Jump to content

Wikipedia talk:Non-administrator rollback/Archive 1

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1

Why isn't this taking place there, or linking there to see past discussion? –Pomte 23:14, 30 December 2007 (UTC)

By all means add it, I just forgot - not sure where the best place to put it is so I'll leave that to you :-) Ryan Postlethwaite 23:15, 30 December 2007 (UTC)

roll back scripts

Is this different than things like Twinkle? Lawrence Cohen 23:19, 30 December 2007 (UTC)

It's simply faster - with twinkle, you go through the edit screen and then it saves the revert (even though twinkle doesn't show the edit screen) - rollback misses out the edit screen completely. It also cuts server load. Ryan Postlethwaite 23:21, 30 December 2007 (UTC)
Oh, neat. You just issue warnings by hand after? Lawrence Cohen 23:26, 30 December 2007 (UTC)
Warnings aren't automagically included with admin rollback, but some clever javascript might fix that. Probably over my meager skill level, though. Currently I give out warnings with a separate script (not quite "by hand", but close to it). – Luna Santin (talk) 23:27, 30 December 2007 (UTC)
There's plenty of sciprs already available that issue warnings after using admin rollback - I use one and it's fantastic. Twinkle will probably be extended. Ryan Postlethwaite 23:34, 30 December 2007 (UTC)
When a rollback is performed (on Commons anyway) the confirmation page that loads includes a summary of the edit made, including a link to the talk page of the user you reverted. Shouldn't make it too hard to remember to dish out a warning that way... Dihydrogen Monoxide 05:02, 31 December 2007 (UTC)

Quick feedback

I generally like the quick, no-nonsense approach to an approvals system -- this is not and should never resemble anything like a quasi-RfA, and I'd prefer to avoid any system that makes it a big deal. Just a few things to consider...

  • Keep a centralized request page (as with WP:AWB or WP:AIV) or make requests via user talk templates (as with {{helpme}} or {{prod}})?
  • Is the edit count suggested too high? Bearing in mind that this is not another RfA. Is there a better measure of activity that is both quick and relatively objective?
  • Is three months too long? That seems to be getting into quasi-RfA territory.
  • "No history of edit warring in six months" sounds very difficult to check. Perhaps something more akin to "no history of frequent or excessive edit warring"? Block log and user talk history might be good indicators.
  • "No blocks for edit warring in the past year"; a year seems to be a very hefty penalty, what about seeking support from multiple users in these cases? Seems to balance the risk without being too draconian.
  • I still somewhat like the idea of giving rollback to everyone, throttled, and then selectively releasing the throttle on request; but it may be late for such suggestions.

Thoughts, anyone? – Luna Santin (talk) 23:25, 30 December 2007 (UTC)

  • Ah yes, there would be a centralised request page, basically taking the format; {{Userlinks|USERNAME}} - brief statement as to why the user wants the tool. ~~~~ followed by {{Done}} or {{Not done}} with an explanation from the admin. With respect to the last point, I guess there's always going to be differnces in opinion! Ryan Postlethwaite 00:03, 31 December 2007 (UTC)

Couple Questions

  • Will there be clear usage guidelines? I'd expect a lot of traffic and new usage, I think there should be some pretty clear standards as to when usage is exceptable, obvious vandalism for example.
  • Do people expect a large rush of applicants up front? Are there plans to handle the initial burst (or to limit it to some extent)?

Thanks, RxS (talk) 23:31, 30 December 2007 (UTC)

    • Obviously as part of the implementation, users would have to agree to use the tool responsibly, and only for vandalism reversion - misuse will be dealt with via removal of the tool. There will most probably be a large number of users requesting it at first, but we're lucky enough to have plenty of admins on hand to deal with these if/when it's implemented. Ryan Postlethwaite 23:36, 30 December 2007 (UTC)
(ec) Hm, an opening rush is inevitable, true; hopefully that rush will mostly be composed of experienced users we can generally trust with rollback already. As far as usage guides, such things will probably evolve over time no matter what initial guidelines are put down -- for example, the use of admin rollback in a content dispute is often frowned upon, currently, in large part because of the lack of any explanatory summary. – Luna Santin (talk) 23:38, 30 December 2007 (UTC)
I added something that deals with limiting it to obvious vandalism. I think allowing usage outside of that might lead to more drama then it's worth, however the usage guidelines end up they should be pretty explicit. RxS (talk) 23:59, 30 December 2007 (UTC)
Agreed, there are few uses outside of vandalism reversions and those should probably be handled by admins. Mr.Z-man 03:28, 31 December 2007 (UTC)
I have added use of rollback to revert your own edits, which is a perfectly legitimate use and one which there is no reason to prohibit – Gurch 10:50, 31 December 2007 (UTC)

←I don't see why only admins should be handling vandalism rollbacks. I think ordinary users can be trusted with it. However, to prevent the "opening rush", the feature should be implemented a) only for auto-confirmed users, and b) not by default, but by a user preferences option, so that not all users would suddenly see this rollback link and perhaps click it out of curiosity. Users should need to explicitly know about the feature beforehand and enable it themselves. Equazcion /C 08:49, 4 Jan 2008 (UTC)

Sorry for being a noob ...

...but how is this rollback function different from the one I already have with twinkle ? If hypothetically I were a huge retard who wanted to revert all of someones recent contributions, what would stop me from just going through their contribs and clicking all the [vandalism] tabs? I don't understand why administrators would want to grant the right to a user, or remove the right either. If someone was really being so abusive why wouldn't the admins just block him ? I really don't understand this, how is the rollback function any worse than just reverting ? People are concerned about edit waring but after 4 reverts you are blocked anyway, so a rollback function wouldn't save much time over only 4 edits. As for people using the rollback function for vandalism just block them like any other vandal. In my opinion it could just be given to auto confirmed users automatically. And I don't understand the use when fighting vandalism, as it says on the page the edit summary is automatic, but normally wouldn't you want to give a reason in the edit summary such as "reverting vandalism" ? The rollback system doesn't sound remotely as good as twinkle for dealing with vandalism. Examples of when it would be used might help thanksJackaranga (talk) 16:20, 31 December 2007 (UTC)

It's true that anyone can install a script such as twinkle an mass revert. The point about rollback is that it's much quicker to revert with as it completely misses the editing screen out (unlike twinkle which edits the version your reverting to and then saves it). It essentially does the same job as a twinkle revert, or undo for that matter, but it does it much more efficiently. Yes it could mean quicker edit wars, but that's not the major issue here. Rollback gives an automatic edit summary such as (Reverted edits by 81.105.251.203 (talk) to last version by HBC AIV helperbot3) - this is the problem with it being used in edit wars, it gives no explanation of the revert and instead implies bad faith meaning that the dispute will be excalated. If someone currently misuses a script, we can remove it, that's why we need to be able to grant and remove this one. In a nutshell, it's a more powerful method of reverting a user (it's quicker) and in the wrong hands, could be used to mass revert and cause a lot of damage so it has to be given sparingly. Ryan Postlethwaite 16:28, 31 December 2007 (UTC)

Rollback vs Undo

Someone please explain to me what the major difference is between the undo button currently availble for non-admins and the rollback feature currently restricted to admins. As I understand this proposal their is no major difference between the former and the latter, so I would like to be enlightened on the differnce before putting my two cents in one way or the other. TomStar81 (Talk) 10:36, 4 January 2008 (UTC)

With undo, you can only undo a single edit. With rollback, you can undo multiple consecutive edits by the same user. For instance, if a vandal made multiple edits that damaged an article, you could fix all of them with a single action, rather than having to undo each of them one-by-one. Equazcion /C 10:39, 4 Jan 2008 (UTC)
In addition to what Equazcion said, when you click rollback it doesn't show what you are changing. This is because it is a server action which makes it faster. Also because it is one server action, instead of multiple (a page request, an edit page request, save the page) it reduces server load greatly. One reason people are less willing to hand it out is that you can rollback a person's edits from their contributions page. Next to each edit of theirs is a rollback link so a persons edits can be reverted very fast (one click each). James086Talk | Email 10:44, 4 January 2008 (UTC)
Further, rollback only works on the most recent edit to a page while undo can revert an edit from earlier in the history. violet/riga (t) 10:49, 4 January 2008 (UTC)
And further to James' point, you not only don't see what you are rolling back to check it, you don't get to add an edit summary, just some pro forma "reverted edits by X" stuff. This is a comparison that seems to have been lost here... perhaps it should be added to that arguments section (which, incidentally, in its current form looks pretty POV to me). --bainer (talk) 12:52, 4 January 2008 (UTC)
I've added the edit summary point to the arguments list (with no answer, since I don't have one). Equazcion /C 12:59, 4 Jan 2008 (UTC)
I agree it is toxically POV, so I added some balancing material. Equazcion didn't like that though, and he removed it. Apparently, I don't have a license for that section of the page, and only those who unabashedly support the proposal may add things. I'm only allowed to write things where I entered the poll apparently, even though my comments'd be totally out-of-context there. Splash - tk 13:29, 4 January 2008 (UTC)

← For those interested, please see User talk:Splash#Argument responses. Despite the strawman argument I can assure you I have problem with opposing statements -- I only have issue with prolonged exchanges taking place in a section that's meant to list the proposal rationale. There is already discussion taking place elsewhere, and people who oppose the listed arguments say so in their votes. There is no need to make this particular section into another discussion. Equazcion /C 13:36, 4 Jan 2008 (UTC)

(Suggest you change the above to "I have no problem with opposing statements". If you do, feel free to delete this post as well) Scolaire (talk) 14:52, 4 January 2008 (UTC)
Then remove the entire section, lock, stock and barrel. But don't remove only the critical portions and expect that to go down as mere housekeeping. It's not the rationale, by the way, because that appears adequately in the 'lead' of the page. Splash - tk 13:39, 4 January 2008 (UTC)
I have already replied to this insinuation on your talk page. Despite your assumptions, my motivation has nothing to do with silencing the opposing viewpoint. It is the rationale as well, in addition to the lead. As in many instances of stating rationale, the author has predicted the arguments against him and supplied preemptive responses. Just because those have been divided into their own section does not change the intent with which they were posted. Equazcion /C 13:43, 4 Jan 2008 (UTC)

In regards to the edit summary question above, I posted a section here on this proposal. I don't see any valid reason why any user shoulnd't be required to add a short summary on using rollback, admins included, to explain their usage of the tool. Lawrence Cohen 18:08, 4 January 2008 (UTC)

So, to sum up: the major differnce is between rollback and undo is that rollback can tackle mutiple edits at the cost of providing a custom edit summary. Is this correct? TomStar81 (Talk) 21:27, 4 January 2008 (UTC)
Pretty much. The main difference is that rollback does roll back multiple edits without the user in question having to check; if I rollback a vandalous edit, all the edits in a row to that article by the same user will be reverted, and I don't have to check or even know about that. --ais523 21:32, 4 January 2008 (UTC)

Argument responses

This section is being created in response to a desire by some users to respond to the arguments made in the "Arguments" section of the proposal page. Rather than have that section become a discussion, I've moved the posted responses, as well as the original arguments, to this talk page. Please feel free to respond to any point. I will also post a link to this discussion in the arguments section.

NOTE: Signatures are welcome and encouraged in this section even though none were present to begin with.

  • Other rollbacks exist, like TW.
    True, but rollback is easier on the servers and the clients.
    Marginal improvements to server load are not something to be worried about. This would reduce the number of page requests clients (editors) have to make.
    Also there is no reason why users should have to use external tools, which may pose a security risk and which are browser dependent.
    Finally this is a poor argument for why this tool should not be granted, if anything it suggests all the claims about problems this tool will create are wrong
  • If someone can be trusted, they should be an admin.
    Yes, but not everyone seems to think so. I know of several vandal fighters who tried to become administrators but failed due to "lack of mainspace edits". Also, I doubt that ClueBot would pass an RfA.
    Are you extending this proposal to include the dispensing of rollback to bots as well as editors who request it?
  • It would encourage edit wars.
    No, an admin would have to assign the privilege only if the user was trusted not to be disruptive and if it was misused, it would be removed.
    This in no way stops people from starting high-speed edit wars in the first place, where at present the tools to initiate at high-speed are restricted to admins. Edit wars are already stopped by blocking.
  • If you do this, might as well give it to everyone.
    It was going to, but there was significant opposition to that. Opposing for reasons like this makes it harder to get anything changed.
    Not all compromises are acceptable to both parties. This proposal adds significant conditions that may persuade some of those otherwise opposed to any kind of roll-back feature being granted to non-admins, but also adds bureaucracy, creates new divisions of editor, and makes a useful feature invisible to the majority of users who'd benefit from it.
    This proposal does not add "bureaucracy" in the ordinary meaning. It simply enables admins to delegate the use of a tool, a relatively harmless one (any use can undo the damage, any admin can revoke the privilege *without having to go through any specific process*. This kind of "bureaucracy," with totally distributed implementation requiring little or no discussion, is exactly what we need to manage the increasing scale of Wikipedia. —Preceding unsigned comment added by Abd (talkcontribs) 17:24, 4 January 2008 (UTC)
    If one has to go through an application process in order to be granted use of a tool, then that person is being burdened with a bureaucratic procedure. If an admin has to evaluate an application before granting use of that tool, then the admin is being burdened with a bureaucratic procedure. In no real sense is this not "bureaucracy" in the ordinary meaning.
  • Administrators should not be allowed to set privileges.
    They can grant and revoke editing rights. Why not something less significant? Administrators who are not trusted to do what is good for the encyclopedia should not be administrators.
    Incorrect - they can suspend editing rights to most, but not all, pages. This usually comes with a time limit that expires naturally. It is misleading to use the grant/revoke terminology. All rollback removals would be indefinite and apply to all pages everywhere. The last sentence of the 'reply' sounds like off-topic rhetoric.
It's actually quite true. The proposal effectively allows any admin to grant or revoke a privilege useful *only* (by the restrictions) for removing vandalism, lessening their load, the load on the servers, and the load on users who want to help. As admins are explicitly responsible for granting this privilege with caution, and likewise for revoking it (and any admin can revoke if abuse or even mere error is seen), this is quite safe and creates no bureaucracy other than one under individual administrative control, and only admins who care to take on the responsibility of making good choices in granting the bit need do so. "All rollback removals would be indefinite"? That's right. They would last until an admin -- any admin -- decides to grant it again. The removal of a right to help with vandalism is not punishment, in any way. The removal of a right to use the tool for edit warring would be automatic, if we could make it so. If an admin makes biased decisions on this, it is like any other abusive administrative action, but, among such, this is way less likely to cause conflict among administrators than other admin privileges.
  • It will introduce wheel wars.
    Any administrator action can introduce a wheel war. We trust the administrators not to wheel war, though.
    Yes, and this provides several new battlegrounds. Doesn't matter that 'we' trust admins not to wheel war, since the fact is they do. This will result in more wheel warring than at present.
    But not often enough to worry. If our admins are really that stupid we have a very big problem that we need to address and it would be far better if we are aware of that. —Preceding unsigned comment added by Nil Einne (talkcontribs) 18:24, 4 January 2008 (UTC)
  • Adds more bureaucracy.
    Not much more. It will be a simple page like WP:RFPP, not something as complex as WP:RFA, as that would defeat the purpose of this.
    It adds more bureaucracy and the argument above doesn't address that, indeed it posits why this solution is completely unacceptable. The added bureaucracy does defeat the purpose of this. So why do it if even its advocates know that, as proposed, the system will fail?
    This introduces no new likely battleground, in fact. Were I an admin, I would not risk my admin bit wheel-warring over granting or revoking a rollback bit. Quite simply, it is not that important, and I'd instead, if serious conflict developed, which seems quite unlikely with any but the most unbalanced admin, use standard dispute resolution if simple discussion doesn't cut it. Rollback is to be used only for removing vandalism, and, while I've seen abusive users call legitimate edits "vandalism," it is rare among administrators. Mistakes will be made, but they will also be easily undone. This introduces *no* "bureaucracy."
    Yes, it adds more bureaucracy, and nobody has argued it adds a "battleground" in the context of this particular issue. Here's the reality: in order to use the feature, a user has to go through an application process which may or may not be successful. The reality is that few people who want a feature designed to make it easier to undo vandalism are going to bother; indeed the most likely affect, as with me, is for us to say "F that, you say you want our help undoing vandalism, but you then purposely give us the choice between either asking for permission or using the wrong tools for the job. Why should I want to help *at all*?" Honestly, if this proposal succeeds, I, for one, am not going to bother reverting vandalistic edits except in the simplest, easiest, occasions, and then only over articles I care about. This proposal is a slap in the face. You have a tool, and you're going to deign to give it to me if I plead for it? WTH?
  • Bureaucrats should be the only ones doing this.
    We don't have enough bureaucrats to do this, and administrators should be trusted anyway. Administrators already can revoke and grant the editing privilege.
    See previous answers re admins, trust and grant/revoke terminology. It is true that we do not have enough 'crats to deal with this proposal as it stands.
  • Users shouldn't get administrator tools.
    Rollback is not really an administrator tool, it is just an editing tool that was originally restricted to administrators for technical reasons, but is no more powerful than the Twinkle rollback, but is less stressful on the server and the browser.
  • Too much form-filling.
    Isn't that what administrators do?
    If that's the best answer to this charge, then well, really. Admins do backlog clearing, tidying up, vandal control etc. This is not all merely form filling. Deliberately adding a new set of forms to the list is unnecessary.
Admins are not required to grant this privilege. So this adds no administrative burden except one voluntarily assumed. —Preceding unsigned comment added by Abd (talkcontribs) 17:44, 4 January 2008 (UTC)
  • Rollbacks don't allow custom edit summaries.
    True, but rollback is intend for obvious vandalism, which doesn't need a custom summary.
    Also, yes they do. It's a bit inconvenient for a human to use, but a script or bot could easily use its own edit summary while still getting the performance benefit of rollback.
  • The case for requiring editors plead for this feature and be granted permission has not been made, it appears to have been added merely as a sop to those opposing offering the feature to non-admins. The concerns expressed by the latter group of people do not appear to be addressed, and when clearly wrong (such as the belief that the tool can be abused in a way "Undo" cannot), are, if anything, given unneeded credibility. —Preceding unsigned comment added by Squiggleslash (talkcontribs) 18:42, 4 January 2008 (UTC)

