(copied from WP:VP):
- Interesting but my initial reaction is that this is too complicated. Technically, it might be a complex overhead, especially because of the proposed delegation system and in practice, I'm afraid you would find that the preferred version is ancient, even though the article has been through a definite progression since or (even worst) articles with low traffic in which an unacceptable version becomes the preferred one because a handful of sockpuppets have deemed it as such. Like it or not, random users will see the preferred page as a mark of confidence. I much prefer the idea of stable versions where the complex process partly ensures that only articles of high importance are tagged. Pascal.Tesson 20:45, 4 October 2006 (UTC)
- The base version shouldn't require much more server/database overhead than the watchlist feature itself, which needs to maintain a large set of {user, article} records in the database. The delegation system might involve overhead; I'm not sure exactly how much. At any rate, I hope you don't mind if I copy these comments over to the talk page. --EngineerScotty 20:52, 4 October 2006 (UTC)
I'm not completely sure this is workable. If I put 'preferred' on everything I read or edited which I thought was a good article, that would essentially mean I had to keep track of hundreds of different articles. If I hadn't got round to updating my preferences, I might have as preferred something which is very out of date. It's an interesting idea, but I'm not sure what it adds compared to improving the protection of Good and Featured Articles.
--Merlinme 13:04, 5 October 2006 (UTC)
Stable version are soon planned to be tested on German Wiki. Weakness of the current proposal is fragility - such a feature is even easier to misuse and war over than text changes. Pavel Vozenilek 23:22, 5 October 2006 (UTC)
- SW support is implied--there is no way to "tag" an old revision of an article with metadata, at least not for non-admins and/or gurus. --EngineerScotty 00:28, 6 October 2006 (UTC)
- May I ask what this proposal is driving at, that doesn't get addressed in the German stable versions system? While I don't see the current suggestion as workable, I wonder whether there's the core of an idea here that might become a good thing. One question from my standpoint has to do with objectivity: for FAs and GAs one type of rollback default would be the version that passed GAC or FAC. I agree that edit creep is a problem, yet I'm concerned about the potential for edit war abuse here. Is there a way this proposal could insulate against exploitation? Durova 04:46, 6 October 2006 (UTC)
- That's what we are here to discuss. :) The page, as it stands, is my brief thoughts composed over a few minutes a couple afternoon; something that sorely needs to be scrutinized. This is nowhere near the "proposal" stage at this point. I've got to sleep now, but I'll write some use cases later. --EngineerScotty 05:39, 6 October 2006 (UTC)
- May I ask what this proposal is driving at, that doesn't get addressed in the German stable versions system? While I don't see the current suggestion as workable, I wonder whether there's the core of an idea here that might become a good thing. One question from my standpoint has to do with objectivity: for FAs and GAs one type of rollback default would be the version that passed GAC or FAC. I agree that edit creep is a problem, yet I'm concerned about the potential for edit war abuse here. Is there a way this proposal could insulate against exploitation? Durova 04:46, 6 October 2006 (UTC)
- I remember that the MediaWiki feature "voting on quality of articles" (with radio buttons (1)(2)(3)(4)) was rejected or postponed because of high potential for misuse by automated systems. Placing tags is only slightly safer.
- A working way to make stable version of an article may be (1) to bring together qualified (or at least some) people who will look at the article during the review period. (2) let them discuss what needs be fixed and implement the consensus, if reached. A next stable version would require the same process. Pavel Vozenilek 01:05, 8 October 2006 (UTC)