I have neither the time nor care to figure this all out, but I figured I ought to point it out.

Basically, the developers are doing away with the reviewer group and letting anyone who is autoconfirmed (10 edits, 4 days of registration) approve revisions for public viewing. This essentially makes flagged revisions quite useless, except for obvious vandalism.

But note the humor, because it's very close to simply sprotecting all articles, but letting IPs make suggestions on the talk page. Which has, ahem, been suggested by certain parties in the past.

Teh Dewelopers might note that you SHOULD in theory be able to custom-set the "promotion" requirement for any article, and have each one set differently according to how often it's vandalized and how important it is. For pokemon, IPs might edit, since who gives a crap?. For other articles, 10 edits 4 days, so essentially only autoconfirmed users can promote. For more heavily vandalized ones, 20 edits, 8 days. Then 40 and 24, etc. until you get to German standards for promotors. You could make this automatic or suggested, and have the promotion-requirement go up by 50% or 100%, everytime an article got "vandalized" by having somebody clearly vandal-promote it, and not by mistake but with the clear intent of committing harakiri with that "account."

The limit would be hypothetical articles like Carolyn Doran, which would require 8.5 years or 400,000 edits, which would mean that only Rjwilmsi, Jimbo Wales, and a few developers could promote a new version.

http://en.wikipedia.org/wiki/Wikipedia:Lis...number_of_edits

That would work well on articles where the tenacity of benefactors exceeds that of malefactors, but I'm inclined to doubt that is the more common case.

Not at all. This is a systemic lack-of-will problem (just like flagged revisions themselves), not a matter of "protection work" vs. "vandal work".

The admin can make protecting an article as difficult as he/she likes, just by how stupidly or intelligently they go about it. You can (for example) sprotect a given article for a year with a few clicks, and that means that, for a whole year, EVERY vandal has to use up a dedicated harakiri nameuser vandal-only sock-account to get at it, each one having to wait 4 days, keep track of a password, and make 10 edits. Of course, I assume here that each vandal-only account nameuser is indef-blocked the first time they vandalize, which is how it ought to be (it's fairly easy to tell a deliberate vandal from a mistaker, and certainly it's easy to tell a new vandalism-only account from an old contributor).

So, for EVERY one of these vandalisms, for a whole year, the work EVERY vandal does is more than the work the admin did, just ONCE, in protecting it. Assume 10 dedicated vandal-accounts who go through the nameuser rigmarole to get at the sprotected article, and work the ratio. The vandals are using up a lot more time than the admin, even at this level.

Of course, as with all things in life, it's possible to fuck this up. The admin can fumble protection of the article, sprotecting it for 72 hours or something, THEREBY causing some OTHER admin to have to do it all over again, in just 3 days, when the vandals start in again. How many times have I seen that? And, of course, vandal-only IP and nameuser acccounts can be coddled too, instead of indef'ed the moment they vandalize after a single warning (assuming ALL their edits have been vandalisms, which is the case almost all the time-- very rarely is an account a mixed bag, and for nameusers it hardly EVER is).

Now, consider the extra power of a flag revision super sprotect feature (which is basically what flagging is), where every time time a harakiri vandal-only nameuser account gets though the sprotection layer, you can simply double the sprotection with a click, from 10 to 20 edits, and 4 to 8 days, or whatever. With a single command you have now DOUBLED the work every vandal-only account has to do, to get at the article. So the admin work per article proceeds linearly, but the vandal work goes up more than exponentially. The latter because not only does each vandal now have to do double, then quadruple, then octuple the work that vandals have to do to be able to edit an article (and so on) but because the protection level applies to many vandals attacking a given article, yet for each level, the amount of protection is set upward only one time, by only one admin.

And yes (again ) it's possible to make this not work, by having admins who refuse to use the system. But that's not a problem with the system. The system I describe would work. The question is whether WP cares enough to adopt it. Right now, admins make each other work very hard sprotecting and desprotecting articles from IPs, which in turn are endlessly coddled. The only exceptions being Jimbo's userpage, the main page, and a few other sacred sites. If WP doesn't care enough to use sprotection properly, they won't care enough to use the even more powerful flagged revisions properly, either. But again, that has nothing to do with individual work on individual articles. That's a policy issue.

I do that, but the vast majority of administrators go through the standard 4 warnings cycle. Maybe that will change with flagged protection, but I think that's unlikely.



I'm confused a bit about your proposal. Would administrators set the "you need X amount of edits before you can edit this page", or would the software automatically reconfigure that based on certain criteria?

Both. The software automatically sets a suggested increase in protection, and some admin simply signs off on it, if it looks good. IOW, there's a standard policy of increasing protection levels, but they can be evaded in individual cases, if somebody wants to spend the loving care and time to do it, or there's a special reason for it.

Isn't it 4 days, 100 edits?

And I should know ... I have had to do some genuinely useful and interested editing, include new topic creation, before you sock banned me yet again.

Or is it only 100 edits if you want to edit on some topic page an admin has locked?

10 edits.

Yes it is 4 and 10 on enwiki, though it can be set differently on a per-project basis, but this is rarely done and the default is 4 and 0 (see config):

My prediction is that even if this flagged stuff is implemented, people here will still be disappointed. The wave of entities who behave idiotically and who can potentially edit is too big to be contained.

I certainly will be, as I will then have lost my bet with One