I'm missing something

Discussions are taking place about Section 1, Section 2, Section 3....er...where are these Sections for me to read? Kingturtle (talk) 14:37, 4 January 2008 (UTC)

I'm not really sure, I think it's so each section doesn't get massive. Ryan Postlethwaite 14:53, 4 January 2008 (UTC)
There seem to be 12 sections that people are referring to. Where are those 12 sections for me to read? Kingturtle (talk) 14:56, 4 January 2008 (UTC)
On the main page of this proposal, each section is broken down into 10 comments. Ryan Postlethwaite 14:57, 4 January 2008 (UTC)
The sections are merely divisions of the long list to make editing easier. There is only one proposal, at the top of the page. Equazcion /C 15:01, 4 Jan 2008 (UTC)

The new sections make it much harder to read....it looks like a mess and has no real use, anyone mind if I remove them? RxS (talk) 15:16, 4 January 2008 (UTC)

Yes, I mind. The section is just too large without the breaks. It would be fine for reading-only, but as soon as anyone needs to edit, it becomes very difficult. Equazcion /C 15:19, 4 Jan 2008 (UTC)
Then make the sections bigger, or just section off the most recent comments or something, the editing activity in the first 12 support sections won't be high enough to justify making such a mess. And in fact after a day or 2 editing on the whole page will have dropped off enough not to justify it. Such little sections make a mash of the whole thing. RxS (talk) 15:38, 4 January 2008 (UTC)
Done, 20 per section now. I hope that's sufficient. Equazcion /C 15:45, 4 Jan 2008 (UTC)
What exactly was wrong with it the way it was? Not sure I follow the logic of making this change. --Kbdank71 15:48, 4 January 2008 (UTC)
I'm not sure either, but apparently it was more difficult to read? I disagree but I don't mind either way. Equazcion /C 15:49, 4 Jan 2008 (UTC)
RFA seems to do fine with the same structure of support/oppose, who's having trouble here? --Kbdank71 16:21, 4 January 2008 (UTC)
I don't think anyone has complained about that, at least not here. The main discussion is still following that structure. The issue brought up here was the division of the vote lists into sections (section 1, section 2, etc). Equazcion /C 16:24, 4 Jan 2008 (UTC)
Perhaps I should restate my initial comment. What was wrong with the way it was before it was chunked into sections? In other words, I agree with RxS. --Kbdank71 16:44, 4 January 2008 (UTC)
Ah. It's too hard to edit such a large section. There appears to be consensus for the sections though, as many users have been editing them for accuracy etc. The latest change was that someone made the section names into a running total. Equazcion /C 16:51, 4 Jan 2008 (UTC)
I guess I see that activity as people trying to make sense of a bad formatting scheme. RxS (talk) 16:57, 4 January 2008 (UTC)
I guess you do. The people editing the headers could just as easily remove them. That's the way I see it. Equazcion /C 17:03, 4 Jan 2008 (UTC)
This is a very active project, edits are constant. Without sections, edit conflicts become far more common, making it more work to comment. Bad idea to remove them. —Preceding unsigned comment added by Abd (talkcontribs) 17:47, 4 January 2008 (UTC)
I'm not against sections (don't really care), but I believe you are wrong that they prevent edit conflicts. There is a proposal to request a change through Bugzilla but it is not implemented. SpinningSpark 10:16, 5 January 2008 (UTC)
He's not wrong. Edits to different sections don't produce edit conflicts, so splitting up large sections will reduce conflicts. The proposal only has to do with the way edit conflicts are handled once they occur (I made the proposal you're referring to). Equazcion /C 10:24, 5 Jan 2008 (UTC)

Control panel toolbar for a rollbacker

Help requests

  • AN alert
  • ANI alert
  • AIV alert
  • UAA alert
  • CVU alert
  • RfR alert
  • WikiDefcon update
What is this? Equazcion /C 15:35, 5 Jan 2008 (UTC)
I don't know what these are, perhaps you have a javascript installed that creates these? 1 != 2 17:45, 5 January 2008 (UTC)
I didn't post this, I'm just asking what it is. I believe User:Igorberger did, as a proposal of some kind. Equazcion /C 17:52, 5 Jan 2008 (UTC)

Question

When I use Twinkle's rollback function, I never roll back from the contribs list. I usually look at a diff and do the rollback from there. I can see how users could abuse a rollback function accessible from the contribs or history list -- users engaged in an edit war, or users who don't think much of another editor, could blindly roll back the edits of a particular user, just because it was that user who made the edit. Is there any way this rollback functionality could be changed for non-admins, so that users must at least invoke them from a diff screen, and not a contribs or history list? Equazcion /C 16:35, 4 Jan 2008 (UTC)

It would be much easier to just revoke rollback access from any person that was not responsible with the tool. 1 != 2 18:00, 4 January 2008 (UTC)
I don't know about you but if I find an IP whose contribs are vandalism I usually check out and manually revert each edit. I don't use twinkle and don't see why I should have to. It wouldn't definitely be somewhat faster if I could just check out each edit and then roll back them all. Similarly if I see two or three edits to a page by an anon or user I sometimes check the general diff of the user and then if it's vandalism either have to manually undo them all (leaving 2-3 edit histories) or I have to manually edit the whole old pre vandalism page, and save which is slow and also means I may accidentally revert new edits. As 1 == 2 says, we already have a system in place to revoke access. In the unlikely event we really have a large number of people who have seemed trustworthy enough to now that admins are likely to give them the function yet will suddenly start doing stupid things and not stop when they're told what they're doing is wrong so that we can't just revoke the occasional problem user then perhaps we will have to reconsider but I see no point getting too concerned about a hypothetical situation which probably won't happen Nil Einne (talk) 18:17, 4 January 2008 (UTC)
It's good that you check first, but not everyone's gonna do that. If we had some kind of assurance that they would check the edits first then I'd be all for it -- but the easiest way to force people to check edits first would be to eliminate the tool from contribs lists. Equazcion /C 18:26, 4 Jan 2008 (UTC)
The point here though, as I understand it, it's intended for people who are trusted. As I've said in a number of places, I don't really see it that likely we have that many people who are trusted yet are going to do something stupid and not learn when warned that we can't just take it away from the odd case that crops up. If we do have so many people, it's probably better we know isn't it, it's quite a disturbing thought. Besides that, the problem is eliminating it from contribs will slow things down. You can't just check a bunch of edits then revert them all with several clicks. You don't have to check, revert, check revert etc. Also, one issue I didn't really mention. Surely there are cases you accept when it's probably unresonable to expect a user to check every edit. It's not that common. But if for example 'IAMATROLL' made 20 edits in the past hour and the first 5 are 'BUSH IS GAY' are you really telling me you think someone should check every edit to make sure none of them were useful? Yes there are some people that lack common sense and will do this in non-obvious cases but again, we come back to the point will these people really be given roll-back in the first place and if they are, will there really be so many we can't just take it away without too much fuss when it happens? Nil Einne (talk) 19:02, 4 January 2008 (UTC)
Trusted in this case means that you've been here a certain amount of time and one admin trusts you based on a cursory look at your editing history. I think people who haven't necessarily vandalized, stalked, or warred before can still start doing that in the future. It doesn't necessarily have anything to do with bad faith edits, but users can develop POV about other editors and decide to revert them simply based on who they are, and without looking at the edit. I think this is likely to happen and when it does it will be irritating. Forcing people to view edits before rolling them back is really not an unreasonable request. Rollback would mean that multiple consecutive edits would be taken care of in one shot, though, so only the most recent edit would need to be viewed. As for your example, something like that is very rare. I use Twinkle myself right now and I haven't found it all that tedious to check the latest diff of an article prior to rolling back, even in cases of blatant vandalism across multiple articles. Equazcion /C 19:07, 4 Jan 2008 (UTC)
I'm not saying it will never happen and that when it will, it won't be irritating. Indeed I'm sure there will be the odd problem. I'm just not convinced it will happen often enough to be a significant problem we can't just deal with such editors by removing the tool. If anything, as others have mentioned this may help us weed out such editors before they become admins. I'm also somewhat confused. If an editor is really bad enough to revert people they don't like, I highly doubt forcing them to view the edit is going to change their mind. The only think forcing editors to view an edit will solve IMHO is editors who are simply careless, stupid or lazy (in other words people who are not malicious but simply don't really bother to use their brain) and such editors tend to cause sufficient problems that an admin will notice before granting the tool. And as I said, your proposal will still require an editor to review and then revert every edit (or edit group) which significantly limits its usefulness rather then being able to review all edits and then revert them quickly. It also won't help in the second example I mentioned Nil Einne (talk) 19:17, 4 January 2008 (UTC)
It's not a matter of changing their mind. It's a matter of ease. If I decide to stalk a user and I have rollback, I can revert his edits from his contribs list almost instantly as he makes them. How long will the process take to get the rollback privilege revoked? How long will this poor stalking victim have to have every single edit undone instantly? Will he even be able to make the revocation request without being instantly reverted? An ill-meaning person could really make someone's life miserable with this tool, and have a good amount of time to be able to do it before he's caught. This sounds like a scary idea to me. Again, take it out of the contribs list, and boom, I'm for it. Bur this way, I would actually be afraid to piss someone off who had that rollback function and wasn't an admin (I'm afraid of pissing off admins as it is, but even more so, someone who can rollback from contribs but that hasn't been through RfA). Equazcion /C 19:24, 4 Jan 2008 (UTC)
But if your suggestion is going to make it a lot harder to revert someone without good reason then by definition it's going to make it a lot harder to revert someone with good reason. And therein lies the problem. I'm sure that there will be a few, very few I hope users as you describe that will get through. But I strongly suspect they will be noticed soon enough and probably not just lose the tool but be indefinitely blocked. And really I don't see any reason for people to be afraid. Yes idiots exist. Sometimes they are smart enough to hide their idiocy for a while. But in the end, we have to strike the balance between taking away highly useful tools because an odd idiot may get hold of it and do a bit of damage and giving useful tools which will help other users make wikipedia better. Also, I highly doubt it'll make a big difference from the undo function anyway (which is available to admins) for the case you mention. The big difference with this is that it enables a user to kill all your edits from one or several days with one or a few clicks. I.E/ a user could check you our every few days and kill all your work mightly quick. That's a potential problem but they won't be too hard to revert and such a user is going to be banned mightly quick since they're so easy to spot. If a user is constantly watching you and reverting your complaints they can easily do that with undo, the tiny 10-20 second difference is hardly going to make a difference. It's not as if you right a complaint in multiple edits. And personally, I would NEVER mind pissing off someone if you had a good reason. If someone is being unhelpful or bad I will try to discuss and failing that I will take it the the various avenues available. If I piss off someone when I wasn't doing anything wrong and they decide to be a fuckhead I will welcome the indefinate ban placed on this user since they are clearly unsuitable for wikipedia. Nil Einne (talk) 19:44, 4 January 2008 (UTC)

Edit wars: prevention

At Wikipedia talk:Rollback for non-administrators#Some suggestions, I made a suggestion for a way to avoid edit wars. Several others had suggested limiting how often a non-administrator could use rollback; I suggested looking at it from the point of view of the article: "Limit rollbacks on an article. For example, two in a row, or two in an hour. Limiting rollbacks need not limit ordinary editing (including reverting). Also, make rollback subject to the three-revert rule just as a revert is." Would this be compatible with the present proposal? Fg2 (talk) 12:17, 4 January 2008 (UTC)

How about leaving an automated message on an roleback receiving editor's talk page telling them that their edit has been rolledback by editor giving a rollback and if they disagree with the rollback please leave a message on non administration rollback notice board and send an alert to the rollback User notifying him or her that their implemented roolback is beeing disagreed. They do not have to answer it but statistics can be kept and used for evaluation to revoke the rollback tool from an editor if their roolback authority is problematic. So this will be anti exploit review mechanism. Igor Berger (talk) 12:31, 4 January 2008 (UTC)
Well, I am already feeling a little uneasy about the proposal. One RC Patroler has been stalking me with mindless tagging (semi-speedy tags are the patroler's particular favorite), making abusive claims unwarrantedly dropping into discussions, and voting against whatever I vote for on AfDs... but, this clever editor "generally" manages to keep the stalking part under 10% of edits, thus hovering on the borderline and evading detection by outside parties. I have learned to live with this much, but "if" the person can lay a hand on the rollback stuff... I can only shudder at the thought. Aditya(talkcontribs) 14:43, 4 January 2008 (UTC)
This editor could be a sockpuppet or involved in social engineering on behalf of a third party. You may want to bring it to a few admin's attention and see how things fan out. Igor Berger (talk) 14:48, 4 January 2008 (UTC)


