Wikipedia talk:Block protocol

From Wikipedia, the free encyclopedia
Jump to: navigation, search


Discussion of 18 November version


For your consideration. Gerardw (talk) 21:17, 18 November 2011 (UTC)

  • I'm confused, what if anything substantive this adds to WP:BLOCK that isn't covered. The purpose appears to be a proposal to make block logs subject to oversight. (is it even technically possible to hide block log entries without a software alteration?) Although I understand that misinterpretation of the block log can have long range negative consequences, as long as unblocks are carefully worded and dummy blocks are used as necessary to update the log, this seems like a solution in search of a problem. If this is addressing a wider concern of which I am unaware, forgive me, otherwise this looks creepy to me. Crazynas t 21:52, 18 November 2011 (UTC)

This is in response to Wikipedia:Arbitration/Requests#Unblocks_and_enabling and Wikipedia:Ani#User:WebHamster_block_and_unblock.3B_possible_wheel_war_-_leaving_it_to_the_community_to_judge_me_and_others. The issue of past blocks is a recurrent, and usually irrelevant, theme at WP:WQA. Gerardw (talk) 22:05, 18 November 2011 (UTC)

Thank you for the relevant reading, I'm still confused why this needs a separate page (as opposed to an amendment to official policy) however. Crazynas t 22:17, 18 November 2011 (UTC)
Me too. You might want to look at some of the contemporary discussion at Wikipedia Talk:Blocking policy. In case this is the intention (though I'm not yet sure that it is), we do not want to make blocking policy more prescriptive, and we definitely don't need a "prescriptive fork" of policy. causa sui (talk) 23:41, 18 November 2011 (UTC)
  • Expunging the record would require software changes. Rich Farmbrough, 00:28, 19 November 2011 (UTC).

I started a separate page for the purposes of discussion -- if the proposal is accepted, it could be merged into the existing one, if it fails, it's moot. Gerardw (talk) 00:46, 19 November 2011 (UTC)

