On 31 May 2010 19:46, Abd ul-Rahman Lomax abd@lomaxdesign.com wrote:
These are issues that I've been thinking about for almost thirty years, and with Wikipedia, intensively, for almost three years specifically (and as to on-line process, for over twenty years). So my comments get long. If that's a problem for you, don't read it.
... Has it really not occurred to you that *you're* trying to convince *us* of something? In which case, conciseness is likely more useful than defiant logorrhea ... Oh, never mind.
- d.
The key is not making it easier to remove adminship. This proposal gets us closer to the real problem, but fails to fully perceive it as does the common call to separate the functions of adminship.
The real solution to the current (and relatively long-standing) problems with RfA and adminship in general is the marriage of the "technical" side of adminship with a "political" side, which is rarely acknowledged. Successful reform will involve separating these two aspects, rather than the more common idea to separate some technical pieces from others. The proposal below is a bit lenghty, but it's the product of years of thought, and I encourage you to read it. If you don't have the time, well then, the take away point is that we should create a distinction between those administrators trusted to intervene in highly-controversial areas and those not so trusted.
The technical bits of adminship are, indeed, no big deal. With a large community of administrators and an alert body of stewards, the possible danger of obvious abuse of the administrator privileges is nearly zero. As an illustration, in the heat of the recent dust-up on commons, an administrator there "went rogue" and vandalized the main page. His edits were reverted in less than a minute: http://commons.wikimedia.org/w/index.php?title=Main_Page&action=historys.... Even in an absolute worst-case scenario of administrator abuse (for example, vandalizing the main page and then deleting a large number of pages with just less than 5,000 revisions in an attempt to lock the servers, especially abusive shenanigans in the MediaWiki namespace, or inserting malicious code into monobooks), the damage done would be reversed in under 10 minutes. Given this, it is highly improbable that any vandal/banned user would attempt to gain administrator status solely for the purpose of carrying out some such abuse. The danger comes from a compromised account or a higly disaffected administrator, and neither of these possibilities can be headed off by any level of standards at RfA, however high.
Why, then, has adminship become a big deal? Because in addition to the purely technical functions of adminship, administrators also have a political function. Administrators are often compared to janitors, but the metaphor is highly flawed. Janitors empty the wastebins, but they don't decide what should go in them. Many of the functions of adminship do not carry a significant political component: blocking obvious vandals, most instances of speedy deletion, fixing cut and paste moves, deleting old userpages, straightforward AfD closures, etc. are simple instances where a trusted user is needed to perform a technical function.
On the other hand, there are cases were administrator functions become highly charged and political - in closing controversial AfDs, blocking in many 3RR situations, and above all, in cases where some sort of intervention is necessary against well-established users who have engaged in some sort of unacceptable conduct. In these cases, the role of the administrator is fraught and ambiguous. He is faced with highly political choices about how to judge consensus, what course of action to take, etc. It is customary for relatively new and inexperienced administrators to stay out of these situations and leave the decision up to an administrator who has more experience and, for that matter, for political weight within the Wikipedia system.
The problem, though, is that there is no formal guidance of any kind as to who should actually make such decisions. From a policy perspective, an administrator sysopped last week has the same standing as someone with years of service. More importantly, a long-standing administrator with a reputation for more questionable judgment has exactly the same standing as a long-standing administrator with a reputation for impeccable judgement. There is no drawn by the community, except in the various most informal way, to separate administrators who should intervene in highly controversial situations from those who should not.
It is intervention in the highly controversial cases that causes problems and allegations of abuse. Our concern is, or at least should be, primarily in who is making highly controversial administrator judgements and on what basis, not who is carrying out F5 speedy deletions or blocking obvious vandals. Concern over these highly controversial judgements, because there is no line separating those administrators who engage in them from those who do not, is what has driven steadily escalating standards at RfA. We are less concerned that a newly-appointed admin will prematurely block a vandal without any warnings tomorrow, than that he will, in 12 months, block a well-established user for the wrong reasons after a heated debate at ANI. In other words, the problem is that RfA is being asked to make a judgment that should not be made at RfA.