Effects on RfA

So how long before this feature starts being used as a yardstick for RfA? Support: User has had rollback for 6 months with no problems. Oppose: User yet to be trusted with rollback, can't be trusted with tools. Burzmali (talk) 17:55, 4 January 2008 (UTC)

Are you suggesting that would be a bad thing to happen? And if so why? Equazcion /C 17:59, 4 Jan 2008 (UTC)
Judging a users at RfA based on how well they made admin-like decisions in the past? I sure hope that happens. 1 != 2 18:01, 4 January 2008 (UTC)
Same here. I'd even say this should be listed as a possible benefit of the proposal. Equazcion /C 18:22, 4 Jan 2008 (UTC)
Good or bad, it is one more hoop that a potential admin has to jump through to get the mop. Not to mention that it is a hoop held by existing admins. Burzmali (talk) 19:39, 4 January 2008 (UTC)
I'm not sure it will become such a major standard in RfA's. After all, rollback is by no means the most significant sysop tool; I think we can all agree that blocking, deleting, and protecting are far more important tools for sysops than the ability to revert more efficiently, especially with tools such as Twinkle, Popups, and Godmode-light openly available. Pyrospirit (talk · contribs) 20:05, 4 January 2008 (UTC)
I'd say the "most important" admin tools are the ability to move a page over a nontrivial redirect and the ability to perform history moves. Admin tools are editing tools, and having to ask for the former can be annoying. The latter, on the other hand, is the only thing an admin can do that's non-trivial to undo. Burzmali makes a good point - too many cliques already. Guettarda (talk) 20:09, 4 January 2008 (UTC)
If someone wants my support in RfA then they need to have a history clear of edit warring, it does not matter which tool they used. Frankly when thinking about how a user will use their tools, rollback is on the bottom of my list. It is the only admin tool that a regular user can undo. 1 != 2 20:44, 4 January 2008 (UTC)
That fact that a user has been trusted with the tool is the important point, the tool itself is irrelevant. RfA already has enough hoops for admins to hop through. Burzmali (talk) 22:46, 4 January 2008 (UTC)
That's a bit of a cynical way to look at it. This would be an easier way for admin hopefuls to prove they can be responsible. Calling it a "hoop that they have to jump through" doesn't make it a bad thing, for anyone. Genuine admin hopefuls will see this as nothing but a benefit -- and if they see it as a burden, then they definitely should not be given the complete set of admin tools. Equazcion /C 06:10, 5 Jan 2008 (UTC)
  1. Support Vox Rationis said, "The feature should only be given to editors who are very experienced with vandalism removal, recent changes patrol, and are decently well-read in Wikipedia policies." While I totally agree with his statement, there needs to be an oversite. Igor Berger (talk) 02:40, 5 January 2008 (UTC)
If they are that helpful and experienced, why not simply make them admins? -- The Anome (talk) 12:54, 5 January 2008 (UTC)
Because maybe some editors do not wish to be admins? I can think of many editors who would responsibly use that tool (perhaps even need the feature) who definitely do not want to be admins. .:Alex:. 13:01, 5 January 2008 (UTC)
Exactly right - we are people - we are not resources to be turned on and off. Saga City (talk) 13:08, 5 January 2008 (UTC)
Sorry this is the wrong section for your talk. This is the proposal for committie to oversee the rollback administration. The committie is to be composed of admins. Igor Berger (talk) 15:25, 5 January 2008 (UTC)
I really have no idea what this is a proposal for exactly, however I just wanted to reply to Anome -- Are you saying that everyone who's experienced and helpful should be an admin? Because I think the criteria is a lot more stringent than that. Equazcion /C 15:31, 5 Jan 2008 (UTC)
Equazcion, that's the idea - RfR! This is to prevent abuse and educate the rollbackers about the program. Igor Berger (talk) 15:53, 5 January 2008 (UTC)
Well I think that's something that would get created by default if the proposal is implemented. There would need to be a place where editors request the rollback function, and such a place would likely contain documentation, rules and advice. I don't think there's any need to propose that separately. Equazcion /C 16:16, 5 Jan 2008 (UTC)
Exactly. But I see why this was requested. The proposal up top says that requests need to be put on this page, but I believe that having a page that explains Non-administrator rollback and a seperate requests page would be a better idea.—Preceding unsigned comment added by .:Alex:. (talkcontribs)

Admin workload?

I'm sitting on the fence here. I realize that tools like TW can do a similar feature. The proposal here says that the non-admin rollback will be used only for vandalism (and self edits), but there is no mechanism to prevent other uses. Discussions above talk about edit wars, but I am wondering what kind of load this will place on admins? I am assuming usage of the tool is community-enforced; if an editor sees someone else abusing this tool, they'll be reported, and they'll be revoked. Where would we report abuse? ANI, which is already very busy. AIV, even tho it is not really vandalism? Admins appear to be under a constant backload to begin with, how much more would this load them? Yngvarr 19:31, 4 January 2008 (UTC)

As an admin I can say it is no problem at all. The backlogs are specific to certain areas and are not all encompassing. 1 != 2 19:34, 4 January 2008 (UTC)
As full as ANI is, reports made there are generally responded to within minutes. If its an absolute, seconds-will-count emergency, find an active admin and use their talk page (check the history of ANI or AIV) or ping one on IRC with !admin. Mr.Z-man 20:34, 4 January 2008 (UTC)

illegitimate gauge of consensus

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.


Time constraints on the poll have been removed as requested, so this is largely a moot thread.--Docg 11:14, 5 January 2008 (UTC)

I'd strongly advise any developer against proceeding on the basis of this poll, regardless of the result. We polled on this 2 years ago at Wikipedia:Requests for rollback privileges/Poll and got no consensus. Now, admittedly consensus can change, and maybe it has. But that poll lasted *6 months* and involved nearly 300 users, this one is scheduled to run six days over the holiday period, and despite the fact the community is far larger, attract a fraction of the involvement (indeed I only stumbled on it by accident). To suggest that the no consensus position coulb be overturned on that basis will be quite invalid.--Docg 02:38, 4 January 2008 (UTC)

From what I understand, this is a rough vote to check which are the weak points of the proposal, and work on them. Also, there are two main differences: the rollback privilege exists now (and it was not something hypothetical like in that poll), and the process to grant it is different (poll vs. direct granting), which was ultimately one of the main negative points in that proposal (as you see, a solution for one of the points there was offered here, and probably in a third poll there will be another for the "misuse" argument). -- ReyBrujo (talk) 02:54, 4 January 2008 (UTC)
I'm not sure about that, the way I read the bolded text at the top of the page it's measuring consensus for implementation. I'm surprised at the low participation here to this point. I'd have to agree with Doc, without more eyeballs this isn't an accurate gauge. I'm on the fence about this, but perhaps the addition of more rollback type tools available since the last debate makes this less of an issue and maybe unnecessary. RxS (talk) 03:14, 4 January 2008 (UTC)
Anyone is free to publicize this wherever they'd like, however it's been on WP:AN, WP:BN, Template:Cent, and WP:VPR. I'm not really sure how to respond to claims that it hasn't been well-publicized. --MZMcBride (talk) 03:17, 4 January 2008 (UTC)
I don't think the issue is how well-publicized the page has been, but the low level of participation that resulted. A change like this probably needs more input...RxS (talk) 03:23, 4 January 2008 (UTC)
Agreed. The previous "no-consensus" position was arrived at with 400 participants. Now, that may well have changed. But showing 49 people supporting it in 6 (holi)days does not demonstrate this.--Docg 03:24, 4 January 2008 (UTC)
(edit conflict) For one thing on how it hasn't been well publicized, you started the poll in the middle of the holiday. How could you do that and expect reasonable awareness and publicity and participation? This is almost a poster child for when not to do things... Georgewilliamherbert (talk) 03:27, 4 January 2008 (UTC)
This isn't a job - most people will edit more in the holiday period, they're off work and college. It's very very well publicired, where else do you want us to mention it? Ryan Postlethwaite 03:31, 4 January 2008 (UTC)
Many people are away though, so why not extend it for a couple of weeks so those who are can opine? The low level of interest here speaks for itself and robs it of legitimacy. If you really want this to happen, it is in your interests to show strong support and not indifference anyway.--Docg 03:33, 4 January 2008 (UTC)
But where else do you want it publicised to encourage more people to comment? Ryan Postlethwaite 03:34, 4 January 2008 (UTC)
The main things is I want you to give it longer, so more people have the opportunity to see the publicity there is. Personally, I only check policy goings on every couple of weeks or so.--Docg 03:37, 4 January 2008 (UTC)
Exactly. I don't think a notice was posted to the mailing list either. It's the only policy-related part of Wikipedia I frequent anymore, so I certainly wasn't aware of this proposal. My opposition to it can be struck off if we extend the deadline by about two weeks, preferably more. Johnleemk | Talk 03:39, 4 January 2008 (UTC)

Well, it's on the mailing list now, anyway. Can anyone think of anywhere else it hasn't been posted that it should? A sitenotice, perhaps? —Random832 04:36, 4 January 2008 (UTC)

Well, actually I would suggest that there be a watchlist banner similar to what was done with WP:ATT last year. That attracted a lot of participation and carefully thought out responses from the broad spectrum of Wikipedia editors. It could go up right after the fundraising banner comes down in a couple of days. The poll really needs to be extended, though. A very significant proportion of editors have been editing irregularly over the seasonal break. Risker (talk) 04:53, 4 January 2008 (UTC)
A poll reflects public opinion. Disregarding public opinion altogether is equally as foolish as only regarding public opinion. I think that both the numbers of people and the arguments made support a consensus here. 1 != 2 17:55, 4 January 2008 (UTC)

The time 'limit' was removed and this has now been published in all currently proposed areas (except sitenotice which probably won't get support). If editors voted oppose because of their concerns of insufficient time for consensus they are welcome to reconsider. Nil Einne (talk) 20:01, 4 January 2008 (UTC)

  • Comment Illegitimate gauge of consensus indeed. This is like asking children, "should all of you get candy but mommy can take it away if she feels like it?" If it's not broke, don't try and fix it. -- ALLSTARecho 23:51, 4 January 2008 (UTC)
  • Comment. I really do think you should keep this discussion open for longer. Here in Australia lots of people are off to holiday cottages and so on for the whole of January pretty well. They return just after Australia Day on 28th this year. It is time to leave computers behind. I might be with them if I was not retired and the kids have left home. I can go on holiday at other times. This is too important to rush. Leave the debate open for the rest of the month. --Bduke (talk) 08:40, 5 January 2008 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

random

just some random thoughts, But I agree that admins should have the ability to grant/remove the rollback, But let me toss in another wrinkle that might make things easier, users who have more than 10,000 edits and have been with the project for over 6 months automatically get granted rollback, (by a software config, that already exists) but admins can still remove the auto given right. βcommand 23:15, 30 December 2007 (UTC)
Hmmm, not sure about this, we have a lot of people with over 10,000 edits that really couldn't be trusted with it and would use it soley for edit warring. Ryan Postlethwaite 23:17, 30 December 2007 (UTC)
Then it can be removed from them. Majorly (talk) 23:22, 30 December 2007 (UTC)
Trust and Authority? Igor Berger (talk) 05:21, 4 January 2008 (UTC)

You could pass an RfA with those fixed requirements up there, making this whole thing pointless. Do away with them and let administrators exercise their judgement; they're supposed to be trusted members of the community, not dumb automatons that get spoonfed instructions with no room for discretion – Gurch 23:22, 30 December 2007 (UTC)

I am not sure if someone could pass a RFA, but another month and he could for sure. My fear is that we will be seeing people using the rollback feature without taking the time to warn the user in their page (since rollback should only be used when dealing with vandalism). Who would be assigning the rights? Administrators? I would prefer having bureaucrats do it, as to give them some more work, especially if they will have to review the users' last hundreds of edits. However, I am not against the idea of non-admins using the feature. -- ReyBrujo (talk) 23:36, 30 December 2007 (UTC)
The problem with the bureaucrats granting is that they simply haven't the man time to do this - it would be too much to handle for such a limited resource. We already have plenty of scripts available that allow the use of admin rollback and follow with a warning, so there wouldn't be a great change in that respect. Ryan Postlethwaite 23:40, 30 December 2007 (UTC)
You seem to be worried about people dealing with vandalism incorrectly. That is really neither here nor there. A large proportion of vandalism is already dealt with by non-administrators. This would change only the method by which they do it. If people warn users now, I can't see why they would suddenly stop if they were able to use rollback. If they don't, I can't see any reason why they would suddenly start if they were able to use rollback. So the situation would be no different to how it currently is – Gurch 23:43, 30 December 2007 (UTC)
Automatic tools right now allow to revert and warn at the same time. I am worried that these people would either not use the rollback feature at all (since the scripting solution gives them more than a simple rollback) or migrate to the new system and stop warning users (just like some admins rollback without warning, or users in general undo others without explaining why or leaving a note in the other's talk page). -- ReyBrujo (talk) 23:56, 30 December 2007 (UTC)
There are tools available to do this both for the admin-revert and non-admin-reverts. And it's no problem e. g. to include the admin-revert in Twinkle. --Oxymoron83 00:38, 31 December 2007 (UTC)
Rey, warning is not essential. It's preferable, but not required. It's better the vandalism is removed faster more efficiently. Majorly (talk) 04:31, 31 December 2007 (UTC)
I have my doubts about me being able to pass an RfA. -- ClueBot (talk) 05:05, 31 December 2007 (UTC)

Somewhat related but in an almost opposite tone is Wikipedia:Limit the undo function, a proposal I haven't really organised properly yet. I have concerns about the use of the undo function, but mostly about its use by IPs. violet/riga (t) 23:44, 30 December 2007 (UTC)

