I understand that this is a planned feature for the MediaWiki software.
I heard recently that a co-founder of Wikipedia has become highly dissatisfied with it on account of it containing so many factual errors that it was useless (and beyond repair), and he's quite right - this is a major issue that needs to be addressed. Obviously, the ability to mark revisions is the perfect solution. If there was a way to pick out a revision as being error-free (I assume, synonymous with "stable"), Wikipedia could potentially progress towards being an academically-citable encyclopedia.
I was just wondering who would feasibly *do* the marking as a stable revision? Obviously if this can be done by any users then there will be no advantage to it (as just the same liability toward inserting errors will transfer into a liability towards marking stable revisions which aren't actually stable). If you restrict it to registered users then there will still be no advantage, as even long-time registered users often vandalise and get things wrong. If you restrict it to admins then there will be too few of them.
The real problem is that it will take proper peer-reviewing - by experts - to really mark an article as "stable" in the sense of containing none of the errors and mistakes that caused the aforementioned co-founder to give up on Wikipedia. Obviously this is because any average editor (even an admin) is not necessarily qualified to declare an article error-free. Certainly, if nothing else, it will take expert-reviewing to bring an article up to "citable" standards.
So how do we currently suppose this will all work? Will the Foundation hire experts to check articles? Will we rely on expert volunteers contacting the Foundation so that they can be given "expert" accounts that can mark stable revisions? Or will we just allow long-time trusted editors to mark versions as stable, which leaves us in the same position of not knowing whether the article is *mistakenly* stable or not?
One feasible way I can see this as working is defining an arbitrary amount, say 100, that has to be reached for an article to become stable. If one person marks a revision as stable, it gets +1, and if they are a more trusted editor (been around for longer, done more major non-reverted edits) then it may get +5. If someone marks it as unstable it gets -5 (weighting towards holding back). And so on. Then if the article reaches 100 it becomes stable. This method roughly solves the problem of there being vandal or mistaken stable articles, but assumes that one revision of an article will stick around for long enough to be evaluated in this manner. Will we have to freeze the page after an admin puts it into "evaluation mode", or perhaps set it aside into a subsidiary page where it is evaluated, after that revision has been nominated for Stable Revision Evaluation? Obviously this is all a very tricky issue because we're dealing with a wiki!
I was just wondering what people thought of these issues, and what plans there are, if there are any.
I once suggested permanently semi-protecting all featured articles.
2007/4/13, Virgil Ierubino virgil.ierubino@gmail.com:
I understand that this is a planned feature for the MediaWiki software.
I heard recently that a co-founder of Wikipedia has become highly dissatisfied with it on account of it containing so many factual errors that it was useless (and beyond repair), and he's quite right - this is a major issue that needs to be addressed. Obviously, the ability to mark revisions is the perfect solution. If there was a way to pick out a revision as being error-free (I assume, synonymous with "stable"), Wikipedia could potentially progress towards being an academically-citable encyclopedia.
I was just wondering who would feasibly *do* the marking as a stable revision? Obviously if this can be done by any users then there will be no advantage to it (as just the same liability toward inserting errors will transfer into a liability towards marking stable revisions which aren't actually stable). If you restrict it to registered users then there will still be no advantage, as even long-time registered users often vandalise and get things wrong. If you restrict it to admins then there will be too few of them.
The real problem is that it will take proper peer-reviewing - by experts - to really mark an article as "stable" in the sense of containing none of the errors and mistakes that caused the aforementioned co-founder to give up on Wikipedia. Obviously this is because any average editor (even an admin) is not necessarily qualified to declare an article error-free. Certainly, if nothing else, it will take expert-reviewing to bring an article up to "citable" standards.
So how do we currently suppose this will all work? Will the Foundation hire experts to check articles? Will we rely on expert volunteers contacting the Foundation so that they can be given "expert" accounts that can mark stable revisions? Or will we just allow long-time trusted editors to mark versions as stable, which leaves us in the same position of not knowing whether the article is *mistakenly* stable or not?
