On Fri, 2003-12-05 at 22:13, Delirium wrote: Jimmy Wales wrote: This will probably be opposed by those who'll see it as just another step into hierarchical organization, but I think it might be a good idea to create a new, more-inclusive class of users, that has authoritative significance but no technical powers. Basically any user who has been here for some period of time (maybe 2-3 weeks or so), and shown him/herself to be editing in good faith. As we get bigger, I think there will be a lot more "fake" users trying to influence things, so restricting things like policy formation and votes to "real" users in some sort of formal way might be a good idea. I'd envision it being very easy to gain this status: even someone who's invovled in lots of edit wars should be considered a "real user", so long as they aren't purely a troll or vandal, or someone who just signed up 3 hours ago. This might actually have the effect of reducing the hierarchy somewhat, because right now sysops are a sort of de facto group of "trusted users", since the only defined groups we have are "sysops", "logged-in users", and "anonymous users". Sysops are too small a group, and logged-in users are too big a group (anyone can create 100 accounts if they wish). Making a larger group of trusted users without technical powers would reduce sysops to being just a subset of that group with additional technical powers, but no additional powers of any other sort. Ideas? -Mark I for one love this idea. In addition to addressing the problems that Delirium lists above, it would have the added benefit of giving users a modicum of validation from very early on, thus maybe reducing the number of walk-in walk-outs, and also reducing the number of folks applying for sysophood after 2 weeks and 100 edits who only really want validation for their existence on wikipedia. Now at the other end of things, maybe at this point on the English wikipedia we have enough sysops in attendance at any one time, that it would not be unreasonable to require two concurring sysops to perform sysopactions which are uncommonly needed but potentially particularly controversial. I am thinking about unprotection of pages and undeletion as well. Perhaps protection of page as well, with the proviso that only one of the sysops participating in the protection be uninvolved in editing the page or having taken a strong view about it. Is this technically feasible? Jussi-Ville Heiskanen (aka Cimon Avaro)

Toby Bartels wrote: And generally less, right, since after all a person can always 'undelete' by just starting the article over and copying and pasting the text? --Jimbo

Jimmy Wales wrote: Not entirely. For one thing, your solution destroys the edit history. Also, only admins can even view the text that they want to copy and paste (although that point is not relevant to current issue, since it is only about relative numbers of administrators). In short, undeleting by your method is not as good as undeleting by software. That is one reason why undeleting shouldn't be unduly restricted. -- Toby

[Sent both the wikiEN-L and directly to sannse.] sannse wrote: I think that you may be interpreting my remark out of context. The "undue" restriction on undeletion that I'm arguing against is to increase the requirement from 1 administrator to 2, ''while at the same time'' keeping deletion at only 1 admin. I'm ''not'' suggesting that undeletion be ''completely'' unrestricted, while keeping deletion restricted. Although in another post, I mentioned the ideal of making ''both'' of the operation unrestricted, I do believe that this would require some software changes. At the moment, neither deletion and undeletion is handled very smoothly. -- Toby

sannse wrote: No problem! -- Toby

Cimon Avaro (Jussi-Ville Heiskanen) wrote: Yes, you are. I will clarify. The word "never" indicates that I consider a matter of principle. I don't merely mean that I don't think that this should happen ''now'' -- I mean that it should ''never'' happen. On principle. Of course, principles can be challenged, and I might change my mind. Thus this is not the same as dogma. I left out stating any justification, because I hoped (and believed) that the principle, once mentioned, would be widely viewed as correct. If somebody challenges the principle (or its application to this case), then I would want to respond with justification for my position. If I expected people to change their minds ''without'' doing so, then that would be dogamatic. But I don't. I half expected you to say «When you put it that way, I agree.» -- on grounds of principle. And I half expected you to say «I disagree with that principle. What are your justifications?» -- and then I would justify it. As it is, you began a meta-discussion about my attitude. That's OK, The Cunctator does the same thing. ^_^ But I'll elaborate on the principle anyway, since I'm talking. The main principle is this: : On a wiki, undoing an action should always be as easy as doing it. We violate this in various ways -- page creation, for one thing, which is what deletion undoes. But that is handled by the software in completely different ways, so there's a reason to violate principle. But that's because software needs trump other principles: : When the Right Thing is impracticable, then do the Wrong Thing for now. -- Toby

Eclecticology wrote: Ideally, all of these features would be available to everybody. Deletion was made an admin-only process because undeletion was impossible (check old talk history on deletion policies). Now it might be expanded. Unfortunately, deletion is handled so differently by the software (and what's most relevant here is the inexperienced user sees it) that I doubt that people would easily treat is as just another edit. If the software can be made to treat it as just another edit, then there would be no reason to restrict it (hence also not undeletion). Another problem is our reliance on ''page'' deletion as a way of deleting ''specific revisions'' that infringe copyright. Since this is only effective when the infringement is a page creation, it's not a permanently workable method. That software may change too. -- Toby

is not Actually, the law ends up being forgiving for internet hosts who have copies of copyrighted material as long as it's clear that they're not intending to publish; for example, even though deleted pages are available to sysops and in the backup dumps, as they're well hidden they're not a threat to the legitimate publisher/copyright owner. Obviously, we need to make a good faith effort to respect copyright, but we don't need to suffer from copyright paranoia. This is not to say that the current system of dealing with copyrighted, unlicensed material on Wikipedia isn't broken. But I don't believe that the only legally sound solutions require total guaranteed erasure of any copyrighted text. Think of Google and its cached copies of websites. might You can add my rah rah here. With the stub-marking feature, deletion of entry histories just to get back the ? (or red) link is unnecessary.