What is the rationale that this rises to the level of suppression policy currently the closest policy that covers it is #4 "Hiding of blatant attack names on automated lists and logs". I don't believe that (even if I accept for the moment the necessity of removing entries at all) it should rise to the level of suppression. If we are going to remove entries they should at least be reviewable by other administrators no? I agree with DGG that this is a corner case, (most blocks are not overturned, most overturned blocks are not controversial, something that effects .01% of editors should not be policy). However as in my reply to Malleus bellow, I feel that the intent of this proposal could be achieved by making block log apologies no longer optional (at least in cases where clear consensus is found that the block was unwarranted). Crazynas t 21:27, 24 November 2011 (UTC)
  • I do not think it represents the consensus at those discussions linked to above. I certainly do not think it was agreed that "a previous administrator statement to the effect that a block isn't appropriate shall not be considered an "action" in the sense of wheel war reversal" What I think was agreed was in this particular situation the block was a good block and should not have been unblocked. Most of the discussion concerned these special aspects. It is a very poor idea to make elaborate codified provisions on the basis of a special situation--it leads to unpredictable results and can create more disturbances in practice than anyone realizes at the time. That's why we have IAR, so we don't have to do this sort of general proposal. The admin involved knew and said he knew he might have done what amounted to a technical wheel warring violation and came just as he should have to ANI to discuss it. (and his decision was widely though not uanimously supported, not on the technicalities, but on the basis of the best way of handling the specific situation. DGG ( talk ) 01:40, 19 November 2011 (UTC)
I understand what I am proposing is not the current consensus. The current consensus is -- there is no consensus. Please see Wikipedia:Arbitration/Requests#ball Gerardw (talk) 02:25, 19 November 2011 (UTC)
  • I'm not convinced. Some said the block was bad and the unblock was bad. Some said both were fine. Some seemed to think one or other or both were dubious without actually being "bad" or "good". I do agree strongly with your comment " It is a very poor idea to make elaborate codified provisions on the basis of a special situation" I would even simplify to "It is a very poor idea to make elaborate codified provisions". Rich Farmbrough, 23:27, 19 November 2011 (UTC).
No: As a regular editor (and hopefully future admin candidate) this proposed policy still gives the first "positive" mover a significant advantage. If a complaint comes up, several administrators decline to block, and then one rouge-admin comes in and evaluates it the other direction it's now become a case of the community consensus not being respected. As you said in the essay, the block log is a permanent brand of shame. Unless someone from meta sneaks into the database to obliterate the block log record, it will always show up on the editor's permanent record (which has been used against them when applying for advanced privileges). If the editor is truly being as disruptive as they are claimed to be it should be blatantly obvious (and equally easy to build a consensus for the block) and/or they should have the 4 full warnings. In short, I feel this has some good clarification on policy, but makes an individual admin's action more important than notices of admins declining action rather than "sysop is not a big deal". Hasteur (talk) 14:16, 19 November 2011 (UTC)
Thanks for the feedback. It's not intended to change the criteria for when a block is applied -- I've added an FAQ above to try to clarify that and the software situation. It's not meant to say a "block" vote overrides a consensus of "not block," it's meant to provide an orderly protocol for Wikipedia to navigate the period from initial assessment by the first few admins and the final consensus. Gerardw (talk) 16:53, 19 November 2011 (UTC)
Right Question, Wrong Answer: I think "Accidents of timing, specifically which administrator evaluates a case first, should not be used to guide policy." is a worthy goal, and would like to see a proposal that fixes that particular problem. Alas, I don't believe that the present proposal will meet that goal. No policy is going to change the basic fact of human nature that we are a social species and that we will pay attention to the actions of others who have accepted or declined before us. The paying attention to what others do part is a Good Thing - it's the fact that first responders get more attention that is a problem. The only workable countermeasure to this sort of problem would be some sort of scheme that prevents you from seeing what others decided to do until after you commit to your own decision, which may very well introduce a new set of problems. --Guy Macon (talk) 17:24, 19 November 2011 (UTC)
  • Drive-by comment. Not really up on all the details of this, so just a couple of comments.
Re third principle, "A transiently inappropriate block does not do significant harm to Wikipedia". Don't agree and cannot support this in this form. Any block must be assumed to be permanent, as probably most people, on being blocked once for any duration, probably walk away from the project, see WP:HURTS for exposition on this. This does not apply to people playing Wikipedia as a MMORPG and so forth but it's probably true of most editors.
Moving on to the meat of the issue... this involves Malleus Fatuorum, right? Well, Malleus Fatuorum is kind of sui generis I guess, and as they say in law, "hard cases make bad law". Not to sure about the details, but it might be a good idea to make a special policy for Malleus Fatuorum. This would be highly unusual, but we have dealt with special cases in special ways in the past I think. Stating as policy something like "Malleus Fatuorum may not be blocked except on the assent of three admins" or conversely "Malleus Fatuorum, if blocked, may not be unblocked except on the assent of three admins" or something (not really up on Malleus Fatuorum or the details of the situation, so whichever solves the problem).
If we don't want to make a policy naming a particular editor, then achieve the same effect by creating a class of editors, like this: "Absent emergency such as suspected compromise of the account etc., a user with 100 points may not be blocked except on the assent of three admins. Points are accrued as follows: X for a GA, X for a FA, X for DYK, X for each 10,000 edits, X for each year of service, -X for each previous block, etc." or whatever.
This would possibly be a good idea, but are unlikely to be accepted. So we are left (if I understand the situation rightly) of grappling with a way to write policy so as to deal with with a few very exceptional cases like this. I submit that this is not possible to do well so would tend to support no change to policy here. Herostratus (talk) 18:29, 19 November 2011 (UTC)
It would be my hope that the duration of inappropriate blocks would be short. Under this proposal, unlike current practice, the inappropriate blocks would be removed from the block log. It's not intended to be about Malleus, it's intended to make what is happening now happen more smoothly. As far as Malleus, I'll note that if this proposal was adopted, Malleus block [[2]] would be much shorter with the overturned blocks removed. Gerardw (talk) 19:07, 19 November 2011 (UTC)
OK. I don't really want to interfere if it's a technical matter of making things run more smoothly. I'm not super keen on messing with block logs (except in extraordinary cases and by going through an involved process) because they are historical facts. They happened. People then can make the case that particular ones were uncalled for and should be discounted. However, I see the point about someone using them to game the system, I guess, so I dunno. I have one question though: if an admin makes a number of questionable blocks, and these are expunged from the record, wouldn't it be harder to call that admin for making questionable blocks? Herostratus (talk) 20:20, 19 November 2011 (UTC)
Please don't consider commenting interfering, hashing things out is part of the collaborative process, and will only make the proposal better. My initial draft (in my user space sandbox) did start to get into separate views of the logs, but I felt that was making things overly complicated. It appears to me many of the editors who frequent WP:ANI have fairly good memories, so I think that the issue would self correct. I do assume that some users will have access to the full log -- similar to WP:CHECKUSERs.Gerardw (talk) 20:31, 19 November 2011 (UTC)
Hiding inappropriate blocks is another phrase for "brushing it under the carpet". Parrot of Doom 20:49, 19 November 2011 (UTC)

A question, in the Webhammster case there was a block an unblock and then re-block, if the unblock (along with the block it was unblocking) was removed from the list would an admin still be able to re-block? Also if an admin reversed a nunmber of blocks and they are removed how could you detect a patern of unblocking?Slatersteven (talk) 21:05, 19 November 2011 (UTC)