I wonder if it would be better to require 2-3 admins to approve granting, rather than one. Ditto on the removal. I would also like to see that if someone has rollback removed for cause, it can not be granted again for some period of time (2-3 months)? Thatcher 06:25, 31 December 2007 (UTC)
Unnecessary bureaucracy. – Gurch 10:06, 31 December 2007 (UTC)
Comment'This is a misunderstanding of "bureaucracy." What this really amounts to is allowing admins to designate assistants, no community load is increased. As Wikipedia grows, this kind of delegation is going to become increasingly necessary, and there are other applications possible. The proposal lessens the load on administrators, it lessens the load on the servers, it lessens the load on users, and it risks practically nothing. "Bureacracy" means the creation of processes requiring increasing community attention or creating barriers to distributed decisions. This is actually the opposite. It does *nothing* but allow admins to recruit help, to enable a trusted class of ordinary users who can more easily combat vandalism. Only admins who want to do this need do it, none are forced or coerced in any way. I'd recommend that admins either watchlist all contribs by an enabled user, or *delegate that to a very trusted user.* In other words, administrators *can* create a bureaucracy of a kind, directly responsible to them, requiring no discussion. Admins properly should be held responsible for their actions, and delegating rollback (or the monitoring of rollback) is quite the same as any other action, it is subject to review, and if done carelessly, an admin could be slapped with a wikitrout, and clear abuse would be, like any other abusive administrator action, subject to loss of the admin bit. I do not expect this to cause much hassle in that way, since use of rollback is explicitly limited, at this time, to vandalism, and, more than occasional errors, which can be easily fixed, not for removing good faith edits. Essential in this understanding is that admins can remove what they can grant; without that, this proposal would be dangerous and would indeed be creating a new "bureaucracy." --Abd (talk) 17:03, 4 January 2008 (UTC)
On the contrary, I think that admins should be allowed to remove the permission without prior WP:ANI discussion. After all, admins can already block without discussion, which is a much sharper sanction. Unnecessary bureaucracy in the implementation of this feature should be avoided. Sandstein (talk) 08:33, 31 December 2007 (UTC)
Said unnecessary bureaucracy has been removed. – Gurch 10:06, 31 December 2007 (UTC)
I would prefer if administrators could "block" rollback usage for a determined time, like a block. Having two options only (give and take) is problematic, because some admins will prefer to only punish serious offenses. We can block someone for a hour, a day or a week, but we would have problems if our only options were unblock and block indefinitely. -- ReyBrujo (talk) 14:47, 31 December 2007 (UTC)
You can always restore access to the tool as soon as the period of suspension is over. I don't see that a major obstacle. Spartaz Humbug! 18:01, 31 December 2007 (UTC)
Yeah, but it is not automatic. That means we will have to have some list of users with temporary removed access as to not bother people to request again once their "block" is finished. -- ReyBrujo (talk) 20:24, 31 December 2007 (UTC)
I don't think rollback is such a vital tool (like being able to edit at all) that there will be many uses for a very-short-term removal. If a user is using it to edit war, why give it back after a day or 2? IMO, they should have to re-request it and convince people that they can be trusted with it again. Mr.Z-man 20:48, 31 December 2007 (UTC)
Why not simply make rollback an autoconfirmed feature, like page moving? It doesn't seem weighty enough that it should need a special approval process, since it doesn't let the user do anything that couldn't be done by hand with a few seconds more work. *** Crotalus *** 01:58, 1 January 2008 (UTC)
That has been discussed, however, sleeper accounts could be used to vandalize using rollback, or users who have shown they are clearly incapable of controlling their actions and would abuse rollback would also gain access to it. Autoconfirmed accounts mean that someone has been around for four days; that doesn't really seem like enough time to understand reverting, much less rollback and its intricacies. --MZMcBride (talk) 03:34, 1 January 2008 (UTC)
I don't think that the vandalism concerns really have that much weight. Anyone who wants to vandalize can do so in a dozen different ways that have nothing to do with rollback. That will be the case as long as this is the free encyclopedia that anyone can edit. We can set a different period (say, 60 days) for autoconfirmation if 4 days is insufficient. But I do think it should be automatic. We have too many bureaucracies already. *** Crotalus *** 04:29, 1 January 2008 (UTC)
I personaly dont really see how giving rollback to autconffirmed users would be harmfull. The vast majority of vandalism (over 97%) is commited by IP's. Even if someone was to wait a while how much harm could they do? Sure they could roll back people contributions faster but they could have their vandalism undone just as fast. -Icewedge (talk) 04:44, 1 January 2008 (UTC)
I should also note that if autoconfirmed users get rollback automatically, it means it would be hardcoded into the settings and could not be taken away from a user like with this system. Mr.Z-man 05:02, 1 January 2008 (UTC)

Would IP's be permitted to use the tool as well? CambridgeBayWeather (Talk) 08:07, 1 January 2008 (UTC)

Nope, you can't change the rights for an IP. Ryan Postlethwaite 15:12, 1 January 2008 (UTC)
And we wouldn't want to, in case the IP got reassigned. Hut 8.5 15:37, 1 January 2008 (UTC)
If it was given to autoconfirmed, as if they were to vandalise in other ways, they should be blocked for abusing their editing right (rollback is an edit). Tiddly-Tom 15:40, 1 January 2008 (UTC)
The main abuse of this would be edit warring, not vandalism. We don't block for every edit war. Also, if a user is going to wait for 4 days to vandalize, there are far more destructive things that rollback. Mr.Z-man 22:03, 2 January 2008 (UTC)
Yes, but what happens when an experienced sock of a vandal rolls back all of clue-bots edits? Malinaccier (talk) 00:48, 3 January 2008 (UTC)
In that case we would block, but as far as abuses of rollback go, that would be a 1 in a 1000 occurrence. Edit wars happen far more frequently than things like that. Also, if it was given to all autoconfirmed users, it would be slowed by the rate limiter (the addition of which to the code was the genesis of all this discussion). Mr.Z-man 01:17, 3 January 2008 (UTC)

As a Spam patrolman I would like to know who will get the rights to this tool. Will it be given to all Spam patrolman or you have to be one of the regulars? And what does one of the regulars mean in the first place? Is it someone who has been at WikiPedia as long as Jimbo? Igor Berger (talk) 10:27, 4 January 2008 (UTC)

Consensus

I wonder how whomever makes the decision will determine "consensus" from this "discussion". I note the "voting is evil" mainstay of such polls, as well as several people who have commented but not "voted". - jc37 06:08, 3 January 2008 (UTC)

Over two thirds of this sample agree with the main idea of giving rollback to users other than admins. The main complains are "Administrators should not be the ones handling it", that "Users will abuse it", and that the prerequisites are somewhat weak. There may be some steps to try to fix those three points before implementing this. -- ReyBrujo (talk) 12:17, 3 January 2008 (UTC)
Currently, 81% of those who have left a numbered comment / vote have support the idea of giving rollback to non admins. That, by most standards, would indicate consensus among the community. However, ultimately, the final decision rests with the sysadmins (those capable of making live code changes). In regard to the comment about those who commented but did not "vote," I assume you're referring to those who didn't use a bold support. For those comments, it seems clear that those users supported the idea, even if they didn't spell it out.
In response to ReyBrujo's third point, we seem to have hit a catch-22. If you look at the old versions of this page, you'll see that there used to be specific criteria for granting +rollback. However, users were quick to complain about those "strict" criteria, so the prerequisites were modified to be more loose and open to administrators' discretion. Now that the prerequisites are not so narrow, users are complaining. It seems to perfectly fit the mantra that "you can't make everyone happy." However, the current proposal is the one being discussed, and it seems that the majority of users agree that administrators can be trusted to assign +rollback.MZMcBride (talk) 20:56, 3 January 2008 (UTC)
I'd like to add at this particular point, that it seems apropo that a proposal regarding rollback may itself be in need of rolling back. If we are to use this as a case study, though, I hardly think that we would want some non-admin user rolling something like this back simply because he/she is able to. Because of the broad implications of such a tool, I am going to restate my opposition and say that I am not sure that anyone is ready to handle 6+ million (that's someone else's estimate) people having a more efficient way to edit someone else's work (yes, even if they are only supposed to use it on their own edits, there is no doubt in anyone's mind that this tool can and will be abused some time in the future if it is implemented, and would only make revert war more thorough and damaging).Ecthelion83 (talk) 07:07, 4 January 2008 (UTC)
I will say, as a safeguard, if this turns out to be a failure, there is nothing permanent about the software change. Should this change be made and deep concerns emerge that leave the community wanting to revert this change, it can, and would, be done. --MZMcBride (talk) 20:56, 3 January 2008 (UTC)
As far as "Users will abuse it" goes, only giving it to users who have shown they have some experience and no history of behavior problems, combined with WP:AGF should make this not much of an issue. Mr.Z-man 01:14, 4 January 2008 (UTC)
And who will decide that a user has "no history of behavior problems"? A single admin. Based on what criteria? No specific criteria. A recipe for trouble, if you ask me. Freederick (talk) 13:30, 4 January 2008 (UTC)
What do we currently ask admins to do all the time? Stop edit warring, respond to incivility, enforce behavior rules. Mr.Z-man 20:21, 4 January 2008 (UTC)
If this tool is to be implemented, I would really like to see how it works in a beta version test by a limited test pilots under developers and supportes mentor program. Please be resposible for all the editors and for WikiPedia. What seems easy today may be a social engineering Malware tomorrow. Maybe best to form a special committie for its deployment and monitirng. Igor Berger (talk) 16:10, 4 January 2008 (UTC)
I completely agree with McBride here. It seems to me as if the concerns have be adequetly addressed as best they could. People are opposing based on their ideas of everything that could go wrong but there is simply no way they can know these things will go wrong and the experience of twinkle etc suggest the big catastrophes probably won't happen. In the unlikely event they do, we can just kill the whole thing. I don't see any reason to assume this is going to so badly kill the project we won't be able to recover. Sure we have to be cautious but there's a difference between be cautious and not doing anything because of fears about the million and one things which perhaps, maybe, could, possible, might go wrong. In tersponse to Igorberger, I don't see any point for 'limited test pilots' or a special commitee. It's not as if this is a completely new tool which may cause the servers to explode, just one that has been limited to admins for performance reasons. Since it's likely the only people involved in such a pilot will be so highly trusted and everyone will be on their best behaviour I find it highly unlikely a test pilot will tell us anything. If things are going to go wrong, they won't be detected in the test pilot. Nil Einne (talk) 18:39, 4 January 2008 (UTC)
I've seen this argument several times. The most basic response is: You don't know that it will go "right" either. And I'd rather err on the side of caution on this. (There is no way that anyone can claim that this tool is a need. It's a nice "want", but not a need. Wikipedia won't crash without it.) - jc37 18:43, 4 January 2008 (UTC)
Yes but the point is it's not as if we are making an irreversible decision which may blow up the servers. In other words, if it goes wrong, then kill it. Yes this tool isn't a need but there is a very good chance IMHO this will make wikipedia better. A small chance it will make it worse. It seems to me we should try it and if me (and the majority who support it) are right and it goes fine then good. If it crashes and burns then bad, we live and learn. Agaim I emphasise this is not a permanent irreversible change. Unless there is good evidence this will go wrong, I don't see any reason we can't just give it a try. Not doing it because it might, maybe, perhaps, could, possibly go wrong seems simply a bad idea to me. Nil Einne (talk) 19:52, 4 January 2008 (UTC)
Nil, "Yes but the point is it's not as if we are making an irreversible decision which may blow up the servers. In other words, if it goes wrong, then kill it," I hope you, as Alexander the Great did, are not leading us on march eastwards in order to find the end of the world. Igor Berger (talk) 02:15, 5 January 2008 (UTC)

Rejection of dichotomy

Mixed response. • I support the general idea of non-administrator rollback, but I oppose the current proposal as the wrong solution. • I think the alternate idea of "Assign it automatically; revoke if abused" is a much better one. It fits the wiki spirit of "anyone can edit". "Automatically" could be every named user, or some objective and automatic criteria (edits, age, combinations, whatever). Adding more bureaucracy is to be avoided. • If we must have the bureaucracy, then step 2 in "The way it works" needs to be better defined. WP:RFPP works as a "simple request" because WP:PROT is clearly defined. • I object to the current wording of the "Arguments and counter-arguments" section, which I feel has unwittingly becomed structured as "Reasons why objections should be overruled". I especially object to "Opposing for reasons like this makes it harder to get anything changed." That verges on "Your opposition is invalid because it opposes this proposal." • I've created a new section for more complicated responses because I think the above Support/Oppose needlessly polarizes the discussion. This has become "One must either suppose or oppose", which does not help advance consensus. —DragonHawk (talk|hist) 14:17, 4 January 2008 (UTC)

There is a section for this below, under "simpler proposal". I support this as well. Equazcion /C 14:21, 4 Jan 2008 (UTC)
The Arguments section above is just another part of the proposal rationale. The proposer has predicted common arguments and preemptively responded to them, as a kind of FAQ. Just because it was divided into its own section and made into a list form doesn't mean it needs to be balanced. It's part of the proposal itself. Opposing arguments are stated in peoples' !votes. Equazcion /C 14:25, 4 Jan 2008 (UTC)
What you going to do with sleepers and sockpuppets? Igor Berger (talk) 14:27, 4 January 2008 (UTC)
I'm not sure what you mean or what you're responding to, could you clarify? Equazcion /C 14:30, 4 Jan 2008 (UTC)
What I mean if you allow anyone who is an editor not an anon IP or less than 4 days, you can have a person registering for an account, doing a few edits than lurking in a background stalking an editor. When an editor has a disagreement with another editor, the sleeper jumps out and rolls back the whole article. The editor will have to go to ANB to restore the article, imagine if it happens a few times? The editor may lose his or hers credibility. This is very often done by Sockpuppet (Internet). How would you guard aginst such abuse? Igor Berger (talk) 15:24, 4 January 2008 (UTC)
I don't see why you would need to go to ANB to restore an article to a revision prior to the vandalism. A rollback still registers as an ordinary edit in the article's history. Simply undo it and report the user as a vandal/sockpuppet. Equazcion /C 15:28, 4 Jan 2008 (UTC)
You still need to respect 3 edits rule per user to the article per 24 hours and block the editors after that from editing that article. Igor Berger (talk) 15:38, 4 January 2008 (UTC)
Not in the case of overt vandalism. 3RR doesn't apply there. Equazcion /C 15:47, 4 Jan 2008 (UTC)
Can there be some sort of test given to an editor wishing to use this tool? You can build questions and answers from this discussion. The test should not be multiple choice but short answers so a small administration committie can guard against power hungy editors and possible abusers. Igor Berger (talk) 15:05, 4 January 2008 (UTC)
That would be the same as the current proposal in that users would need to put in a request and be approved before acquiring the rollback function. Equazcion /C 15:10, 4 Jan 2008 (UTC)
Good, I think we are all learning how this tool should be used and who should use it. If the editors wishing to use this tool can live with this consensus that we developing I see no problem for it. But if I see a first respondent editor power hugging an article I will file an incident report at ANB, RfC, and even at RfA. Igor Berger (talk) 15:50, 4 January 2008 (UTC)
  1. I still do not know. I am seeing too much support give it to me! Maybe this should be taken to RfC and even RfA to gain approval and see what the administrators have to say about this, being they the one's have to deligate authority. Igor Berger (talk) 16:23, 4 January 2008 (UTC)
    RfC means that you request comments on an existing discussion. It isn't a separate place on Wikipedia. RfA is where new administrators are elected. It isn't a place where proposals or disputes are handled. Equazcion /C 16:27, 4 Jan 2008 (UTC)
    So should we just give you the keys to the doors and trust you? I had to fight a sockpuppet for three days at ANB before he died! Do I need to fight a patrolman for 1 month to get to be able to edit an article? Igor Berger (talk) 16:34, 4 January 2008 (UTC)
    No, if you have a problem, just take it to an appropriate venue. If you are not vandalising then any user is likely to have the tool removed mightly quick. The thing here is that this tool is only going to be available to restricted set of people so it's not as if they can keep coming back at your like a sockpuppet. BTW, this was already posted to the AN. Administrators are welcome to (and I'm pretty sure are) coming here and discussing the proposal. Nil Einne (talk) 19:49, 4 January 2008 (UTC)
    Nil, thank you for alleviating my consern somewhat, but looking at the vote tally, I feel this thing is being shoved down our throughts! We are consensous society not a voting society! But that aside, I as a Spam patrolman and a security expert for many years am being treated as a Spammer by my security detail WikiPedian team members. An admin calling me a Troll, a senior user having an editor war with me because I added some relevent security deffinitions on WikiPedia Spam project article page. Are these not the trusted individuals who will get this tool? Will it not be the admin who called me Troll, who will let his regulars get this tool? Hey honestly I am scared. When I visit their user pages, I feel like I am in an interigation room. But I know what I am doing with regards to computer sience investigation and can deal with that. How about the regular Joe editor who has no idea why all his edits are being reverted faster than a twinkle, how will they feel? They will run a way from WikiPedia faster than the mice jomping a sinking ship! If you cannot educate your patrolmen, do not give them the deadly weapon!!! Igor Berger (talk) 23:29, 4 January 2008 (UTC)