What we need, then, is not a way to desysop more easily, but rather a way to delineate highly-charged and controversial administrator actions, and the administrators qualified to perform them, from uncontroversial administrator actions, and the administrators qualified to perform them. I will not presume to provide a full criteria for what separates controversial from uncontroversial administrator actions, but I would suggest something along the lines of the following. Controversial: Arbitration enforcement actions, blocks of established users for any reason other than suspicion of account compromise, close of AfDs where the consensus is not clear (this of course becomes itself a murky distinction, but could be well enough set apart), reversal of the actions of another administrator except when those actions are plainly abusive. Non-controversial: All others.
As for deciding which administrators are qualified to make decisions in the most controversial areas, I would suggest that we already have a group of people, the bureaucrats, in whose judgement the community has expressed particularly high confidence. I would propose that the bureaucrats become the group who are expected to undertake the controversial administrator actions; this would almost certainly entail some expansion of the current bureaucrat pool, but personally I like the idea of tying the controversial administrator actions to the ability to promote administrators - it underlines their seriousness, and at present, the bureaucrats do not have many functions. If, however ,the community is unwilling to combine the two groups, another group, say "sub-bureaucrats" could be created, but I must emphasize the importance of a bright-line distinction between those administrators trusted to perform highly controversial tasks and those not trusted to do so. Obviously, the ordinary administrators would still have the technical ability to intervene in the highly controversial areas, but doing so would obviously entail serious consequences or desysopping.
This brings up a final point: the issue of administrators with insufficient knowledge to appropriately follow policy on, for example, speedy deletion. I firmly believe that if we separate the political and non-political aspects of adminship, this becomes less of an issue. While an administrator taking the wrong course in a controversial area is akin to a janitor, who is empowered to decide what to throw out, deciding to throw away your important papers because he doubts their importance, the mistakes of lack of policy knowledge and inexperience are more like a janitor who, because he doesnt' know any better, throws away the recycling and attempts to recycle the rubbish. The second category of mistake is more easily rectified. The old idea, of some sort of mentoring for new administrators, does nothing about the political aspects of adminship (making controversial decisions) which is why it has failed in the past, but it is a perfect solution to the problem of inexperience/ignorance. New administrators who do not have a full grasp of the speedy deletion policy, or the blocking policy for vandals, or the criteria for granting autoreviewer status would be encouraged, perhaps through a formal process, to get up to speed on those areas by a more experienced mentor. If we carry through this proposal, there is every reason to believe that the crowd at RfA would be much more willing to promote more candidates and the process would become much less grueling. Our shortage of people to perform technical tasks could be easily reduced, if not eliminated.
This proposal is not process creep or the introduction of unneeded bureaucracy. It is also not an answer in search of a problem. There is clear acknowledgement that we have a problem, and this solution is a minimalist one. As I have proposed it, it simply takes advantage of an existing process (RfB) and group of users (bureaucrats) and would require only minimal amendments to policy, setting aside those areas of administrator conduct that are highly controversial and requiring that only bureaucrats act in those areas.
Administrators differ in competence, and perhaps even in trustworthiness, but I think experience has shown that not even the most experienced and trusted of all will always correctly interpret the view of the community, and that nobody whomsoever can really trust himself or be trusted by others to be free from bias. I see no reason to think that the long-term administrators are any more likely to show neutrality or a proper self-perception as the newer ones. If anything, they are more likely to have an over-extensive bview of the centrality of their own ideas. Consequently, I think there is no other basis by which any administrator can make a decision except by consensus, implied or express . For those who are willing to read beyond the first paragraph:
in general I do not think it is the business of the closer to decide between conflicting policies. Their job is to discard arguments not based on any policy, or, sometimes, by SPAs, and then judge consensus. The questions asked at RfAdmin are enough to identify admins who know enough to tell what is policy and what is not, as long as things don't get too complicated. It is not enough to identify admins who understand all policies well enough to judge which of conflicting ones to apply, or how to interpret them in difficult situations. A good thing, too, or we'd have chaos, because none of us agrees for all of that. The only people here competent to judge conflicting content policies or how to interpret them are the interested members of the community as a whole, acting in good faith. It is by the community's express consensus that BLP and Copyright trump other policies if the situation is unambiguous. But how the BLP and copyright policies are to be interpreted and applied in any particular instance is a question for the community, not individual administrators.