If the proposed protocol was followed, there would have been a block, a discussion to reach consensus and the block would have remained in place. Gerardw (talk) 22:13, 21 November 2011 (UTC)
Re "[S]ome users will have access to the full log -- similar to WP:CHECKUSERs", I still have a question, although not too likely to come up much. If I'm watching an admin with an eye to making a case against him if his behavior becomes too problematic (and I do do that), then I want full access to his full blocking log, especially any problematic blocks. Herostratus (talk) 21:26, 19 November 2011 (UTC)
Its oonly an assumption, I would rathr see a concreate commitement to the idea that ther will be a kind of Blockchecker.Slatersteven (talk) 22:08, 19 November 2011 (UTC)
How about the following: An admin RevDels the "Action and target" field in the block log. This way, there's a significant number of users who can confirm that there are no problems; several users can see the full block log if necessary, although these entries will be clearly marked as "wrong"; it's easy to fix mistakes or abuse; no new features need to be developed - only use of the old ones. עוד מישהו Od Mishehu 17:49, 21 November 2011 (UTC)
I guess. My question is: is this a problem? I understand -- if I'm getting this right -- that in the particular case of the particular editor Malleus, he has a lot of blocks because, if I'm following the argument, some admins don't like him or something, and the claim that adding another, which apparently was unjustified, just makes him look worse and thus more likely to blocked in future. OK, but aside from this one editor, is this a general problem?
I have been blocked twice, and one was a simple mistake, akin to the person typing in the wrong username. It'd be OK with me if this was expunged, but on the other hand it doesn't much matter. If I had 12 blocks, and should only have 11 in the log because this one was unjustified, I guess I would like that one expunged so I'd look a little less bad, but I would think that in that case I should spend more energy cogitating on the 11 blocks that were justified.
So I'm just asking, is this a solution in search of a problem? And if unjustified blocks are a widespread general problem, I wonder if we shouldn't put more energy into addressing that -- tightening up the guidelines or procedures or whatever -- rather than worrying about block logs. Herostratus (talk) 18:08, 21 November 2011 (UTC)
Arbcom seemed pretty clear in suggesting an RFC for reasons (FAQ #4). Inappropriate blocks, in my opinion, are a separate issue, but one that is not easy to solve. Gerardw (talk) 22:13, 21 November 2011 (UTC)
Unjustified blocks are clearly a widespread problem for anyone with eyes to see it. And apologies after the event just don't cut it, as the log looks just the same as before; the apology isn't logged Malleus Fatuorum 00:09, 22 November 2011 (UTC)
Would mandatory recorded apologies be sufficient? Something along the lines of "Very brief blocks may shall be used in order to record an apology or acknowledgement of mistake"? Crazynas t 21:27, 24 November 2011 (UTC)
It would make the present situation even worse IMO. People don't actually read the block log, they just look at its length. Malleus Fatuorum 22:21, 24 November 2011 (UTC)
Seems like a reflection on their competence more than anything(also, see my current rant about not reading on Wikipedia), however, I agree that it's a problem. Maybe we should have an RFC about making competence a requirement</joke>.(that would be rather humorous to read, blocked: (incompetent))In all seriousness I think that oversighting the logs is a bad idea, and mandatory recording is a step in at least letting those that really want to know that the block was in error (while still retaining the record)Crazynas t 22:58, 24 November 2011 (UTC)
  • No As I stated in my comments before, the principle of symmetry is better than this attempt to codify a deliberate asymmetry oin dealing with blocks or other decisions. To wit:
Any decision by any administrator, posted as such, whether to block or 'not' to block, to undertake a specific action or 'not' to undertake a specific action, shall not be reversed by any other administrator without full discussion, and subject to clear consensus.
Is my specific proposal, and I suggest that this be what is adopted. Cheers. Collect (talk) 23:59, 21 November 2011 (UTC)
Since we already have technical changes to the software on the table, how about this: Whenever there is the slightest chance that a block will be controversial (automatically satisfied for civility blocks and for editors who were unblocked for whatever reason in the recent past), admins may not block but may only leave a new "block request" entry in the block log which will also change the user's editing interface in an eye-catching way. (More so than the orange bar, which some editors manage to overlook.) Then a different admin can confirm or reject the block request. Rejected block requests and confirmations are not displayed in the default view of the block log. Hans Adler 01:03, 22 November 2011 (UTC)
While I agree with the Collect's concept as a policy, as a protocol is is problematic. This policy is, as I understand it, essentially the current policy and subject to Race conditions. That is to say, there could easily be an edit conflict between blocking and not blocking. Blocking and not blocking are inherently asymmetric:
* There is a single, unique location where a block is logged. A user may be blocked for multiple reasons ... 3rr, ANI and SPI come to mind. Therefore admin statements regarding not blocking a user may be found in multiple locations, making a missed "not block" statement likely to lead to good faith violations of the policy.
* A blocked editor produces a static contribution state for evaluation; an unblocked editor can continue to make changes while consensus is being achieved. Therefore editors making comments later in the discussion will be discussing a different situation than earlier commenters, and it will be unknown whether an earlier commenter's viewpoint would have been different based on the updated situation.
Additionally, such a policy, in practice, encourages admins to make quick decisions to be "first to the punch." This is especially true because the proposed wording, "clear consensus" implies that in the case of a split decision, the first actor's decision will stand. The impact of one admin's judgement should not be greater than the other 1500. Wikipedia is better off with thoughtful, considered decisions.Gerardw (talk) 14:57, 24 November 2011 (UTC)