Archiving

I personally really don't care about this issue, as I am not voting, but for those that are we need to do something about the size. As I am editing, it is about 283 kilobytes long and can't get much longer before people won't be able to edit this page anymore. I think we need to archive the sections but I need some feedback. Editorofthewiki (talk) 18:00, 5 January 2008 (UTC)

I was thinking the same thing, doing it now.
I also moved general discussions to the talk page, where they presumably belong. Equazcion /C 18:06, 5 Jan 2008 (UTC)

Voting is evil

  1. Oh BTW, what's all this "support" and "oppose" about anyway? While it's convenient to put one's comments in one section or the other to easily mark one's stance on this, I hope nobody comes up with a brilliant idea of actually making the results of this poll binding. Or did we start to enact policies by voting and I missed that? Миша13 09:49, 31 December 2007 (UTC)
    True, voting is evil, but this case is different (and it's not a clear cut vote). The devs want to see consensus clearly demonstated and this method is far better in showing consensus rather than a long convoluted discussion that conclusions can't be brought from. You also miss out on the views of people that simply support or oppose is but don't have anything really extra to add to the discussion. Ryan Postlethwaite 14:35, 31 December 2007 (UTC)
    Yep, this is an often-overlooked benefit of voting: It is relevant to know if a particular point of view is held by a few people, or a few hundred. Having a few hundred people say exactly the same thing does not lead to a better discussion than having a few people say it, and the rest say "yes, I agree".--Srleffler (talk) 06:31, 4 January 2008 (UTC)
    The problem with voting in most situations on Wikipedia is that there is usually no ballot. In a "real" vote, the thing you're voting on is well-defined. "Who do you want to elect?" "Should smoking be banned in government buildings?" On Wikipedia, things are almost always constantly changing. Most pages continue to be heavily edited while the supposed "voting" is going on. What one would supposedly be "voting on" is continuously evolving. For a vote to be an accurate reflection of consensus, the thing being voted on has to be well-defined. • I note that this proposal appears to be changing as people "vote" on it. —DragonHawk (talk|hist) 21:36, 4 January 2008 (UTC)

What happens is we get a big mess, where some people are voting, but others are discussing.

  1. This poll is absolutely absurd -- with little advertising, and a short window, I don't think this should be enacted without further discussion. Ral315 (talk) 10:02, 4 January 2008 (UTC)
    Little advertising? It is/was on the village pump, the administrators' noticeboard, centralized discussion, and the watchlist notice. There aren't many more places to put it. Mr.Z-man 10:51, 4 January 2008 (UTC)
    Yeah well, hehe - I didn't see it on my watchlist notice until yesterday. And maybe I should put a reminder on my left menu to visit all those pages you so happily stated ;) The first time I saw this page - Supports led opposes 3:1. --- Then again - there's only so many users you can notify 8)  — master sonT - C 15:48, 5 January 2008 (UTC)
    There are almost 300 participants so far, clearly it has been advertised well. I can assure you that it has not been selectively advertised to favor one side. If the people in favor are showing a lead it is because more people want it, not how the discussion was held. 1 != 2 15:56, 5 January 2008 (UTC)
    Further discussion, like the kind going on at this page? Equazcion /C 11:03, 4 Jan 2008 (UTC)
Not everyone keeps an eye on every one of those places. The watchlist notice is what brought 90% of the people here, and that was placed very recently. KnowledgeOfSelf | talk 13:36, 4 January 2008 (UTC)
No-one has to keep an eye on every one of those places, it's not like finding this page is a puzzle that requires pieces from stuff posted at each location. This is what WP:CENT is purportedly there for in the first place, the fact that it alone isn't enough is leading me to seriously consider nominating it for MFD (I've always considered it a bit of a backwater)—Random832 15:13, 4 January 2008 (UTC)
People are not just "voting", they are giving English descriptions of their opinions and responding to each other, that is a discussion. The poll is the format of the discussion, not the means of determining the result. 1 != 2 17:53, 4 January 2008 (UTC)
That does not appear to be the case. When I tried posting an "English description", it got moved to "discussion", apparently because it wasn't a simple vote, but a reasoned response. I could have moved it back, I suppose, but I didn't want to get into an edit war. —DragonHawk (talk|hist) 21:30, 4 January 2008 (UTC)
Well, whoever did that was wrong. You are supposed to explain your opinion. 1 != 2 15:41, 5 January 2008 (UTC)
  1. I realize that it's not part of Wiki policy to consider voting good, but... well, voting IS good. We should be using votes for everything. Debate and discussion usually just goes around and around because attempts to put everyone on the same page will always fail. Voting is pure democracy with little room for anger and flames. It should be embraced. --Bishop2 (talk) 19:55, 4 January 2008 (UTC)
Well, that's nice, and I might even agree, but you're talking about changing a fundamental aspect of how Wikipedia works. If you want to propose that we overturn and strike WP:NOT#DEMOCRACY, WP:Consensus, and WP:POLLS, you should propose it at the Village Pump. I will warn you that there is likely to be a large outcry against the idea. —DragonHawk (talk|hist) 21:42, 4 January 2008 (UTC)

The fact is, admin has already decided what they will do. Consulting users is a mere formality. Democracy within Wikipedia is a facade. I say this as someone who enjoys Wikipedia. But I enjoy it realising that it is not democratic: it is controlled at every level.--Gazzster (talk) 11:12, 5 January 2008 (UTC)

Whether this is implemented is not up to the admins, its up to the developers. Mr.Z-man 11:15, 5 January 2008 (UTC)

the case for bureaucratic expansion has not been made

The proposal under discussion makes no case for the provision of rollback except to remark that other methods are "less efficient in bandwidth and time." The only effect of this proposal would be a saving of bandwidth (totally unquantified) and time (exceedingly minor), and an (easily quantifiable) increase in wikipedia bureaucracy as people apply for the tool and administrators reject/accept the applications and prosecute violations.

We have enough bureaucracy on wikipedia; I suggest we not create more without excellent reasons that are here completely lacking. Sdedeo (tips) 18:47, 4 January 2008 (UTC)

Actually having to do multiple undo makes the edit history more confusing and does take a fair amount more time particularly with a slow connection (try using a modem with a large page). It's clear that some RCers are asking for this and that some admins are willing to partake in the process. If you don't want it and don't want to grant permission that's fine, but is there any reason why people who want it shouldn't get it when it doesn't negatively effect you?Nil Einne (talk) 19:30, 4 January 2008 (UTC)
Multiple undos? If you need to undo multiple edits by a single editor, then a normal revert is more efficient. If you can't do that, because there are subsequence edits, then rollback won't work either. Rollback is far harder to use on a slow connection than "undo" or normal reverts. Guettarda (talk) 20:02, 4 January 2008 (UTC)
Are you telling me having to undo multiple times is faster then rollback? I'm somewhat doubtful but never having used rollback I can't actually say. What is a normal revert? You mean editing an older version of a page and saving? I do this sometimes but it can be annoying and is risky since it is relatively easy to accidentally revert a lot of good edits. Furthermore, unless I'm mistaken a rollback can work whenever you can undo. There are lots and lots and lots of cases when you can undo with intervening edits. For example, when someone modifies 'BUSH IS A GAY' but is unable to spell and has bad grammar so has to edit this vandalism several times. Nil Einne (talk) 16:56, 5 January 2008 (UTC)

I'm opposed to bureaucracy in general, and in particular on wikipedia, which is gradually accumulating an absolutely massive amount and, as a consequence, shifting increasing amounts of power to small subgroups and not the editors who make the bulk of mainspace contributions. Giving more authority to admins to certify people and punish violations of a new rule seems like something you need a great deal of reason for. Sdedeo (tips) 19:42, 4 January 2008 (UTC)

I see very little if any bureaucracy, there is not even a proposed system other than admin discretion. This is a new tool, not a new procedure. Perhaps more codified procedures will form if they are seen to be needed, but it seems we are starting out with system devoid of red tape. 1 != 2 20:47, 4 January 2008 (UTC)

As per the proposal: you apply for the tool, and an admin reviews you and, if satisfied, chooses to approve your application. Admins are also responsible for prosecuting violations of the rules for this tool by revoking permission. The increase in bureaucracy and the concomitant increase in administrative powers granted to a subgroup (both the "approved user" and the admin who approves her) seems to give no significant benefit to the community. (The proposal is divided as to whether rollback is much more efficient [doubtful] or a "slight performance benefit".) Sdedeo (tips) 20:54, 4 January 2008 (UTC)

Too often in Western society bureaucracy is viewed in a derogatory sense. While overbearing bureaucracy can harm efficiency, some bureaucracy is needed. Imagine how unorganized and unstandardized the processes would be, if we did not have the admin's noticeboard, RfA, and AfD pages. I think that as these are necessary for Wikipedia's efficiency, so would a "well-oiled" policy on a rollback feature help the efficiency. Mind, in my previous support comment, I mentioned hat this feature should go to very experienced users, and as there are not that many of users (comparatively) who would be able to make effective use of the feature, the feature's maintenance would not take away from the efficiency gained. Having used a rollback script, I can say that it is more time consuming than I would like.--Vox Rationis (Talk | contribs) 02:21, 5 January 2008 (UTC)

While the benefit may be minor in your opinion, those who do not wish to gain from it do not need to even notice any new process. And I really don't view asking and then someone making a decision an overly bureaucratic decision. 1 != 2 15:43, 5 January 2008 (UTC)
Exactly. No one is saying it will be compulsory for admins to partake in the process. If they don't want to fine. Nor is anything saying you need to have the tool. If you don't want to fine. But for those users who do want the tool and those admins willing to partake in the process why should we be denied just because people feel well I don't think it's necessary? Nil Einne (talk) 17:00, 5 January 2008 (UTC)

Just felt like mentioning something in the Wikipedia:Wikipedia Signpost. Apparently the code has now been changed so admins now have the power to grant individual rights to users (such as rollback, which is used as an example...). .:Alex:. 16:03, 5 January 2008 (UTC)

Sounds good, so now we can all go home. There is no need for non-admin rollback tool. Igor Berger (talk) 16:09, 5 January 2008 (UTC)
Unless I'm mistaken, one of the whole reasons for this discussion was to decide whether admins should grant the tool and to who which hasn't be resolved. We already knew the devs we're probably going to add the feature and they had asked us if we wanted. What the admins added was basically what we think we want but haven't decided yet whether we do. In the event there is no consensus, we don't actually have to use the feature Nil Einne (talk) 16:59, 5 January 2008 (UTC)No you're wrong. The signpost thing is precisely why we're discussing this. Basically if I understand it correctly, the devs told us they were going to implement the feature and whether we want some sort of non-adminstrative roll back. Eventually it was decided that an admin like roll back for non admins granted by admins is the best idea. The devs have implemented the feature which will enable them to allow admins to grant roll backs (and other tools). However it remains up to us to decide whether we want to ask the devs to grant admins the right to give roll backs. Nil Einne (talk) 17:15, 5 January 2008 (UTC)

I believe bureaucracy is "guilty until proven innocent", and just don't feel that the case has been made here. Put most broadly, this proposal is an opt-in variant of "trusted user", something that the community has objected to many, many times. I don't want to take up too much time with my objections here, so I'll leave it at that. Sdedeo (tips) 17:14, 5 January 2008 (UTC)

But the issue there was there was no defined reason for a 'trusted user'. Here we are saying many users want this tool. Many users feel they have need for this tool. No one IMHO has been able to explain why they don't need this tool except to say 'well I don't feel this tool will be useful' which is somewhat unfair IMHO to those who feel this tool will be useful. I can understand even if I don't agree with those who argue that adding this will create too many extra problems but it seems excessive to deny this simply because you don't feel it will be useful Nil Einne (talk) 17:19, 5 January 2008 (UTC)
If the user needs this tool so bad they should become an admin. If they do not want to be an admin than they could use twinkle It reverts the same way as the proposed solution and you can input a reason on non obvious vandalism. It is also better the way it is because it removes unneeded bureaucracy Alexfusco5 18:26, 5 January 2008 (UTC)

Edit summary must be mandatory for all users

I don't know if this is technically possible, but could the rollback for all people, admins included, be changed to require a mandatory subject entry? That would force all people to explain why they're reverting, even if it's just "rvv". That could help keep a ton of new people having this ability on the up and up. Lawrence Cohen 18:06, 4 January 2008 (UTC)

Seems like a good idea to me if possible Nil Einne (talk) 18:18, 4 January 2008 (UTC)
This would be a pain with one of the more helpful features of rollback. If there is a user (or vandalbot) which does a large amount of vandalism in a very short space of time, their edits can be quickly reverted by clicking on all the rollback links in that user's contributions. This would take a lot longer if you had to specify an edit summary for each one (even TWINKLE doesn't force you to enter a summary). It would slow down RC patrol for some people too. Under the current proposal, only a small group of trustworthy users will be given the rollback button, and if they use it for anything other then blatant vandalism they will have the tool removed, so if we can't trust that every edit they revert is vandalism and need to make them give an additional summary they should not have the tool. Hut 8.5 18:28, 4 January 2008 (UTC)
Sorry but I've neved had access to rollback not being an admin. But wouldn't a general edit summary be acceptable and not really slow things down much? If you're reverting a user for vandalism then a rvv one time wouldn't take that long. If you're reverting a user for vandalism once and the second time because you simply don't like the person, well you shouldn't be doing that :-P Mind you I'm not convince this is that important but some people seem to think it is so I supported it. Nil Einne (talk) 18:52, 4 January 2008 (UTC)

Just spotted this. I think that the current version is probably useful for admins. But I agree with it for non-admin usage. I've started a #Counter-proposal below. - jc37 18:39, 4 January 2008 (UTC)

If you need an edit summary then do not use rollback. Rollback is for reverting bad faith edits where the content of the revert explains its purpose. If you need an edit summary then rollback is not the correct tool to use. 1 != 2 19:33, 4 January 2008 (UTC)
Agree with 1 != 2 - if you need an edit summary, don't use rollback. Guettarda (talk) 20:06, 4 January 2008 (UTC)
I strongly agree with this too, should there be now edit summary, rollback won't work. Marlith 20:07, 4 January 2008 (UTC)
I thought the revert was only intended for obvious vandalism? I probably misunderstood. Lawrence Cohen 23:21, 4 January 2008 (UTC)
I've also used rollback for self-reverts - quite frankly, that's one of the best places for it, since you don't need to look at the diff and there's no fear of offending the person you're reverting. Guettarda (talk) 02:39, 5 January 2008 (UTC)