The assumption in closing is that after discarding non-arguments, the consensus view will be the correct one, and that any neutral admin would agree. Thus there is in theory no difference between closing per the majority and closing per the strongest argument. But when there is a real dispute on what argument is relevant, the closer is not to decide between them , but close according to what most people in the discussion say. If the closer has a strong view on the matter, he should join the argument instead of closing, and try to affect consensus that way. I (and almost all other admins) have closed keep when we personally would have preferred delete, and vice-versa. .
When admins delete by Speedy, it is on the assumption that what they are doing is so unambiguous that the community has given implied consensus in advance. If someone challenges this is good faith, the proper response is to simply send the article for AfD, and find out the express consensus.
If I wanted a place where my view of proper content would prevail, I'd start a blog or become an editor of some conventional publication.
On Mon, May 31, 2010 at 5:51 PM, David Lindsey dvdlndsy@gmail.com wrote:
The key is not making it easier to remove adminship. This proposal gets us closer to the real problem, but fails to fully perceive it as does the common call to separate the functions of adminship.
The real solution to the current (and relatively long-standing) problems with RfA and adminship in general is the marriage of the "technical" side of adminship with a "political" side, which is rarely acknowledged. Successful reform will involve separating these two aspects, rather than the more common idea to separate some technical pieces from others. The proposal below is a bit lenghty, but it's the product of years of thought, and I encourage you to read it. If you don't have the time, well then, the take away point is that we should create a distinction between those administrators trusted to intervene in highly-controversial areas and those not so trusted.
The technical bits of adminship are, indeed, no big deal. With a large community of administrators and an alert body of stewards, the possible danger of obvious abuse of the administrator privileges is nearly zero. As an illustration, in the heat of the recent dust-up on commons, an administrator there "went rogue" and vandalized the main page. His edits were reverted in less than a minute: http://commons.wikimedia.org/w/index.php?title=Main_Page&action=historys.... Even in an absolute worst-case scenario of administrator abuse (for example, vandalizing the main page and then deleting a large number of pages with just less than 5,000 revisions in an attempt to lock the servers, especially abusive shenanigans in the MediaWiki namespace, or inserting malicious code into monobooks), the damage done would be reversed in under 10 minutes. Given this, it is highly improbable that any vandal/banned user would attempt to gain administrator status solely for the purpose of carrying out some such abuse. The danger comes from a compromised account or a higly disaffected administrator, and neither of these possibilities can be headed off by any level of standards at RfA, however high.
Why, then, has adminship become a big deal? Because in addition to the purely technical functions of adminship, administrators also have a political function. Administrators are often compared to janitors, but the metaphor is highly flawed. Janitors empty the wastebins, but they don't decide what should go in them. Many of the functions of adminship do not carry a significant political component: blocking obvious vandals, most instances of speedy deletion, fixing cut and paste moves, deleting old userpages, straightforward AfD closures, etc. are simple instances where a trusted user is needed to perform a technical function.
On the other hand, there are cases were administrator functions become highly charged and political - in closing controversial AfDs, blocking in many 3RR situations, and above all, in cases where some sort of intervention is necessary against well-established users who have engaged in some sort of unacceptable conduct. In these cases, the role of the administrator is fraught and ambiguous. He is faced with highly political choices about how to judge consensus, what course of action to take, etc. It is customary for relatively new and inexperienced administrators to stay out of these situations and leave the decision up to an administrator who has more experience and, for that matter, for political weight within the Wikipedia system.