Wouldn't we be reducing transparency if we start purging overturned blocks, and wouldn't it act as a double-edged sword, i.e. there would be no record of the original blocking admin making the block? –MuZemike 23:34, 23 November 2011 (UTC)

For me, the primary goal is to improve the smooth, effective functioning of Wikipedia as a whole. There would still be a record of discussions surrounding the block that could be located. For me, removing the stigma from the in hindsight inappropriately blocked user is more important than ease of gathering information regarding the administrator. Gerardw (talk) 15:00, 24 November 2011 (UTC)
This is a tricky one, and we're never going to get a "one size fits all" answer, because we are all individuals, and some of us are very different indeed ... . I don't like the idea of making what would effectively be a separate set of rules for a particular class of editor, at all. It would be too easy to game the system and start pushing an editor you don't like into that class; yes, it would take patience, but sometimes someone with an axe to grind can be quite unbelievably patient. I like Collect's idea in principle but agree with Gerardw's "race conditions" point, too. Hans Adler's idea of a block-request option is good. How about the idea that an editor with a long string of questionable blocks could ask for that string to be checked for "reasonableness", and the dubious ones moved to a "supplementary block log" which could also be seen, but wouldn't have the immediate eye-impact of an arm's length of "previous convictions"? Pesky (talkstalk!) 10:15, 25 November 2011 (UTC)
These are all ideas worth considering, and another would be to keep questionable/disputed blocks in the main block log but mark them as such (with an icon and different color for instance). The advantage here is that if an admin has many questionable/disputed blocks this would also stand out in his blockining log. Herostratus (talk) 18:07, 25 November 2011 (UTC)

A question (sorry if its been asked already) Some un-blocks are adminustrative in nature (they are for 'time served, or becasue the user has agreed to behave) not due to mistake or bad timing (or malicious blocks). As I understand it these would also be removed correct?Slatersteven (talk) 12:27, 27 November 2011 (UTC)

No. Wikipedia:Block_protocol#Types_of_block_removal classifies two types of blocks, and only the overturned blocks would be removed. If an editor receives a block supported by community consensus, it would remain in their log. Gerardw (talk) 12:54, 27 November 2011 (UTC)

Default is block[edit]

Er, I could not disagree with any section that I've ever seen in an essay more than I disagree with this. If a block is contentious, then the default should be to UNBLOCK. First, we have the right to address our accusers. If a person is blocked under questionable circumstances, they should have the ability to respond and to address the criticism. Second, there is the principle of innocent until proven guilty. Leaving a questionable block in place assumes guilt before innocence. Third, Unblocked is the original status. If consensus doesn't exist for a change, then it should remain in the original status---which is unblock. this section is so egregious that I discount the rest of the entire essay.---Balloonman Poppa Balloon 22:38, 4 January 2012 (UTC)

Which suggest the essay is poorly written as the primary point, which is to eliminate this so-called "first" and "second" mover conundrum, isn't coming through. I've explained the logistical reasons for suggesting "block" be the default state; a temporally independent protocol where "unblock" was the default state could work as well if the nuts and bolts could be worked out. My biggest concern, based on my observations of past discussions, is that the editor under discussion frequently escalates the situation during the discussion, so editors who comment earlier are looking at a different set of edits than those later in the editing sequence. Nobody Ent 22:52, 4 January 2012 (UTC)
I have no problem with the editor escalating the situation during the discussion. If they want to hang themselves, give them the opportunity to do so---it will take away the contentiousness of said block. If a person is dumb enough to escalate the issue during the discussion, let that come out. It's kind of like the person who gets overly defensive and rude during their RfA. While people might have been on the fence to begin with, by the end of that type of RfA, they are opposing. If a block is bad, then it needs to be lifted until consensus says otherwise. This gets rid of the so call first/second mover advantage, but maintains the status quo.---Balloonman Poppa Balloon 16:30, 5 January 2012 (UTC)