On 06/07/2010 09:41 PM, John Du Hart wrote: Hi, John. I am indeed aware of the big poll and much of the discussion. My question wasn't, "do people still want it?" It was, "Are people ready to go in a week?" As I said, I think the answer's yes, but I wanted to offer people one last chance to raise any issues. It turns out that I do follow wikitech-l. I and the rest of the team believe we are technically ready. If you have cause to believe otherwise, wikitech-l would be a good place to lay out your concerns. Thanks, William

On Tue, Jun 8, 2010 at 6:29 AM, Rob Lanphier <robla(a)wikimedia.org> wrote: Might it be worth gathering all Flagged Revs pages and moving them to [[WP:Pending Changes/Historical discussions/...]] with redirects, to make clear what's what? FT2

On Tue, Jun 8, 2010 at 7:34 AM, FT2 <ft2.wiki(a)gmail.com> wrote: Would my explanatory graphic be useful? Does someone want to help me sync up the language with what is currently in use? (or does someone else just want the SVG source, I've given copies to some other WMF communities for their own usage). [Responding to FT2 because I'm looking to his leadership on coordinating this kind of thing]

I think I have an idea how to do it. The "Flagged revisions" main page summarizes FR and where it went (patrolled revisions, pending changes etc), with links to all major pages on the topic. So its a reference to provide back information on past discussions, proposals etc. Those pages which are historical in the sense of "for reference only / not live proposals or policies" floating in project space, can be moved under that page and acknowledged on it. Also a link from that page to "Pending Changes" for the live proposal. One thing needed - can someone reply to this thread with a list of all Flagged Revs related pages (whether RFCs, proposals, or major threads) so we can see what's out there? FT2 On Tue, Jun 8, 2010 at 6:57 PM, Rob Lanphier <robla(a)robla.net> wrote:

On Wed, Jun 9, 2010 at 10:26 AM, Rob Lanphier <robla(a)wikimedia.org> wrote: http://en.wikipedia.org/w/index.php?title=Special:Search&fulltext=Searc… Will get most of the pages (in the WP ns at least). -Peachey

Just for the sake of understanding better for next time, would people have preferred that we launched later? We gave a date as soon as we were reasonably confident that we could hit a date for the minimum feature set, based on the theory that people wanted this ASAP. But naturally, we could have announced a later date. If in the future the community would like X weeks of cushion to prepare for a feature change, I think that's a very reasonable thing to ask for. The time wouldn't be wasted; there's plenty of good stuff left to do. William On 06/12/2010 06:59 AM, Cenarium sysop wrote:

On Sat, Jun 12, 2010 at 10:13 AM, William Pietri <william(a)scissor.com> wrote: Honestly? We could have picked any delay and there would still be some people it will suprise when it hits. The Vector skin change had been in the header for months and still suprised people. It's been announced and discussed in advance to an adequate degree, IMHO. -- -george william herbert george.herbert(a)gmail.com

As I'm actively involved in the preparation of the trial, I assure 10 days is short. Most people don't get involved until a launch date is fixed, especially in this situation where we had to wait for a year with nothing coming so people just waited for something consistent to get involved. And now we need to quickly find consensus on remaining policy issues, write documentation, etc. So I wouldn't have preferred that it be launched later, because we've waited enough, but that the launch date be given at least 3 weeks in advance. On Sat, Jun 12, 2010 at 10:04 PM, Amory Meltzer <amorymeltzer(a)gmail.com>wrote;wrote:

On 12/06/2010 18:13, William Pietri wrote: Personally, just launch the damn thing already! KTC -- Experience is a good school but the fees are high. - Heinrich Heine

On 06/12/2010 02:22 PM, David Gerard wrote: Great! That is certainly how I feel about it, so I'm relieved the community feels likewise. For those similarly enthusiastic, there is, as Cenarium points out, still some community work that needs to be done: http://en.wikipedia.org/wiki/Wikipedia:Requests_for_comment/Flagged_revisio… One thing that definitely needs sorting out is when the limit of 2000 Pending Changes articles gets raised. We'll get some numbers on performance impact after it has been in use for a bit, but there has been some talk that we shouldn't raise the number purely because the servers can handle it. It'd be nice to have a clear signal on that by the time the technical issues are sorted. William

On 06/13/2010 03:59 PM, David Goodman wrote: Ok. Since the 2000 limit initially came from the Foundation side of things rather than from the community, I was being especially careful not to presume. But from the mailing list and on-wiki goings on, it looks like the community prefers a software-enforced numeric limit regardless of technical capacity, so we'll plan to leave the limit in place until we hear otherwise. William

