 2007 (UTC)
Of course, you can't expect a wide discussion on Wikipedia to reach a consensus on a proposal which might change things. There will be many people reading a few lines and writing twice as many lines objecting. -R. S. Shaw 22:06, 24 August 2007 (UTC)