Twinkle and other tools

Am I missing something? Two of the most common arguments appear to be either that twinkle and other tools can do this already (which doesn't really explain why we shouldn't add rollback) or that adding this tool is going to unleash a scourge of editors who will edit war, revert people they don't like and generally wreak havoc. It seems to me these are mutually exclusive claims. If twinkle and such can do it all already then I fail to see how adding this is going to somehow unleash all these bad editors. So far, the only thing I'm getting that's different from twinkle etc that may help stop abuse is that they force an edit summary. Is there something else I'm missing? Nil Einne (talk) 19:26, 4 January 2008 (UTC)

In case people ask if twinkle can already do this why add it, well the point is other then the potential performance gain (and let's not forget it was the devs who asked whether we want it) both for the server and the user (especially for RCers using modem and other slow connections) there's also the issue of browser compatibility and the fact that people simply don't want to have to install custom javascript which does carry some small risk when the function can be easily and more effectively provided by the wikimedia software. However, if users really are the abusive kind, it seems likely to me they will take the effort to get a working browser and may not care so much about security risks. Nil Einne (talk) 19:29, 4 January 2008 (UTC)
TW is used for edit warring all the time. Expanding rollback makes it even easier. Guettarda (talk) 20:04, 4 January 2008 (UTC)
If someone usess rollback to edit war, he/she will lose that privilege quite quickly. --Conti| 23:51, 4 January 2008 (UTC)
People use TW all the time, and keep the privilege. People use rollback to edit war even now - you just need to be discrete and well-connected. People misuse admin tools and keep them. There's too much politics and there are too many cliques in Wikipedia. We will see long arguments "yes, he was edit warring!", "no, you're just trying to take away the tool because he reverted your friend", "no, you're just protecting him because he's your friend". Crap like that happens all the time (albeit, usually using coded language). The more people that have rollback, the more it will be used to edit war. Even if the tools are lost immediately upon edit warring. Guettarda (talk) 02:44, 5 January 2008 (UTC)
I can see how it would make it moderately easier to edit war, but how would it encourage it or cause more edit wars? Most people don't edit war just because they can, there is some sort of content dispute. Are users really that lazy that if they have to do a manual revert or use undo they will decide to just drop the issue but they will edit war if they have rollback? Edit wars rely on multiple people to do the reverting, so it will only make the individual reverts faster, not the edit wars themselves. Mr.Z-man 03:35, 5 January 2008 (UTC)
I think the consern maybe more with a renegate editor, who might have suffered centrain levels of an anxiety attack and just goes on a rollback mission! I recently seen a blocked editor do an edit to a user talk page. Are the dead Resurrected? Or the admin who implemented the block did not do it the right way, and an exploit has been created? How about if an editor with rollback has his user id and password hijacked, and the gohst editor goes mayhem? There are so many posibilities of exploits that are not even touched by this discussion. Admin participate in the community and have relationships accross all genre, but a recent change patrolman or conter vandalism member may exist in a limited sub group of the community. Shit happens. Igor Berger (talk) 03:50, 5 January 2008 (UTC)
If something like that were to happen, we would revert, block, and move on. Rollback is just an edit. If its used maliciously or for vandalism we would handle it like any other situation that didn't involve rollback. Mr.Z-man 07:20, 5 January 2008 (UTC)
Can you revert curl_init cronjob, how many computations will the rollback tool do in 10 minutes, 60 minutes. 60 minutes probaly 40,000 depeding on the server load. Make sure to put preventive mesures to restrict abuse. Say 3 operations per 5 minutes or ten minutes. The variable can be adjusted per user experience level. Maximum rollbacks 30 an hour, but this can also be increased with Trust ™! Igor Berger (talk) 07:31, 5 January 2008 (UTC)
Let's face it someone who starts mass rollbacking everything is going to be blocked within minutes. If it really get so bad, the devs can get involved. If this really happens often then the devs will probably tell us we need to do something. If the devs don't then I don't know if we should worry about it. IMHO the far greater worry are people who use this to edit war, wikistalk or 'own' an article since it may take a while for them to be noticed but even then I doubt it'll be that often and when it happens they're gone Nil Einne (talk) 17:12, 5 January 2008 (UTC)
But why would people be more likely to use rollback then TW? Perhaps I'm wrong but the vast majority of people who cause problems tend to be dedicated enough that they don't care if they have to change browser and use external code to do their dirty work so they are probably already using TW. It's the ones who want to improve wikipedia who don't necessarily have the time to dedicate to installing JS (which as I've mentioned is inherently risky even if the risk is small) and who may not want to be forced to use a certain browser. Also, the way I see it one of the problems with TW is that it's effectively somewhat unregulated by wikipedia. Yes I believe people can be banned from using TW but since it's not actually part of wikipedia it's somewhat of a tricky process. Rollback is part of wikipedia, granted to people who can be trusted with it and if it turns out you can't you lose it. Nil Einne (talk) 17:08, 5 January 2008 (UTC)

Rollback privilege criteria

I'm concerned about the vagueness of step 2 in "#The way it works", which currently states, "Administrators should check the history of the contributor to see if they can be trusted with the tool." I guess the theory is that admins can just look at the edit history of a user, make a decision, and flip the switch (or not). I've got nothing wrong with entrusting any given admin to make a determination. My concern is that this is a recipe for strife, discontent, and argument. Admins are people, and different people will have different takes. It is virtually a given that, in some number of cases, one admin would say "Yes" and another would say "No". Given that there's no "locking mechanism" on Wikipedia, I can totally see one admin explaining why he said "No" while the other is granting the privilege. What then? I expect a discussion between the admins, but without any defined criteria, it's going to come down to personal opinion, which makes for lots of long, drawn out discussions, wasted time and effort, the occasional wheel war, etc. And what happens when a user claims (in good faith) one admin's decision is unfair, that the admin is being too harsh? I'm concerned this will turn into WP:RFA all over again. As this says, "while nearly everybody seems to think there's something wrong with [RFA], literally years of discussion have yielded no consensus whatsoever". I would really hate to have the same happen to "Requests for rollback". —DragonHawk (talk|hist) 22:54, 4 January 2008 (UTC)

I would strongly suggest standards. With standards we will have no arbitrary giving of rollback according to one persons standards. Marlith 01:39, 5 January 2008 (UTC)

Power trip toy

Rollback is not a creative tool; it is a tool for exerting authority. It is nearly useless to those editors who do the actual work of writing articles. It is admittedly useful to admins, trusted and vetted to police others' work. The trouble is, there are too many wannabe cops around the Wikipedia; people whose idea of "contributing" is to boss others around on their own private power trips. Thankfully, there is the RfA hurdle to keep them from living out their fantasies at honest editors' cost. Well, guess what, there is now a proposed way for them to be "little admins" without all these pesky checks and requirements: the user rollback! Whee! Potential for abuse? You bet. Wikipedia is already groaning under the weight of all these pests, who get a kick out of deleting, reverting, chastising, and pushing other editors around, but who never bother to improve an article themselves. Give them rollback, and watch the constructive contributors walk away. Freederick (talk) 01:29, 5 January 2008 (UTC)

You realize that rollback is only ever used for cases of clear vandalism, and that an editor who used it for anything else would have his or her +rollback removed? --MZMcBride (talk) 01:54, 5 January 2008 (UTC)
Would he? The criteria for removal are as vague as those for granting it in the first place. Besides, removal is hardly as automatic as you pretend: it takes time and places a heavy load on editors and admins. Someone has to post a complaint requesting admin assistance, an admin has to take the time evaluating the request, and then decide, quite arbitrarily, to remove or not. No thanks. Freederick (talk) 02:05, 5 January 2008 (UTC)
Admins are not police. We are regular users with special tools. Every user is equally entitled to enforce policy, and the vast majority of vandal fighting is not done by admins, but regular editors. Little admins, big admins, stewards, 'crats, we are just users with different tools, same authority. Writing articles is of course essential, but the fact is that it just would not be possible if not for all the hard work others do keeping this place together. 1 != 2 02:33, 5 January 2008 (UTC)
Edit wars are already reported regularly if page protection or blocks are requested, removing rollback would just be a minor extra step and not the point of not giving it out to everyone is so that we don't have to remove it very often. How is giving out rollback going to drive other editors away? Do we now assume that people will just immediately start abusing tools once they get them and rolling back positive contributions en masse? Mr.Z-man 03:41, 5 January 2008 (UTC)
I'll tell you that I have started to do a lot of anti-vandalism work and if I had to do hundreds or thousands of manual rollbacks before I was somehow given a special privilege to allow it, I would simply not have bothered in the first place. I can assure you, staring at a feed of updates so you can rollback people spamming profanities and racist or homophobic comments all over the place is not only a thankless undertaking, but perhaps the least power-tripping endeavor you could possibly imagine. I sense a bit of "only people who actually write articles are important" power trip in your comment, though. I've only created and edited a few dozen articles, but I am grateful to the people who "don't do the actual work of writing articles", but have monitored and reverted outrageous vandalism on said articles. As for a regular user exerting "power" or "authority" over trouble-making vandals looking to disrupt a resource we all appreciate? I can think of far worse things. Now, exerting "power" over someone that you simply disagree with is another, but then for those people who commit such abuses, how about *removing* the ability, rather than stripping everyone of it until they've somehow proved themselves. For many who have other things occupying their time, but like to drop i and lend a hand now and again as circumstances allow, there is a point where they won't care to bother. Cordell (talk) 23:51, 5 January 2008 (UTC)

This keeps getting said at various places but I think a lot of people are still missing the point. Any vandal can already revert by using "undo". This could not be any easier from a vandal point of view. Rollback is not making it any easier for the vandal. It makes it slightly easier for the editor reverting vandalism and apparently a whole lot easier on the server. It also keeps the history tidier, one rollback is easier to understand than multiple undos when another editor looks at what has been happening. From the vandal point of view, tidiness is not a great concern; much easier to just carry on pressing the undo button than go to the trouble of getting yourself approved for a tool that does the same thing (and hopefully would be denied anyway). SpinningSpark 12:39, 5 January 2008 (UTC)

Please don't Limit the Usage

I disagree with the following bit: "Usage is limited to rolling back vandalism and reverting one's own edits." Here's why:

  1. There's no reason to disallow reverting mis-informed edits. For instance, please look at the following edits I made: diff 1a, diff 1b (rv) diff 1c (rv) and diff 2 (rv). I believe my reverts were totally appropriate (and certainly successful in dealing with mis-informed users), but the edits I reverted were fundamentally well-intentioned, and definitely not vandalism.

    By the proposed policy, I'm supposed to use the undo function instead of the rollback function just because of the type of edit I'm reverting. That's unnecessary (and not very helpful) — the existing measures, like 3RR (and my common-sense), do a fine job of preventing edit-wars, and rollback is different from undo only on a technical level. Also, since I frequently revert mis-informed edits like the ones above, I'd end up still using userscripts since the policy doesn't allow me to use rollback in all cases. What's the point of this?

  2. Another thing I'm worried about is that some day a troll I rollbacked causes hassle for admins and me by complaining that I rollbacked some edit that's not clearly vandalism. As I said, the existing measures for preventing un-justified reverts and edit-wars are just fine; no need to add policy-cruft.

So I suggest to remove the "Usage" section (1.3). Section 1.4, "Removal of the permission" still applies for cases where editors abuse the feature for edit-warring. -- Ddxc (talk) 01:57, 5 January 2008 (UTC)

Agree. In November, an editor made some hundreds of good-faith edits that ran counter to a manual of style point. When I called it to the editor's attention, he undid them, generally using AWB, over a period of a couple of hours during which I was too busy to do any more than a few manually (I didn't have time to learn to use AWB). This seems like a good application of rollback. With rollback I could have helped, but it's neither vandalism nor my own edits that would be rolled back. So there can be constructive uses of rollback for uses the proposal prohibits. Fg2 (talk) 02:13, 5 January 2008 (UTC)
In essence it should only be used when an edit summary is not needed. There are times where many good faith edits need to be reversed fast and rollback can be used as long as proper communication exists between the parties. These are the same limitation admins are expected to follow. 1 != 2 02:39, 5 January 2008 (UTC)
So do you agree that Wikipedia:Non-administrator rollback#Usage should be modified to allow this? It presently says "Usage is limited to rolling back vandalism and reverting one's own edits. Editors using the rollback tool for other purposes will be subject to having the rollback tool removed." This explicitly prohibits the use of rollback on someone else's edits even when "proper communication exists between the parties." Fg2 (talk) 04:11, 5 January 2008 (UTC)
That'd be a deal breaker for me. There really needs to be a really bright line around proper usage. With significantly more editors with the tool comes a greater chance for conflict...what does "proper communication" mean? To avoid that conflict, take an extra couple of seconds and undo with an edit summary. It wouldn't come up that often to justify opening it up for more possible mis-use. You may be sure your reverts will always be totally appropriate and that they will only be used with mis-informed users but we can't count on everyone being so sure or so commited to that limit. RxS (talk) 04:57, 5 January 2008 (UTC)
I'm just not convinced that the usage policy of a technical tool is the right place to enforce this. It happens just as easily that people undo without comment. WP:RV#Rollback says, "If you use the rollback feature for anything other than vandalism or for reverting yourself, it's polite to leave an explanation on the article talk page, or on the talk page of the user whose edit(s) you reverted." after some explanations about good communication. I think this kind of admonition is the right way to go. No need to have a stricter usage policy for non-admin users. -- Ddxc (talk) 12:09, 5 January 2008 (UTC)
Reversion of another user's edits after communication with them comes up very, very rarely. I've been an admin for 7 months and done it twice. Neither time was an emergency. Admins are generally willing to listen to good reason. If for some reason someone complains about something like this, just ask the user whose edits were reverted. If they say they gave permission or there is a diff that clearly shows them giving permission then it should be no big deal. Mr.Z-man 06:17, 5 January 2008 (UTC)
Because rolling back doesn't allow summary (like the undo feature), how can you know if an IP rolling back a 5 to a 8 in an article about the average size in inches of some extincted bird found in the Everest is hidden vandalism or not? Rolling back should only be for vandalism that is easily identifiable. -- ReyBrujo (talk) 17:13, 5 January 2008 (UTC)
As I understand it, IP editors cannot get rollback privileges. Fg2 (talk) 01:04, 6 January 2008 (UTC)
That's my understanding as well. Users will need to at least be autoconfirmed. Equazcion /C 01:11, 6 Jan 2008 (UTC)

Query?