You'll soon have your answer here: http://en.wikipedia.org/wiki/Wikipedia_talk:Reviewing#Proposing_a_delay_to_…. There are many outstanding issues to address and still quite a deal of preparation to be made. Again people didn't get involved until a launch date was fixed, it may be hard to define one in advance, but that's how it is. On Mon, Jun 14, 2010 at 2:46 AM, Thomas Dalton <thomas.dalton(a)gmail.com>wrote;wrote:

On 14 June 2010 09:12, Cenarium sysop <cenarium.sysop(a)gmail.com> wrote: Or we could just do it, since objectors have had *three years* to faff about in. - d.

On 06/14/2010 01:12 AM, Cenarium sysop wrote: I am going to stay quite thoroughly out of the discussion as to community readiness or the actual date; the community asked for this ASAP, and if the community changes its mind and wants it enabled later, that's entirely up to the community. However, I do want to say two things. One, delaying isn't free. There has been a lot of work in prep for this, and some of it will have to be done again for a new date, especially on the ops and press sides. If we cancel the June 15th rollout, then once the community is happy that things are sorted, we'll have to go back and find a new date that works for the FlaggedRevs people, the ops people, and the communications people, and hope that we can get time for reporters on Jimmy's calendar again. Two, the community has been asking for this ASAP all year, so any request to delay has to be clear enough and strong enough to obviously override that long-established and widely supported consensus. So far it's 7 to 5, which is neither clear nor strong. Regardless, we will be rolling out the FlaggedRevs code changes to all wikis tonight as scheduled. That should have no effect on enwiki and hopefully small effects on current FlaggedRevs users, so there's no reason to delay that part of it. William

Despite the fact that I do have reservations about several aspects of this trial deployment, I do recognise that this is indeed a *trial*, and that the purpose of a trial is to learn, and to try out new ideas to see whether they work in whole or in part. The opportunity to learn is the reason that I feel we should proceed at this time. Having said that, I feel extremely strongly that we need to be willing to take extra steps to protect the editors (particularly the non-admins) who are putting their editorial reputations on the line to help us find out if this is a workable process. We need for the participants to make mistakes (so we can figure out how to fix them), to give us honest feedback from the "ordinary editor" perspective, to make sure that the tool works in the way it's expected to, and to figure out whether the parameters we've set for tool use are realistic and viable. I think one thing that is missing from this entire trial process is that there are no broadly accepted objectives, and conflicting supposed measurables for determining whether or not the tool has made a difference. These are some of the things I've heard bandied about: *Pending changes will encourage more non-editors to try to edit, and these new editors will become part of our community. ---Just because someone edits an article doesn't mean that they actually helped. We will have no realistic way of measuring how many new editors made useful edits and how many made vandalistic ones. Nor, if they are IP editors, will we be able to say with any certainty whether they actually stick around to become contributing editors. ---I'd like to hear from someone in the know whether or not we will be able to determine if new accounts created during this period had their first edit on an article under pending changes protection. If we can't tell that, then we cannot attribute any higher-than-usual number of new editors to the use of this tool. *Pending changes will protect more BLPs. ---The same criteria for protection continue to apply. If the article does not qualify for semi- or full protection, it does not qualify for pending changes either. Pending changes is being billed as an alternative to semi- or full protection and is explicitly not to be used as a means to extend protection to articles that would not otherwise qualify. At the end of the trial, there should be no significant difference in the total number of articles covered by one of the three forms of protection than there is at the time we start the clock. *Pending changes will help stop edit wars ---Edit wars are content disputes, and need to proceed through our normal content discussion process; pages that have been protected because of edit wars are not eligible for pending changes. The only exception is if an article is semi-protected to keep anons/unconfirmed users from repeatedly adding the same vandalistic or BLP-violating material, and that is vandalism control as opposed to edit-warring. *Pending changes will reduce visible vandalism ---Um, no. If every review of a pending change is carried out correctly, there should be no difference in the amount of vandalism viewable by the general reader. That's because otherwise the articles would have been semi- or fully protected, and almost all vandalistic edits would have been rejected by the software. *Nobody's being prevented from editing in the way they always have ---We won't know until we try this part. If we see autoconfirmed editors having their edits caught in the pending review queues of articles on Level 1 pending changes, then this is patently false; their edits have always been publicly visible from the time they hit "save". This is data that would be really valuable to capture, if there is a way to do so. ---As well, editors who take on "reviewer" permissions will automatically have their edits accepted, even on formerly fully protected articles (should we decide to try Level 2 pending changes). The reviewer permission goes with them everywhere, so they will now have to review any pending changes before making their own edits to articles that may be part of the trial. *Anonymous editors will now be able to edit the [[George W. Bush]] and [[Barack Obama]] articles ---No they won't. This was actually one of the first, and easiest decisions made by the on-wiki team looking at trial implementation processes. There is no reasonable chance that the number of useful edits will make up for the incessant vandalism and BLP violations in what are already a {{good}} and {{featured}} article respectively. They certainly won't be part of the trial, and even if the community decides to continue using this tool, almost every other BLP in the entire project would be a better candidate for pending changes than these ones. Even the German Wikipedia still has some protected articles. So....now that I have deflated everyone's expectations....We really do need to think about what we would consider to be a useful outcome in this trial. It's time to stop thinking pie-in-the-sky, and get down to what we'd consider a sufficiently positive outcome to proceed. Incidentally, I think it's important that we reinforce repeatedly that this is a trial. Trials end, and this one ends in two months. Unless there is a newly minted community consensus to keep this trial deployment going, I fully expect it to be turned off on August 15th, along with all the other bells and whistles that go with it (such as deactivating the reviewer permission). If there is no intention at this time to stop the trial and deactivate the extension on August 15th, I'd like the WMF and the developers to say so now. Because if that is the case, then this isn't a trial, it's a seat-of-the-pants deployment, and the very large section of the community that is already concerned about how this tool will be used will have every reason to believe they have been handed a pig in a poke. Risker