One feasible way I can see this as working is defining an arbitrary amount, say 100, that has to be reached for an article to become stable. If one person marks a revision as stable, it gets +1, and if they are a more trusted editor (been around for longer, done more major non-reverted edits) then it may get +5. If someone marks it as unstable it gets -5 (weighting towards holding back). And so on. Then if the article reaches 100 it becomes stable. This method roughly solves the problem of there being vandal or mistaken stable articles, but assumes that one revision of an article will stick around for long enough to be evaluated in this manner. Will we have to freeze the page after an admin puts it into "evaluation mode", or perhaps set it aside into a subsidiary page where it is evaluated, after that revision has been nominated for Stable Revision Evaluation? Obviously this is all a very tricky issue because we're dealing with a wiki!
I was just wondering what people thought of these issues, and what plans there are, if there are any. _______________________________________________ Wikipedia-l mailing list Wikipedia-l@lists.wikimedia.org http://lists.wikimedia.org/mailman/listinfo/wikipedia-l
On 13/04/07, Virgil Ierubino virgil.ierubino@gmail.com wrote:
If there was a way to pick out a revision as being error-free (I assume, synonymous with "stable"), Wikipedia could potentially progress towards being an academically-citable encyclopedia.
My understanding is that "stable" is simply that. Not right, not approved, but... stable. A non-vandalised version of the current consensus on the article - you'll see the revisions from before someone vandalised it, or before a violent edit war kicked off, as the stable version, but it doesn't mean they need to have been individually factchecked.
My understanding is that "stable" is simply that. Not right, not approved, but... stable. A non-vandalised version of the current consensus on the article - you'll see the revisions from before someone vandalised it, or before a violent edit war kicked off, as the stable version, but it doesn't mean they need to have been individually factchecked.
That's exactly my understanding too. It would be good if using the same system an article could be marked as "fact checked". We could come up with a system for experts in the field to read the article and agree that it is factually correct. I think we should implement such a system for featured articles (existing ones, not candidates - this isn't intended as another hoop to jump through to become a FA) at the very least. It doesn't need to wait for the software to know how to do it, a system using templates should be fine.
I agree with what's been said. There's a "stable" revision - not vandalised, not being warred over, etc. - which we can trust most users to select - but much more interesting would be the ability to mark out "fact checked" revisions - ones without errors. It's that issue which I'm getting to.
I think it makes a lot of sense to put the two systems together - really, the ability to reach the fact-checked revision needs to be integrated into the software. There can be no academic citation of " http://en.wikipedia.org/wiki/.....&oldid=324234&curid=234234&bla..." - this URL is of precisely the same form as any NON-fact-checked revision. People will, however, be able to cite something more like http://en.wikipedia.org/stable/Biology (as opposed to http://en.wikipedia.org/wiki/Biology) - the former URL presents the falsifiable credibility required.
No I agree, things on Wikipedia are definitely going well. But this kind of article verification will definitely be needed at some point in the future, if Wikipedia ever hopes to be useful to anyone other than the leisurely reader.
Thomas Dalton wrote:
My understanding is that "stable" is simply that. Not right, not approved, but... stable. A non-vandalised version of the current consensus on the article - you'll see the revisions from before someone vandalised it, or before a violent edit war kicked off, as the stable version, but it doesn't mean they need to have been individually factchecked.
That's exactly my understanding too. It would be good if using the same system an article could be marked as "fact checked". We could come up with a system for experts in the field to read the article and agree that it is factually correct. I think we should implement such a system for featured articles (existing ones, not candidates - this isn't intended as another hoop to jump through to become a FA) at the very least. It doesn't need to wait for the software to know how to do it, a system using templates should be fine.
Most discussion about "stable" versions has indeed been in the anti-vandalism context, but I would personally like to see the feature work a lot harder than that. As the saying goes, "You must learn to walk before you can run," so I will be happy to see anything functional in this area. Hopefully it can be taken further after that. Getting something like this going would be a great substitute for the rude behaviour that one often encounters.
Ec
wikipedia-l@lists.wikimedia.org