Got this in the mail so, late to the party – please forgive if I got it wrong, but is there a way to offer it to people with a record of lots of de-vandalising edits handled manually? I'm not sure how else to find vandalism than when I just come across something obviously silly (so not very techie there) and just take it out. It's probably happened once or twice. So though it sounds like a wonderful opportunity for doing lots of reverts, it hasn't been my project to date. Guess I'm saying not for everyone, but with a selection process if that's not too clunky for admins to do. Thanks for getting the notice out there. Julia Rossi (talk) 09:38, 4 January 2008 (UTC)

I didn't get a message about it, and it's usually considered good form to do so. -Halo (talk) 18:46, 4 January 2008 (UTC)

Is there any bot which could do this? Nil Einne (talk) 20:17, 4 January 2008 (UTC)
I'd say User:Ral315 with his User:Ralbot (delivers the sighpost), is probably your best bet. Good idea though. MBisanz talk 04:59, 5 January 2008 (UTC)
AWB could do it, I think. Create list of links on the old proposal page. (Optional:To prevent messages being sent to those users who have already voiced their opinion in this one, you create a list of links from this proposal, compare lists and use the ones only in the first proposal) Filter out non user pages. Filter out duplicates. Convert to user talk. Add at the end of the user talk page the message. I do not have AWB downloaded, but someone who does could do it. I doubt it would take too long. Tiddly-Tom 09:52, 5 January 2008 (UTC)
I was sitting in a meeting this morning and had that very thought. Could someone compose a brief notification message pointing users here who had commented previously? I'll do the AWB legwork, as I just learned how to do it for another proposal I sent out earlier today. MBisanz talk 19:43, 5 January 2008 (UTC)

Standards

Should this proposal be adopted and enacted what ought be the standards for NAR? I would suggest, three months online and a familiarity with policies as well as a devotion to anti-vandal work at the CVU. Marlith 20:00, 4 January 2008 (UTC)

If you look deep into the history of the proposal, standards were rejected in favor of admins' discretion. --MZMcBride (talk) 20:31, 4 January 2008 (UTC)
ah. Marlith 01:29, 5 January 2008 (UTC)

Other wikis?

Is it likely that other wikis will get this (rollback for non-admins)? Since there seems to be great concern that it will crash and burn, in the event that this proposal fails perhaps if other less conservative wikis get it and it turns out fine, consensus will change? Nil Einne (talk) 20:16, 4 January 2008 (UTC)

From what I understand, it merely requires on-wiki consensus and a bug request. --MZMcBride (talk) 20:32, 4 January 2008 (UTC)

Restore this version

Is the useful Twinkle tool 'restore this version' part of admin rollback? If not it should be bundled with it and with this proposal. Obviously many vandals make more than one consecutive edit, or there are consecutive vandals. Perhaps rollbacks could also become colour coded (e.g light green) so the unreadable mess that is a page history could become slightly easier to navigate. Richard001 (talk) 00:04, 5 January 2008 (UTC)

No, and neither of those are currently possible with the software AFAIK. Rollback reverts all the consecutive edits made by the last editor to a page. If a vandal makes 2 edits in a row, and no one else has edited it since then, rollback will revert both. In other cases one would still need Twinkle or manual reversion. Mr.Z-man 00:49, 5 January 2008 (UTC)

Drawing board?

Maybe it's time to go back to the drawing board here. I don't think a consensus is going to be developed here, the discussion is getting pretty fragmented and various alternate versions are being presented. Time might be better spent tuning the proposal in light of the discussion to this point instead of just adding supports/opposes. Editors that participated early on in the debate may find that the current version has drifted out from under them. Bottom line, I don't think a consensus is going to arise from the current discussion. I think we're looking at a pretty small target here consensus-wise so some fine tuning is probably in order. RxS (talk) 05:44, 5 January 2008 (UTC)

I agree that the original proposal should be tweaked, but I think we need a centralized discussion on the possible variants first. If the proposal were changed now I'm not sure what it would be changed to, as everyone has a different idea of what it should be. Equazcion /C 06:33, 5 Jan 2008 (UTC)
Keeping the discussion/tuning going would be a good thing. My point is that perhaps the polling part of it be suspended for now until a variant emerges that has a chance at consensus. I don't know what that would be, that's the small target I was talking about...but I think discussing and tuning the proposal is a better use of time then keeping the voting open. Especially if the scope starts to drift, we'll find that people will have to keep checking back to make sure they still hold the same opinion as they did when they first expressed it...why not settle on something and reset the polling? RxS (talk) 07:05, 5 January 2008 (UTC)
Better to let the vote roll on and establish in principle whether or not we want this feature. If you wait until it is perfect nothing would ever get off the ground. Tweaks can always be proposed and implemented later. Let's deal with the big issue first - should it exist at all. Besides which, only actually running it for a while will show what the problems really are going to be; I suspect that much of this discussion will prove to have been a waste of time looking at problems that are not really there in practice. SpinningSpark 10:46, 5 January 2008 (UTC)
Yeah, I agree that the vote should be allowed to "complete", whatever that means. There should be some kind of time limit placed though, as we have 6 million registered users and voting could conceivably drag on much longer. But if we just clear it out and restart it, people won't like that (especially the supporters, who undoubtedly are excited about their clear "victory"), and it would be good to determine if the general idea is accepted before we fine-tune it. Equazcion /C 13:42, 5 Jan 2008 (UTC)
Equazcion, do not be in a hurry, if you want this to work! And I see you begining to think ergonomically, which is a good sign. Igor Berger (talk) 14:28, 5 January 2008 (UTC)
We can let it roll on then, but there's no way this is going to reach consensus in it's present form. But just so it's clear, I'm not suggesting that it gets cleared out and restarted, I'm suggesting that since it's not going to achieve consensus the poll is ended until something that is more likely to find more common ground is found. But I feel like I'm repeating myself so let it go on... RxS (talk) 14:38, 5 January 2008 (UTC)
I understand the concern, and like I said, there should be a time limit set... wonder if we could get away with posting some arbitrary limit now, like 3 more days? Equazcion /C 15:12, 5 Jan 2008 (UTC)
We had a time 'limit' but it was removed because people felt there wasn't enough time (particularly since it was so near Christmas/New Year) for discussion and were opposing solely for that reason. Mind you, I guess people won't object if we put back a similar message and explain that it will be closed down in 3 days (or whatever) because consensus is obviously not going to be reached and we need to tweak the proposal (although personally I don't hold much hope for sucess) Nil Einne (talk) 17:31, 5 January 2008 (UTC)

←Then it looks like this is just gonna go on until it gets so ridiculous that people stop bothering and forget about it. Like I said: 6 million registered users, all seeing this ad in their watchlist, and so far its taken a week for only 300 of them to vote. The math doesn't support a timely resolution. Equazcion /C 17:42, 5 Jan 2008 (UTC)

I don't think there is just some easy fix you could make that would sway a bunch of opposers without losing many supporters. Most of the sensible opposition centers around bureaucracy; its either too much bureaucratic instruction creep or it needs more rules to prevent abuse. You can't have both. Mr.Z-man 22:41, 5 January 2008 (UTC)
You know very well that there are only a few thousand active users at any one time. Splash - tk 23:27, 5 January 2008 (UTC)
It doesn't matter how many are active at any one time. If only a thousand users are active per month (which I sincerely doubt), that still means next month another set of users can vote. This can conceivably go on forever, no matter how you slice it. Equazcion /C 01:17, 6 Jan 2008 (UTC)
Take a look: Wikipedia:Non-administrator rollback/Creating a new proposal Equazcion /C 15:32, 6 Jan 2008 (UTC)

Bot rollback

Does the rollback feature being given to non-administrators include the bot rollback functionality? Is it possible to disable that capability for non-administrators? Christopher Parham (talk) 06:47, 5 January 2008 (UTC)

Do you mean how admins can modify the rollback url to make it a bot edit and thus hide it from recent changes? I think that is an admin privileged not sure though. I agree if should not be granted to non-admins with the rollback, but I don't think the way the tools are laid out will result in this, I think they are separate permissions. 1 != 2 17:45, 5 January 2008 (UTC)
This capability was bound to the 'rollback' permission, but I dug out this change by Werdna (from Nov '07), which shifted it to the 'markbotedit' permission. Миша13 17:58, 5 January 2008 (UTC)
Good to know. 1 != 2 19:43, 5 January 2008 (UTC)

Request to rewrite part of the proposal

There is a part of the proposal that doesn't quite make sense:

  1. Administrators should check the history of the contributor to see if they can be trusted with the tool.

I assume that it means to check for history of abuse, but it's not very clear. Should it be rewritten to make more sense? .:Alex:. 16:34, 5 January 2008 (UTC)

I think "history of the contributor" just means the user's contribs list (special:contributions/johndoe). It seems obvious to me that they'd be checking for evidence of abuse. Equazcion /C 17:38, 5 Jan 2008 (UTC)
It does not mean just abuse, there can be good intentioned bad edits, a failure to understand policy, a tendency to misinterpret policy. I think the sentence is just as general as it needs to be. 1 != 2 17:42, 5 January 2008 (UTC)
Well it's not very specific in either case, I still feel it should be rewritten to cover all these things. .:Alex:. 17:59, 5 January 2008 (UTC)

Closure

I've been paying attention to this page for some time and it seems to me that consensus has been reached that some type of non-admin rollback should be implemented, but that it shouldn't be implemented exactly as worded in the proposal.

I propose that:

  • We close this poll, as we've gotten all we're going to get out of it.
  • We archive the entire page.
  • We begin a new discussion on how the proposal should be changed so that it can become something people can generally agree on.

Equazcion /C 17:50, 5 Jan 2008 (UTC)

I won't really consider 200 to 100 consensus. Marlith 17:53, 5 January 2008 (UTC)
No, of curse not, not consensus support at least. Consensus support would mean we implement the proposal as currently worded. But even most of the opposers seem to be for the general idea, just not with the specific details as proposed. Equazcion /C 17:55, 5 Jan 2008 (UTC)
Modifying the proposal to take account of the requested recommendations will bring the consensus closer to fruition. You will need the approval of the bureaucrats. Once ready file it as a request for a Bot account. If you do not have the consensus, the bureaucrats will not grant you the account, because it will undermine their authority which is granted to them by the WP:stewards. Please make sure it is done correctly, and if you need any further assistance give me a hallo. Thank you, Igor Berger (talk) 18:54, 5 January 2008 (UTC)
Modifying the proposal has nothing to do with bureaucrats, bots, or stewards. Equazcion /C 18:56, 5 Jan 2008 (UTC)
If you put up a site notice asking if the sun should rise tomorrow a third will object. If you take into account the arguments then you will see that there is clear support for this feature. Many of the opposes are based on false assumptions and misunderstanding, and they are not uniform in their reasoning. 1 != 2 19:43, 5 January 2008 (UTC)
I think many have good reasoning but most of them still don't seem opposed to the fundamental idea of a non-admin rollback. they just seem to object to the current proposed method of implementing it. That's why I think we need to start a discussion on how to change it. Equazcion /C 19:49, 5 Jan 2008 (UTC)
Well, I think once we have the technical ability activated that the community can propose and ratify any policy related to it they want, just like any other tools. 1 != 2 19:51, 5 January 2008 (UTC)
This is more than a tool, it is about assigning admin rights to a non admin editor. If you try to pass it just as a tool, you will get a big no! And I would say you better make a hard copy of this before it is marked for speedy deletion.. Igor Berger (talk) 22:43, 5 January 2008 (UTC)
True, but I don't think the poll was meant that way. The poll is to vote on this specific proposal, and that's why many people are opposed (myself included). It'll be hard to convince everyone that the poll shows enough consensus for the fundamental idea that we can enable it, and discuss changes later. Equazcion /C 19:55, 5 Jan 2008 (UTC)
I strongly agree with ! != 2. There have been a lot of oppose comments that are entirely without merit, some of which seem to escape the bounds of reality. On the whole, non-admin rollback is supported. This proposal is a culmination of weeks of discussion and two other project pages in which community input was sought. Are there issues? Of course, but none that would require an entirely new discussion. That would simply be disastrous. --MZMcBride (talk) 19:58, 5 January 2008 (UTC)
I really don't think this poll can be considered consensus for this particular proposal. I think most of the opposers will probably say something along the lines off "I wasn't aware of any discussion". I'm pretty much saying that too. But, good luck convincing everyone...Equazcion /C 20:06, 5 Jan 2008 (UTC)
Any time anyone tries dismiss 100 wikipedians as being entirely without merit, they immediately lose the argument. Finding 'consensus' from a poll is not a process of electrocuting those that you find to be meritless. It is about finding a position that will not cause rift and division and being able to recongise whether and when a given position does or does not enjoy that status. If that means starting over to be genuine about it, then that is undeniably the best way to go. Consensus forms naturally; it is not cut with an axe. Take some advice from the supporters, opposers and the various discussions and see if you can't work out what would be a better, cleaner proposal. But please don't go around labelling such large numbers of people as 'entirely without merit' when they almost certainly are not. Splash - tk 23:24, 5 January 2008 (UTC)

(unindent)When wikipedians vote on things there will always be some comments that are "entirely without merit, some of which seem to escape the bounds of reality" - that's the problem with polls. However, I could equally say the same of some of the support votes, many of which are equally based on knee-jerks and have not read through the issues. Whatever way you skin it, there is no consensus for anything demonstrated here. I suppose I may be biased by my opposition, but then I suspect that we all are biased by our own views when reading consensus - that's why we look for uninvolved people to sum up debates and judge consensus. However, I doubt any neutral observed would fail to conclude that the results of this are looking like being similar to the poll of two years ago - and it was without consensus. But, I see no harm in letting this run for a week or two more. I suspect nothing much will change, but I could be wrong.--Docg 20:07, 5 January 2008 (UTC)

I agree, as far as many of the support votes being equally without merit. I wasn't here for the proposal two years ago but there seems to be enough interest in this that something will happen... but again I'm not sure if it'll be implemented the way this proposal states. Equazcion /C 20:11, 5 Jan 2008 (UTC)
I could be missing something, but I think the entire Support and Oppose sections have been archived. Isn't it a little difficult for newcomers to the discussion to express a view if there isn't anywhere to do it? MBisanz talk 20:37, 5 January 2008 (UTC)
Found it, my fault MBisanz talk 20:39, 5 January 2008 (UTC)

Consensus isn't about people agreeing to an outcome. It's about everyone agreeing to abide by the outcome (WP:PRACTICE). Many users who opposed this said they only did not support it because of a few things in the proposal or because of the way this would be put in practice. We've determined that many people want this, now we should attempt to find a reasonable method of putting it into practice and see what users think of it. .:Alex:. 22:03, 5 January 2008 (UTC)

Exactly my point. Equazcion /C 22:13, 5 Jan 2008 (UTC)
I'm unclear whether you are suggesting that this should be implemented now regardless of the content of the project page or not. If you don't currently have a 'reasonable method of putting it into practice', then surely further thought is needed until you do? Splash - tk 23:24, 5 January 2008 (UTC)
No. I mean that there is consensus to have rollback, provided that the method of putting it into practice is revised. We need to now discuss which way of implementation is the most effective and reach consensus on it. Basically we need to revise the proposal, so we now start discussions on new ways of implementing this feature that everyone can agree with rather than continuing the "Do we want it, yes or no?". .:Alex:. 11:06, 6 January 2008 (UTC)

(undent, reply to doc) The problem here is that many of the opposers have completely contrary ideas of what should be done. For example, some people are completely opposed to the idea and feel only admins should be given roll back. There has been extensive discussion and multiple reasons given why people feel non admin roll back is a good idea but they remain unconvinced. While they are welcome to their opinions, it's unclear to me what will change their minds. Others feel the process is too bureaucratic. Again, while they are welcome to their opinions, the problem here is alternative proposals like granting roll back to everyone or automatically after a certain time clearly do not have support of even a majority (at least the current proposal has the support of a large majority). Other people have proposed various technical solutions. Some of them might be possible and I don't think there's anything wrong with investigating them further. But some of them limit roll back to such an extent that they IMHO risk defeating the purpose. Others solutions are so complicated that it seems unlikely devs would be willing to implement them. Finally, many others are afraid of the potential consequences since they feel it will be too easy for people who accidentally or malicious misuse the tool to get it. Relating to earlier posts, there may be some technical solutions to allay their concerns but ultimately there is no way we are going to be able allay many since it's impossible to prove something won't happen. There is no way we can do a 'dry run' as some people have suggested. This is something that can be resonably easy killed off and I have seen no opposition to changing or removing the thing if it does spectaculory fail. But there is obviously no way we can guarantee this won't fail and the only real way it seems we can alleviate the concerns of quite a number is if we restrict roll back to admins and we get back to point 1. Remember it's not as if there has been no discussion before this proposal was reached. There has been extensive discussion and extensive tweaking. Also, I would argue it's not simply up to the supporters to come up with ways to improve this proposal. In cases such as this, when there is a very clear majority of people supporting the proposal, opponents need to describe what they feel is wrong and how they feel it can be fixed. You can't just tell someone well 'I don't like it, you need to fix it' especially when there has been extensive discusssion and multiple attempts to fix it already. Note I emphasise the solution part as well since if your proposed alternatives won't work for some reason or have even less support then the current proposal then clearly we're going in the reverse of where we need to go. You're right that it is wrong to simply discard all those opposed to the idea. But it is just as wrong to simply ignore those who support the idea (which in this case is at least 2/3 majority from what I can tell and it is a large number of people). In other words, it is inherently unfair if the only solution is to completely kill this proposal, without any alternatives. What I'm saying is that while I don't know if the current proposal is what we should ultimately start with (I emphasise start because while we should aim to have the best possible proposal implemented, it can be tweaked after implementation) from the way things are now I think we need to inherently accept that non-admin roll back will happen (unless the community has a sudden change of heart). Therefore while people are still welcome to be inherently opposed to the idea, their opposition is outweighted by the support. For those who do accept the idea but want to change it, we definitely shouldn't ignore them, but they need to have ways that work towards making the proposal better. Finally I emphasise one more time I'm not saying that we should ignore anyone's opinion, simply that when people consider the same thing and come up with different conclusions, consensus doesn't mean no one objects but is perhaps closest to what gets least opposition. (Also another problem is that when you keep things going for too long people eventually tire of participating on both sides and eventually you end up reaching a tiny 'consensus' which is far from what has the most support and usually ends up being very controversial) Nil Einne (talk) 11:35, 6 January 2008 (UTC)

Automatic granting

Some have suggested that the rollback right be automatically granted. With the current software it is possible to set it so that rollback is automatically granted after a user account has reached a certain age, has a certain number of edits, and/or has a confirmed email address. With a minor software change it would be possible to check if a user is or isn't in a specific usergroup. Unfortunately, it doesn't seem to be possible to manually remove autopromoted user groups. However, if the software change that checks if a user is not in a certain usergroup is done, it would be possible to "remove" rollback from a user by adding them to a usergroup for users who have had rollback removed (so its a bit of a hack). What would people think about this? Mr.Z-man 21:33, 5 January 2008 (UTC)

I like the concept, but I don't like the idea of having to cheat the software in order to remove it. --Charitwo talk 21:35, 5 January 2008 (UTC)
I'm not crazy about this because even autoconfirmed users can fall into immature battles with others. How long you've been here, the number of edits you've made, and the fact that you've confirmed your email address really say nothing about your ability to handle bad situations. I understand that it could be revoked, but until it is, users could do some very annoying damage with it. I've said this before: If the rollback links could be removed from the contribs list, and made so that people must look at least at the last edit before reverting, I would be all for this. But as it is, someone could keep refreshing a user's contribs list, instantly reverting every edit their "opponent" makes. I see this happening, and the annoyance that occurs prior to revocation is not worth it, in my opinion. But again, remove it from contribs lists, and I'd be all for it. Equazcion /C 21:55, 5 Jan 2008 (UTC)
Well, the threshold could be set much higher than autoconfirm and could require a certain edit count as well. I should also note that the "hackiness" of this method would mean that the developers would much less likely to implement this... Mr.Z-man 22:03, 5 January 2008 (UTC)

Grouping of 'votes'

I've just been looking through the discussion, and I've been attempting to work out why the 'votes' are grouped the way they are, does it make a difference which group you place a support or oppose comment in? Can anyone shed a little light please? -- Geoff Riley (talk) 22:27, 5 January 2008 (UTC)

Place your vote at the end of the last group. Votes have been split up, 20 votes to a group, for ease of editing, just so that they're not all in one huge section. Equazcion /C 22:28, 5 Jan 2008 (UTC)
Ah! Thank you so much.... hmm... okay, it's obvious when I think about it, but it had me confused. Cheers. -- Geoff Riley (talk) 22:34, 5 January 2008 (UTC)
No problem :) Equazcion /C 22:34, 5 Jan 2008 (UTC)

Archiving votes

What's the point of archiving "the first 100/60 votes"? Why not just put the whole poll on a separate page? —Ashley Y 00:03, 6 January 2008 (UTC)

The whole poll is pretty long itself. The proposal page is 99% poll. Archiving some of it seemed to be the only way to shorten the page into a more manageable size. It was so long that there was getting to be a delay in loading it (it approached 300k). Equazcion /C 00:11, 6 Jan 2008 (UTC)


Responding per request

Hi, I received a request to come over and comment here.[1] Non-admin rollback has been discussed before, if I recall. As someone who's walked on both sides of that road, rollback is actually the administrative tool I miss the least. Maybe if I were a regular new changes patroller this would be different. I tend to step in when certain pages turn up on my watchlist and if there's a problem I address other recent edits by the same account or IP. It takes about two seconds longer to undo vandalism without rollback, and at the rate I address the problem that's really not a hardship.

From what I've heard, this proposal got raised again in poll form during the holidays. That strikes me as a not very good way to raise a basically viable idea. I'm glad to see the editors here have archived that poll and posted a notice to the community bulletin board. If you haven't already filed a content-based WP:RFC that would be good also. Thank you very much for asking my input; I'm flattered. Best wishes, DurovaCharge! 00:09, 6 January 2008 (UTC)

Not sure I understand

1. Where I read in FAQ/How does rollback differ: it makes reverting all the edits made by a given account or IP address relatively simple - does this mean that a single button would revert ALL the top edits made by an editor, or the button has to be pushed article-by-article?

2. If I want to revert a series of edits to a page, can't I already just open the history and click the bullets to encompass a range of diff's, then Undo them as a group?

I'm not sure what functionality is being added and I'm not especially buying the "reduce server load" argument either. Can someone help me out? Franamax (talk) 00:27, 6 January 2008 (UTC)

1 - It needs to be done article-by-article.
2 - No, the undo link will only revert a single edit.
The reduction of server load is in comparison to script-assisted rollback methods, such as Twinkle.
Equazcion /C 00:29, 6 Jan 2008 (UTC)
Its also faster for people with slower internet connections than scripts. Mr.Z-man 00:32, 6 January 2008 (UTC)
I'll buy the slow-connection arg, for sure. But I'm confused at the answer above to my #2 - I just made 2 edits in my sandbox and undid them both with one step. Am I missing something? Franamax (talk) 00:36, 6 January 2008 (UTC)
I don't see the standard undo edit summary in your last edit. Did you remove it and replace it with the word "test"? Equazcion /C 00:38, 6 Jan 2008 (UTC)
Okay now that's a surprise. You're right, if you load multiple diffs you can undo them all at once. However, rollbacks can accomplish this in a single click, while this undo method requires multiple steps. Rollbacks can also be done via a link in a users contribs list -- you don't even need to load the article, let alone its history or the diff screen. Equazcion /C 00:41, 6 Jan 2008 (UTC)
e/c'd and lost my answer. Yes, you get a blank summary on group undo.
I prefer the option in any case to add my own summary rather than an auto-summary a la TW or popups, I often ;) have something to say :)
And the capability to sit at a special-contribs screen and blam away with rollbacks, to me is especially scary. What about rotating IP's? Usually you can see a gap between the bad instances and the useful edits from the previous IP lease, but only if you're looking. I'm worried about seasoned but jaded vandal-fighters overdoing it, then we get into the argument over whether it's more important to protect the innocent or prosecute the guilty. The danger is subtle overuse of the tool, which might end up discouraging IP editors who might eventually join up, and this would be very hard to detect.Franamax (talk) 00:55, 6 January 2008 (UTC)
Agreed, all-around. Just one note: TW allows custom edit summaries in its rollbacks. Equazcion /C 00:57, 6 Jan 2008 (UTC)

Again with the concerns on the voting

As the risk of beating a dead horse, I wanted to re-raise some of my concerns with the straw poll. • The thing people are voting for/against has changed since the voting started. Look at the proposal at the first vote (diff). A ballot question shouldn't change during the voting. • While the poll is spilt into "Support" and "Oppose" votes, not everyone in "Support" seems to be supporting the proposal as written, and not everyone in "Oppose" seems to be against the general concept. This makes simply counting votes especially misleading. • My original response to the poll did not fit cleanly into the category of "Support" nor "Oppose", so I created a third section, "Rejection of dichotomy". That got moved into a "Discussion" section, and ended up archived at Wikipedia talk:Non-administrator rollback/Archive 1#Rejection of dichotomy. I noticed this, and so put down a concise "Oppose" vote. But the occurrence left a bad taste in my mouth. If this was a "real" vote, and I hadn't noticed and voted again, that would amount to (unintentional) vote tampering. Fortunately, it's a non-binding straw poll intended to encourage discussion. • I really do believe everyone has been working in good faith here, but I think not always towards the same immediate purpose. That can lead to confusion and misunderstanding. • Anyway, enough out of me. Happy wikiing! —DragonHawk (talk|hist) 00:40, 6 January 2008 (UTC)

I don't think anyone is going to close this by merely counting votes. I think all of the points you raise will be taken into consideration. Equazcion /C 00:45, 6 Jan 2008 (UTC)
PS I archived your "Dichotomy" discussion. If you'd like to restore it please feel free. Equazcion /C 00:47, 6 Jan 2008 (UTC)

What's the big deal?

I still don't understand why this is such a big deal. Any user can Undo multiple edits today. On the History page, I click a "cur" button from several edits earlier or I select a beginning and ending edit and compare selected versions. Then after examining the cumulative differences - a good thing to do when reverting multiple edits - I click Undo and the multiple edits are all gone. And I have the opportunity to add an appropriate edit summary. All in all it is more flexible then rollback seems to be, lets me view the group of edits before deleting which rollback does not, and lets me add an edit summary which rollback does not.

I don't see that rollback gives me anything that I can't do today. Is it a few seconds faster? Does it save a few microseconds of server time? As a vandal fighter what takes my time is putting the warning message on the user's page, not reverting the edits. And all too often I see other editors reverting vandalism without posting a warning message. So rollback seems to be solving the wrong problem.

Just out of curiousity could someone post some screenshots of how rollback works? Sbowers3 (talk) 15:28, 6 January 2008 (UTC)

Archiving opinions is evil

Wikipedia:Non-administrator rollback/First 200 support votes and Wikipedia:Non-administrator rollback/First 80 oppose votes begin with the archive template, which says the pages should not be modified. Does that mean someone who supported or opposed the proposal cannot change his opinion? I think we should have all the opinions back in the main page, as we have done with all the proposal since the beginning of time, regardless of size. That makes easier to switch opinions back and forth, and to bring discussions to previously posted ones. If you tell me nobody changes opinion once given, I would not be against it, but there are examples of people changing sides just because someone corrected a mistaken thought. -- ReyBrujo (talk) 17:35, 6 January 2008 (UTC)

Got rid of the archive template. Equazcion /C 17:37, 6 Jan 2008 (UTC)
Yeah, that is the simple solution. But the fact that you can't give opinions on previously posted opinions stands. That is what "discussing" and "searching for consensus" means. -- ReyBrujo (talk) 17:42, 6 January 2008 (UTC)
this decision MUST NOT go threw. while most people dwouldnt not abuse there power some of the new people who recieve this power will use it accidentaly to cause edit wars that they dit not predict. i fear that the end of any possibility of neutrality on psychic and alternativative science -related articles could ensue of this power goes through without being carefulyl controlled by admins. Smith Jones (talk) 17:44, 6 January 2008 (UTC)
The page would be very large and delay loading if all the votes were put back on the main page, especially for people with slower connections. And I think this poll is so massive that no one is going to pay much attention to the older votes, and will likely only comment on the more recent ones. Besides anyone is still free to read or comment on the votes that have been split off. Equazcion /C 17:47, 6 Jan 2008 (UTC)
Since it was archived with a copy paste and not a move, only the people who edited those archive pages have them on watchlist. People will never know someone has commented on their opinion unless they are informed in their talk page, which is something that is not usually done when everything is in a single page. -- ReyBrujo (talk) 17:58, 6 January 2008 (UTC)
With all the activity on the main page, no one is going to know anyway. The watchlist is almost useless for a page with that much activity. Equazcion /C 18:01, 6 Jan 2008 (UTC)
A page that takes 4 minutes to load is evil, archiving is standard procedure on long discussions. This is a non-issue. 1 != 2 19:00, 6 January 2008 (UTC)
I guess so. Can you point me another where the opinions have been archived while the discussion was still going on? -- ReyBrujo (talk) 19:22, 6 January 2008 (UTC)
I can't offhand, but it shouldn't matter whether it's been done before, long as it works now. Equazcion /C 20:00, 6 Jan 2008 (UTC)
I would recommend you to make reference, in a clear and unimbigous way, to the discussion at the top of the article.
Wikipedia:Non-administrator_rollback#Discussion Igor Berger (talk) 20:20, 6 January 2008 (UTC)
Archiving polls during mid-poll is standard procedure? I'm not aware of any examples; could you provide some? I am, however, aware of a counter-example: Wikipedia:Attribution/Poll. --Iamunknown 21:54, 6 January 2008 (UTC)
Archiving any discussion page, regardless of its content, when it gets to long to reasonably load is normal. Those pages are not being hidden, just put off to the side so we can talk here with reasonable page load times. I really don't see what the big deal is here. 1 != 2 22:03, 6 January 2008 (UTC)
No, it is not for polls, because they may influence results, as it makes harder to update modifications and continue discussions, and segments the discussion itself in three different pages, one of which is not watched by anyone. As I said, show me a single poll out of the hundreds we have had where the results had been archived before the end. That is why we have subsections, to ease editing. There is no need to remove all opinions to make the page load faster. It has never been done. -- ReyBrujo (talk) 22:12, 6 January 2008 (UTC)
(ed con) I suspect, because this was labeled a poll, it's being seen differently than a mere "discussion". Some people will inevitably have the very human reaction that the archiving diminishes their opinion. And archiving part of a poll, while the poll is still going on, is a rather unusual action. (I cannot say it is without precedent, since I haven't been witness to every poll on Wikipedia.) I do think the page size was becoming unmanageable, though. I'm not really sure what the right thing to do would be. I suspect this will be an increasingly bigger problem as Wikipedia continues to grow -- polling via editing wiki pages isn't going to scale well. But this talk page isn't really the right place to tackle that problem. —DragonHawk (talk|hist) 22:20, 6 January 2008 (UTC)
(to Rey) A poll is just a different format for discussion. As everyone keeps pointing out, it's not a vote. In fact this particular poll is even less formal than an RfA. No matter how you slice it, this is still a discussion, and one that got very long. If there's a better way to keep the page at a manageable length, please clue us in. Equazcion /C 22:37, 6 Jan 2008 (UTC)