William Pietri wrote: You say crude, I say simple. If there are articles there needing full protection, nature will take its course, and they will end there in due time. Yours, Jussi-Ville Heiskanen

On 15 June 2010 01:12, William Pietri <william(a)scissor.com> wrote: Thank you, William; although I believed this was the intention, it is important to see it in black and white. I have lost count of the number of times someone has told the community "oh, let's just try this, if we don't like it we can go back to the other way," without any realistic intention to consider turning something off/reverting a policy/reinstating a practice. I look forward to seeing what all we've learned in the coming two months. Risker/Anne

Cenarium sysop wrote: This is a very dangerous view on the issue. This is what people who strenously opposed the new mechanism were most afraid of, and the supporters originally said would not be a danger. If this really happened, I could easily see many of the people originally in support of the new mechanism, could do a full volte-face and come strongly in opposition of the mechanism. Supporters of the original agreement often voiced the proviso that using the mechanism for semied/BLP's or whatever their personal threshold was, would never ever be a thin end of the wedge to spread things out to things we wouldn't semi currently. That is the *old* *agreement* on this issue. A huge drive by any tiny group of blow-hard editors to expand use of the mechanism beyond what we currently semi, could back-fire spectacularly. I don't dispute that in the fullness of time; years or decades from now, it might eventually go that route, but that is a completely different issue, and I suspect there would be many more important community supported initiatives that would have to be accepted in the interim, before that could remotely be acceptable. I agree with your point here. The mechanism shouldn't be used as a damper in edit wars. That way, madness lies. You could have hundreds of reverts back and forth never going live, and a Stygian Stable for the person sorting out through all that which revisions and edits to go live finally. Just a total Charlie Foxtrot in other words. Yours, Jussi-Ville Heiskanen

On 15 June 2010 02:38, Cenarium sysop <cenarium.sysop(a)gmail.com> wrote: Actually, it was impossible to try on the testwiki at the time, because the "reviewer" permission hadn't been activated yet. And the tables clearly state that the edit must be "accepted". There was no indication at the time in the documentation that any other option was possible or acceptable, and no way to test it at the reviewer level. Can you please identify methods in which we can measure the improvement here? Are you proposing, even before the trial starts, to start including articles that do not meet the criteria for page protection? Let's be clear, Cenarium; the trial is very specifically only to be used on pages that meet the *current* criteria for page protection; what you're suggesting here is something completely unrelated to the trial of pending changes in and of itself. Those who are making good faith edits (or requesting them) *might* be members of the community, but I'm not particularly inclined to include the drive-by vandals as such. Remember, my list was made up of things that various people have proposed as good reasons to institute pending changes. I completely agree with you that it was never intended, but some people still think it was. I removed it from the draft policy, in fact; I have no idea who added it in. Yes, indeed. Another place where we agree! Unfortunately, the very first press publication about this change specifically suggested that the [[George W. Bush]] article would become accessible to unregistered and newly registered editors. I'm not the enemy here. I have something of a well-earned reputation as a BLP absolutist and I spend a good part of every week addressing the fallout of vandalism. But I've been around this project too long, and seen too many exceedingly buggy software deployments and major attempts to hijack policy and practice. I can turn a blind eye to a fair number of these, if they don't affect matters within my usual area of assumed responsibility. This one, however, is openly being billed as one thing (improved editing accessibilty for non-registered and newly registered users on articles they've previously been shut out of), but it's pretty obvious that there is a significant desire to use this tool to do exactly the opposite, and actually restrict automatically visible edits from non-registered and newly registered users on a much larger swath of articles. Keep the trial limited to articles that currently meet the protection policy, and there is a reasonable chance of success. Start pushing it further, and the outcome will be considerably more in doubt. Risker/Anne