The problem, though, is that there is no formal guidance of any kind as to who should actually make such decisions. From a policy perspective, an administrator sysopped last week has the same standing as someone with years of service. More importantly, a long-standing administrator with a reputation for more questionable judgment has exactly the same standing as a long-standing administrator with a reputation for impeccable judgement. There is no drawn by the community, except in the various most informal way, to separate administrators who should intervene in highly controversial situations from those who should not.
It is intervention in the highly controversial cases that causes problems and allegations of abuse. Our concern is, or at least should be, primarily in who is making highly controversial administrator judgements and on what basis, not who is carrying out F5 speedy deletions or blocking obvious vandals. Concern over these highly controversial judgements, because there is no line separating those administrators who engage in them from those who do not, is what has driven steadily escalating standards at RfA. We are less concerned that a newly-appointed admin will prematurely block a vandal without any warnings tomorrow, than that he will, in 12 months, block a well-established user for the wrong reasons after a heated debate at ANI. In other words, the problem is that RfA is being asked to make a judgment that should not be made at RfA.
What we need, then, is not a way to desysop more easily, but rather a way to delineate highly-charged and controversial administrator actions, and the administrators qualified to perform them, from uncontroversial administrator actions, and the administrators qualified to perform them. I will not presume to provide a full criteria for what separates controversial from uncontroversial administrator actions, but I would suggest something along the lines of the following. Controversial: Arbitration enforcement actions, blocks of established users for any reason other than suspicion of account compromise, close of AfDs where the consensus is not clear (this of course becomes itself a murky distinction, but could be well enough set apart), reversal of the actions of another administrator except when those actions are plainly abusive. Non-controversial: All others.
As for deciding which administrators are qualified to make decisions in the most controversial areas, I would suggest that we already have a group of people, the bureaucrats, in whose judgement the community has expressed particularly high confidence. I would propose that the bureaucrats become the group who are expected to undertake the controversial administrator actions; this would almost certainly entail some expansion of the current bureaucrat pool, but personally I like the idea of tying the controversial administrator actions to the ability to promote administrators - it underlines their seriousness, and at present, the bureaucrats do not have many functions. If, however ,the community is unwilling to combine the two groups, another group, say "sub-bureaucrats" could be created, but I must emphasize the importance of a bright-line distinction between those administrators trusted to perform highly controversial tasks and those not trusted to do so. Obviously, the ordinary administrators would still have the technical ability to intervene in the highly controversial areas, but doing so would obviously entail serious consequences or desysopping.
This brings up a final point: the issue of administrators with insufficient knowledge to appropriately follow policy on, for example, speedy deletion. I firmly believe that if we separate the political and non-political aspects of adminship, this becomes less of an issue. While an administrator taking the wrong course in a controversial area is akin to a janitor, who is empowered to decide what to throw out, deciding to throw away your important papers because he doubts their importance, the mistakes of lack of policy knowledge and inexperience are more like a janitor who, because he doesnt' know any better, throws away the recycling and attempts to recycle the rubbish. The second category of mistake is more easily rectified. The old idea, of some sort of mentoring for new administrators, does nothing about the political aspects of adminship (making controversial decisions) which is why it has failed in the past, but it is a perfect solution to the problem of inexperience/ignorance. New administrators who do not have a full grasp of the speedy deletion policy, or the blocking policy for vandals, or the criteria for granting autoreviewer status would be encouraged, perhaps through a formal process, to get up to speed on those areas by a more experienced mentor. If we carry through this proposal, there is every reason to believe that the crowd at RfA would be much more willing to promote more candidates and the process would become much less grueling. Our shortage of people to perform technical tasks could be easily reduced, if not eliminated.
This proposal is not process creep or the introduction of unneeded bureaucracy. It is also not an answer in search of a problem. There is clear acknowledgement that we have a problem, and this solution is a minimalist one. As I have proposed it, it simply takes advantage of an existing process (RfB) and group of users (bureaucrats) and would require only minimal amendments to policy, setting aside those areas of administrator conduct that are highly controversial and requiring that only bureaucrats act in those areas. _______________________________________________ WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
David Lindsey wrote:
What we need, then, is not a way to desysop more easily, but rather a way to delineate highly-charged and controversial administrator actions, and the administrators qualified to perform them, from uncontroversial administrator actions, and the administrators qualified to perform them. I will not presume to provide a full criteria for what separates controversial from uncontroversial administrator actions, but I would suggest something along the lines of the following. Controversial: Arbitration enforcement actions, blocks of established users for any reason other than suspicion of account compromise, close of AfDs where the consensus is not clear (this of course becomes itself a murky distinction, but could be well enough set apart), reversal of the actions of another administrator except when those actions are plainly abusive. Non-controversial: All others.
In other words, a two-tier system of admins. Against that, I really think there is an area that should be thought through, just alluded to there. The criteria for reversing another admin's actions do matter, and it seems to me matter most.
Admin actions that can be reversed (i.e. technical use of buttons, rather than interaction by dialogue) lack the sort of basic classification we need: into situations of urgency and situations that can wait; situations of key importance to the project (such as involve harassment, for example), and those that can be treated as routine; and into situations where consultation should be mandatory and those where a second admin can use judgement to override. The fact that some people might conflate those analyses illustrates the need to be more careful here.
I think this is something to untangle. We need to get to the bottom of the community's fears about "overpowerful" admins, by talking through and delineating what a single admin can expect to face in awkward situations. I've never been in favour of restricting admin discretion, which is really what is being proposed. We can't anticipate the challenges the site will face (even though it may appear that there is little innovation from vandals and trolls). I do think admins can be held to account for their use of discretion. Right now it seems that a piece of the puzzle is missing: admins don't know clearly how they stand in relation to the actions of other admins.
Charles
At 05:21 AM 6/1/2010, Charles Matthews wrote:
I think this is something to untangle. We need to get to the bottom of the community's fears about "overpowerful" admins, by talking through and delineating what a single admin can expect to face in awkward situations.
Yes.
I've never been in favour of restricting admin discretion, which is really what is being proposed.
It's not what I'm proposing. Discretion should be almost unlimited as to primary action; however, there should be much better guidelines so that admins can know what to expect. WP:IAR is a fundamental and very important principle, but that doesn't negate that if one ignores rules, one should be prepared to face criticism and be required to explain why or face warning and possible suspension of privileges.
We can't anticipate the challenges the site will face (even though it may appear that there is little innovation from vandals and trolls).
There are structural devices which can make vandalism and even editorial review much more efficient, and there are trends in that direction. When Wikipedia starts valuing editorial labor, and sets up systems to make it more efficient and reliably effective, it may get over the hump. I've suggested that it may be appropriate to start channeling labor into what I've called "backstory," i.e., documentation of why an article is the way it is. Then, if a new editor disagrees, that editor can quickly come up to speed on the history, see all the arguments and evidence organized, and would not be imprisoned by that, but rather might be encouraged, if some argument there is defective, to show that, to expand the consensus there. And then that can be taken back to the article. Articles should not slide back and forth, that is incredibly wasteful. They should grow, such that consensus is always that they have improved by a change. Flagged revisions is a piece of this puzzle.
I do think admins can be held to account for their use of discretion. Right now it seems that a piece of the puzzle is missing: admins don't know clearly how they stand in relation to the actions of other admins.
I developed, early on, a sense of how Wikipedia worked, and it made a great deal of sense in terms of the organizational theory I was familiar with. And then I discovered that only some administrators seemed to understand it. Others believed that the structure was something else. I saw no disruption coming from administrators who understood the concepts that seemed obvious to me. It came from the others. Recusal policy should be far more clear. But that's not the first priority. The first priority is establishing consensus process that is more efficient; the inefficiency discourages participation and causes proposals that might actually help to go nowhere. "No consensus."
That should be a clear suggestion for "refer to committee." That's what successful organizations do when faced with a problem where the response is not clear. (And then committee composition and rules and process become very important.)