You know well that there are no objective way to say if an article meets the 'criteria' or not. If you ask different admins about a particular situation, some will say no protection is warranted, some will say temporary semi-protection is, of variable length, and some say that indefinite semi-protection is. The protection policy says 'heavy and persistent vandalism or violations of content policy' for indefinite, and 'Subject to significant but temporary vandalism or disruption' for temporary, this allows for considerable discretion. And since pending changes protection is much less restrictive than semi-protection, admins will naturally lower their personal threshold for applying it. There are several admins who apply a threshold considerably lower than average, their semi-protections are often contested but almost always uphold, or with no admin going ahead to remove them. When several admins started to make use of ' liberal semi' for BLPs, there has been considerable objection (by me among others) but almost all protections stayed. There we see the two contradictory needs, to better protect articles, BLPs in particular, versus to keep articles editable. Excessive protection (of any kind) is bad; but BLPs subject to vandalism or BLP violations to a level where semi-protection would be within discretion, but just below the threshold where most admins would protect, is not satisfactory. By its flexibility, pending changes allows to better balance those two contradictory needs. A great advantage of pending changes protection is that we can see edits, so determine to a certain extent if protection is still warranted. With semi-protection we can only guess. So we'll be in better measure to remove pending changes protection were no longer needed. This means we'll simultaneously be able to handle more cases for protection, and remove protection where no longer needed. The total of protection may not even grow sensibly at all, but protection would be better distributed. We just need to keep an eye on the backlog and adjust if necessary. In the trial we may not readily see this happening, because it would be more limited and controlled, but I'm sure it will occur to a certain extent. This won't handle all issues, especially isolated vandalism and BLP violations, where protection cannot be used per policy, which is why we vitally need better monitoring tools, like patrolled revisions. I would strongly oppose any attempt to no longer regard the protection policy for using pending changes, or alter the protection policy to extend its scope. For discussion of methods, see Wikipedia talk:Pending changes/Trial. This is a very dangerous view on the issue. This is what people People were mostly afraid to see this becoming a FlaggedRevs implementation similar or close to that on de.wikipedia, which is very different from what I imagine. The idea of Yamamoto Ichiro<http://en.wikipedia.org/wiki/User:Yamamoto_Ichiro>to use flaggedrevs as an alternative to protection was a breakthrough because it allows not only more editability than classic protection but also to better control uses of protection, as I explain above, this allows a much finer distribution, to apply it where it is needed, and only where it is needed, more than classic protection would ever allow. Pending changes is now heavily associated with protection, even on the technical side. The protection policy acts as a safeguard against attempts to use it on pages where protection is unjustified, though we can't say what will happen in the future. But the protection policy does allow discretion, and with the flexibility of pending changes protection, I think this will naturally result in a better distribution of protection. Any protection can be reviewed by the community, and there's going to be much attention given to the backlog. The majority of the community wouldn't accept to use flaggedrevs on pages outside the scope of the protection policy, and I don't see this changing anytime soon. And this is imo for the best. On Tue, Jun 15, 2010 at 2:20 PM, Gregory Maxwell &lt;gmaxwell(a)gmail.com&gt; wrote:

Because it is an open, public mailing list where meta-discussion is supposed to be going on about the English Wikipedia. In any case, it's basically guaranteed there will be a portion of the community who will not be ready and a portion who will apparently be caught completely off guard despite the numerous on- and off-wiki discussions, watchlist notices, and anything short of having a bot send messages to all 12 million + registered users. -MuZemike On 6/8/2010 6:15 PM, K. Peachey wrote: