Please leave your comments below. Objections, approval, or suggestions for changes or extensions are all welcome. Dcoetzee 11:59, 16 November 2011 (UTC)[reply]

The tools that an editor would most likely request are also the ones that the community considers the most sensitive and that drive the most concern at RFA. My understanding is that the ability to Block and to Delete are considered the most dangerous abilities to hand out, followed in a distant third by Protection. Yet I think 90% of requests for tool access will end up involving at least deletion, in the case of NPP, vandal patrolling and those active at deletion venues, and in the case of vandal patrolling, they will have a strong need for block, and may also request protection for semi-protect. Now it is true that we can just apply stricter scrutiny when those tools are requested, but as we do so, we turn it right back into the mess RFA is in now.

The only place this would really help is in the really edge cases where someone is highly active in an unusual and specialized niche that is uncontroversial, the only that come to mind would involve editing protected templates and/or edit notices. So I would ask, what are the other types of requests envisioned by this proposal that would not include deletion or blocking and be subject to RFA like scrutiny, and how often would they end up being made? Monty₈₄₅ 14:32, 16 November 2011 (UTC)[reply]

My response: it is true that block and delete are both dangerous tools that one would not grant lightly on an indefinite basis. However, I believe granting them on a trial basis to a user in good standing with an immediate need for them, and removing them immediately in case of abuse, is a much less risky proposition. Not only does this directly limit the damage the user could cause, but it sets up incentives to discourage abuse (the user wishes to continue to participate in the task they requested the tools for, and is aware that they could easily lose the tool if misused). This is why I centered my example around delete, which is seen as a dangerous tool. And yes, I would expect vandal patrollers to request both block and protect (as already noted). Dcoetzee 19:42, 16 November 2011 (UTC)[reply]

• Wikipedia:PEREN#Hierarchical_structures.

Plip!

. That's all I have to say on this subject. Beeblebrox (talk) 17:20, 16 November 2011 (UTC)[reply]

• • I am of course familiar with WP:PEREN. Not only does that entry openly contradict itself ("Can't be done! Already done!"), but the reasoning is easily refutable:

    If we can't trust people to use their tools sensibly, they don't become admins

  • I already respond to this on the page (trust in one area does not necessarily imply trust in another; trust in a particular tool should be based on experience with that tool). Moreover, in my proposal there are no administrators in the traditional sense (only persons who have acquired a certain set of tools). It's certainly not the case that "anyone we can trust with rollback we can trust with all the tools" - this is a simple extension of that idea.

    A "partial admin" process would at least double the already considerable frictional effort expended at WP:RFA, as users debate who gets full sysop powers versus who gets only partial abilities.

  • As argued in the proposal, the discussion effort expended would be considerably less, because the stakes would be lower (we don't have to be as careful and thorough when handing out a subset of powers on a trial basis as we do when giving out the full package indefinitely - the trial component, which is novel, is critical here). Additionally, my proposal dispenses with people getting "full sysop powers" altogether, except if they happen to eventually acquire them all. Dcoetzee 19:28, 16 November 2011 (UTC)[reply]
    • Added some responses to this to the page. I do realise the importance of directly addressing objections to prior similar proposals. Dcoetzee 20:06, 16 November 2011 (UTC)[reply]
    • After further thought I have struck the idea of eliminating RfA at the present time. Dcoetzee 16:26, 17 November 2011 (UTC)[reply]
  • I agree with Dcoetzee here. The rebuttal at PEREN is pretty weak. I think, by suggesting such a measured approach to access to administrative tools, this proposal is a great improvement on earlier proposals with a similar goal. — This, that, and the other (talk) 01:10, 17 November 2011 (UTC)[reply]
  • Although you say above that 'Additionally, my proposal dispenses with people getting "full sysop powers" altogether, except if they happen to eventually acquire them all.' Is this suggesting that RFA would be done away with? It still produces new admins, so I see no reason why it should be abolished. I see this proposed process as a parallel, for users who do not want all the tools, or who believe they would not pass an RFA on the basis that they do not have broad experience in many different areas. — This, that, and the other (talk) 01:31, 17 November 2011 (UTC)[reply]
    • Thanks for the feedback - see the new section "Does this supplant or supplement RfA?", which I think addresses your concerns. Dcoetzee 16:26, 17 November 2011 (UTC)[reply]
• For whatever it may be worth, I 100% support unbundling the tools. Apprenticeship for each tool is a reasonable idea but not nearly as important as eliminating the idea of "one size fits all" adminship. A person may be very well experienced with, say, page deletions from their experience as an editor (nominating/tagging pages, participating in discussions) and thus well-prepared to assume the deletion tool, while having little preparation for blocking. Or vice-versa. --RL0919 (talk) 01:56, 17 November 2011 (UTC)[reply]
  • I wonder how existing admins would be viewed then. Would it reinforce the stereotype that we're superusers because we have the entire toolset, while the 'lower classes' struggle to prove themselves with each tool so that they may 'rise up in rank'? Would you still be in support if the proposal included apprenticing existing admins for each of their tools? -- œ^™ 08:10, 17 November 2011 (UTC)[reply]
    • That's part of the reason why I would want this to remain in parallel with RFA, so that current "total sysops" don't become an elite class. Of course, one mustn't worry about this sort of nicety too much: this tool apprenticeship proposal, in itself, deserves merit. — This, that, and the other (talk) 08:59, 17 November 2011 (UTC)[reply]
      • Yes, I agree that retaining the current RfA process at least for now is important for ensuring a smooth transition. Dcoetzee 16:26, 17 November 2011 (UTC)[reply]
• It'd work on principle, but in practice it would establish the very sort of social hierarchy which most of us oppose. By unbundling the sysop tool set, there'll be more haberdashery and the rights will be viewed as prizes by those haberdashers and it'll reinforce the elitism which does have an active presence amongst our admin and user corps (the current mentality that I can observe is that the more individual rights there are, the more users will want to apply for them, the standardised user rights are already frequently sought out by the unready, imagine what the media of requesting these "mini-admin" rights will be like).

I'm that certain users show exemplary knowledge of certain areas of Wikipedia's administration, but that will be problematic as admins are expected to assist wherever they can, as that is their duty, to assist. By unbundling the rights, we'll have users who are essentially proving themselves unable to do any work outside of their area(s) of comfort/expertise, this would create a problem as these users would then NOT need to learn about other areas other than their area(s) of expertise.

Admins are specialists but they're also expected to lend an extremity wherever possible, their primary job is to maintain the cleanliness of the encyclopedia where standard users are unable to do so, blocking etc. Having a user right dedicated solely to blocking, while seemingly a good idea is pointless, that user can do nothing more, which would effectively limit their usefulness. Furthermore, it would add to the cruft of bureaucracy that is present in our requests for permissions pages and there would be a larger number of overall individual requests, I know admin promotion rates have been steadily declining, but RfA has not had as much of a problem with curbing requests from the unready, but the general daunting nature of running is a deterrent enough. Unbundling the rights would lessen the workload and the responsibilities attached, thus making these individual RfP pages prime targets for haberdashers.

I know I'll probably be criticised for opposing this system, as I advocated such a system in the past, but time on Wikipedia and observing the habits, tendencies and general mentality of haberdashers has made me see there would be no hope for such a system and would limit the efficacy of adminship, each department occasionally has an overlap in duties and may be required to collaborate, that would require several admins at most, but with this system it'll require several users OF THOSE INDIVIDUAL rights, that would create more work than is necessary. —James ^{(Talk • Contribs)} • 9:03pm • 11:03, 17 November 2011 (UTC)[reply]

Hi M.O.X., thanks for your feedback. To address some of your specific points:

there'll be more haberdashery and the rights will be viewed as prizes by those haberdashers

I do expect many requests from users who are unready. This is why I've included the list of requirements for a trial period, evaluated by consensus discussion, which I've now tried to make more explicit. I believe these will filter out "haberdashers" effectively.

admins are expected to assist wherever they can, as that is their duty, to assist. By unbundling the rights, we'll have users who are essentially proving themselves unable to do any work outside of their area(s) of comfort/expertise [...] Having a user right dedicated solely to blocking [...] that user can do nothing more, which would effectively limit their usefulness

The overall number of users having partial rights would be much greater than the number of admins, enabling them to do more work overall, although each individual may do less. Also, as I'm not presently eliminating the RfA process, there would still be admins with a "duty to assist".

it would add to the cruft of bureaucracy that is present in our requests for permissions pages and there would be a larger number of overall individual requests

It would - the bureaucratic overhead is nontrivial. In return, we get more users with more tools to help complete more work in other areas. I think it's a net win.

each department occasionally has an overlap in duties and may be required to collaborate, that would require several admins at most, but with this system it'll require several users OF THOSE INDIVIDUAL rights, that would create more work than is necessary

You are right that some collaborations would require more individuals to succeed, increasing communication overheads. However, most work is contained to limited areas, and I think overall efficiency would increase.

These are reasonable concerns and if I'm missing something still please explain. Thank you! Dcoetzee 16:41, 17 November 2011 (UTC)[reply]

I still have my doubts, having increased collaborations in contrast to the controlled number we have now, to me, doesn't seem like a benefit, if the admin package will exist in conjunction with the individual rights, I believe there should be bundling. The ability to block users, move media and categories and page protection should be bundled, those are basic admin functions and the ability to function with these rights is exemplary of an admin's ability to carry out simple maintenance, notwithstanding the occasional extraordinary incident. Instead of that, you could use certain tools from the admin package and put them into their own package (eg. blocking and page protection, for the anti-vandal users). —James ^{(Talk • Contribs)} • 12:49pm • 02:49, 18 November 2011 (UTC)[reply]

I'm a little unclear about your explanation here. I agree that certain sets of tools are useful for certain tasks, like blocking and page protection for anti-vandal users, but because the set of potential tasks is very large and fluid, I think it's more flexible to permit any set of tools to be requested. But I don't see why it'd be useful to bundle block users, move media, etc. The primary purpose of the system is not to provide evidence that you can capably function as an admin (that's a secondary benefit) - the primary purpose is to help people who have an immediate need for a tool to get work done. Does that make sense or am I missing something? Thanks again. Dcoetzee 09:43, 18 November 2011 (UTC)[reply]

No, that was poor wording on my part. The bundling of rights would mean that the simple tasks can effectively be handled by anyone and the specialist tasks handled by the specific users with the concerned rights, I understand that the reasoning behind this proposal is that admin workload is now somewhat overburdened and that there are more than capable users who are afraid of RfA or the harsh standards of RfA would mean they'd not pass, however, the problem I see is that this would make adminship a triviality and the immediate ramifications of unbundling the package would be that we'd have a surplus of users who effectively cannot work outside their chosen departments, increased and highly trivial collaborations and an increase in overall workload, to me this outweighs any inherent benefits. —James ^{(Talk • Contribs)} • 8:58pm • 10:58, 18 November 2011 (UTC)[reply]

Okay I think I understand that. My response would be that it all depends on how it's used in practice. There will be some users who get a tool or two for a single task and then stop, while other users would acquire multiple tools before eventually shooting for a full RfA. The experience would help users who are intimidated by RfA and would never run otherwise to eventually go for it. Which type of user turns out to be more common is difficult to anticipate. I don't think availability of tools by another path would make adminship trivial, partly because adminship includes a large number of tools that would be quite taxing to accumulate individually (perhaps some won't be available via apprenticeship at all), and partly because adminship (for better or worse) includes a sense of status and authority that partially empowered users would not possess to the same degree. Dcoetzee 17:37, 18 November 2011 (UTC)[reply]

Hm, I see. In a way you've inadvertantly supported my position, adminship does contain an enumerous amount of tools, individual acumulation while taxing would be easier to achieve than to run for RfA straight-out, while it would give users the experience and knowledge necessary, this can be achieved without the need for the unbundling of the package, apprenticeship, just seems like Admin Coaching but without the limitations. From my perspective it just seems like this encourages users to aim lower than they should be, adminship is daunting for a reason, that is to deter the unready from running and to encourage people to serve the community the best they possibly can and show that the tools would assist their work.

Having specialist rights has already been criticised, look how much the admin package has been de-bundled so far, that in itself has attracted scrutiny and criticism, the proposal is good in principle but from what I can see it would only encourage users to aim for second-best, it would create more work and add to the bureaucracy we already have, while you have dispelled some of my doubts I can't see this system working. This proposal is haberdasher-fodder. —James ^{(Talk • Contribs)} • 11:03am • 01:03, 19 November 2011 (UTC)[reply]

Please forgive me, for just a moment, going off the topic of this proposal - so I can explain the angle I'm coming from. I'll be as brief as I can be, and come back to this proposal after a few sentences;

My view of "The trouble with RfA", in brief: Adminship is a big deal, because it's so hard to remove it. If it were much easier to remove, we wouldn't be so super-cautious about handing it out.

We block users pretty quickly - e.g. for a username vio - and we say "hey, this is why you're blocked - it's nothing personal; let's just discuss it and sort things out"

Imagine that for SysOp - "Hey, I've removed your SysOp status, because I noticed a couple of dubious deletions that didn't seem to meet CSD criteria. So, let's discuss them, and if we sort it out, you can have your SysOp back"

For similar reasons, I think 'admin' should be renamed 'janitor'. It's become far too much a 'badge of honour'. Even most admins seem to think that they've got special authority to make decisions in e.g. closure of consensus discussions. They haven't.

Having said that, I doubt it'll ever happen. In part, because most people who make such decisions are either admins, or people heavily involved in the 'pipework' (non-content) parts of the wiki. Sadly, because those people deal with blocks/socks and nastiness all the time, they naturally forget we're here to build an Encyclopaedia.

</gets off soapbox...>

So - if that is unrealistic / unlikely to happen, then perhaps some trial-system of temporary/partial SysOp is a way forward. A hell of a lot of us think some change is needed...so, perhaps we can give this system a try? Even just a very limited run (few people), and with extra-careful-watching by some well-respected admins/CU/'Crat-folks. For example: We could try it out on just 6 people for a month, and could allocate a couple of mentors to check every single action they take. Net result of that could be 6 new admins before Xmas.

The key stat, to me, is: New admins: 2007, 408; in 2008, 201; in 2009, 119; in 2010, 75. 2011, so far, 46. Surely this is a worrying trend. Hence, even just another few via an attempt at something new would be progress.

Of course, there's a lot of problems and complications with this proposal - I'll mention a few first thoughts, below. But hey - it's a wiki - problems can be fixed.  Chzz  ►  08:21, 17 November 2011 (UTC)[reply]

How many potential set/s of tools are there? Are any of them possibilities? Seems like it could easily get complicated. For example, at a quick glance right now, in addition to the more obvious examples of delete/prot/block, I could use;

• "moodbar-admin" (Alter visibility on the feedback dashboard) - I've been actively involved in work on that extension, and can easily make a case for me to see, and hide, comments
• "edituserjs" - I'm frequently trying to help others sort out problems with their script; this would save lots of hassle saying "try removing that third wiggly-bracket", etc.
• "move-subpages", "suppressredirect" - because I very often help people manage userspace pages, and constantly have to ask admins to tidy up
• Add groups: Confirmed - to make it easier when I help new users, and they need to upload a file or move a page

I'm just showing those examples to indicate some of the potential complications of this idea. Personally I'd prefer to KISS and allocate SysOp, and trust the carefully-watched-users to only use what they'd been told they could use; but, I understand if that might be less acceptable.

One particularly problematic user right is, (deletedtext) - View deleted text and changes between deleted revisions - because I can imagine a lot of people finding that handy, but there's probably no way to evaluate if they use it responsibly or not; it won't actually show up anywhere (unless they copy/paste something). So if a user were granted that for a few months, how could we decide if they should keep it?

Re. the Responses to anticipated objections, WRT "If we trust a user enough to give them one tool, we trust them enough to give them all of them."

-this is pretty key, I believe. The refutation overleaf says, Trust in one area does not necessarily transfer into other areas (consider the aphorism "I trust my bank with my money and my brother with my children, but not vice versa")

Well...consider this: if I trust my brother to borrow my hammer, my power-drill, my screwdriver, etc. then it's a lot more simple to give him a key to my shed - per KISS. After all, if he breaches my trust, I can just take away his key.

I can imagine difficulty in telling who-does-what, and indeed who is allowed to do what; unbundling does have a lot of potential for instruction-creep.

Another example overleaf talks of a user undeleting, but unable to delete - that sounds problematic, as they cannot rectify their own errors.

 Chzz  ►  08:21, 17 November 2011 (UTC)[reply]

You say above (after your descent from the soapbox) that part of the trouble with admins is that it is so hard to revoke sysop-ship. But your example above refers to "taking away the key" of the mischievous brother, without any trouble. This seems a contradiction to me: one good reason not to give a sibling (or anyone) a key to one's garage is, once they have it, it can be very hard to get it back from them again, especially if they are up to mischief. They aren't likely to willingly give it up: nor are admins who bend the rules.

Re rectification of one's own errors, a normal user has to rely on CSD G7 to get a page deleted after it was created in error. A "deleter" deleting and then desiring to undelete could go through WP:REFUND just like any other user, and an "undeleter" wishing to delete would of course use CSD tags. I don't see that as being a huge impediment to the proposal, particularly given the suggested trial period would help to weed out candidates who lacked full maturity and responsibility.

Also re deletedtext: Alongside the problem you point out with granting it alone, I don't think the WMF would like it if non-admins were given this permission. So this one might have to stay out of this plan for the moment.

As you say, this could easily get complicated, but where are we now? Rollbacker, file-mover, accountcreator, edit filter manager, importer, checkuser, oversight, and even reviewer and autopatrolled (obviously not so important at the moment)... a few more would hardly hurt...

Many of the points you mention are certinaly worth keeping in mind. — This, that, and the other (talk) 09:14, 17 November 2011 (UTC)[reply]

Re. key - sorry if the analogy wasn't great, but...on my shed, it's trivially simple and cheap to change the (pad)lock.  Chzz  ►  09:41, 17 November 2011 (UTC)[reply]

Hi Chzz, thanks for your feedback. I have blatantly ripped off your idea of a limited trial of the process, something I had not even considered but which is absolutely necessary. Responses to your comments:

Adminship is a big deal, because it's so hard to remove it

While I make no attempt to address this issue for admins, note that tool trials in tool apprenticeship are subject to probation (tools can be removed in case of misuse, then later restored after the user learns their lesson and makes a fresh trial request).

How many potential set/s of tools are there? Are any of them possibilities?

There would simply be a user group associated with each user right. This would enable us to easily grant any combination of tools to any person, and allow a person to accumulate a variety of seemingly unrelated rights over time such as the ones you listed. Note also, I'm presently proposing this as a supplement to RfA, so in practice a user who participates in a diversity of areas would probably start out with just a 2-4 tools, then jump to RfA there.

One particularly problematic user right is, (deletedtext)

This is indeed problematic. It is technical limitations that prevent us from monitoring use of this right, and in the long run I would hope for software changes that introduce auditing (logging each use of viewing deleted revisions). In the short run, I would only introduce this right in combination with a right that obviously requires it, such as undeletion, which can be monitored.

if I trust my brother to borrow my hammer, my power-drill, my screwdriver, etc. then it's a lot more simple to give him a key to my shed - per KISS. After all, if he breaches my trust, I can just take away his key.

This is partly why at the present time I'm not suggesting eliminating RfA. If I've watched my brother use the circular saw and power-drill for a while with supervision, and he knows what he's doing, I might then develop enough confidence to give him the key to the shed.

they cannot rectify their own errors

Replied to this already in the objections - it is a bit awkward to seek help from others to rectify your own errors, but I expect these requests to be processed quickly.

Thanks again and please let me know if your concerns have been addressed. Dcoetzee 17:15, 17 November 2011 (UTC)[reply]

Support the general idea. As it becomes more specific I may have questions or comments about specifics. Pine^talk 09:10, 17 November 2011 (UTC)[reply]

Requirement Provided the user is in good standing and has a need, they receive the tool on a trial basis

I'm concerned that it might attract a very large number of new-ish/relatively inexperienced users, which could create considerable - and probably pointless - stress. So, should there be some specific requirements? e.g. been around for <xx months>, at least <xx> <mainspace? any?> edits? Of course, exceptions could be made...but I'm mostly thinking of avoiding lots of NOTNOW candidates.  Chzz  ►  10:50, 17 November 2011 (UTC)[reply]

I have added the requirement "To the greatest extent possible, the user should be active and have sufficient experience in the area in which they plan to use the tool." This is intentionally vague - I expect each area to formulate its own specific requirements over time. I don't think general requirements make a lot of sense, since different areas use different metrics. Dcoetzee 17:19, 17 November 2011 (UTC)[reply]

After or shortly before the end of their trial, the user can file a request to retain the tool permanently

How would that be decided? Who would make the decision? Would there be a certain opportunity for any others to object?  Chzz  ►  10:50, 17 November 2011 (UTC)[reply]

I can see this being like an entry at WP:RFPERM, except maybe open for 7 days, to allow for objections and discussion; however, AFD-style consensus and RFA-style !voting are probably not necessary. I would imagine that bureaucrats would make the decision, after perusing the user's logs; but I'm not sure about this. — This, that, and the other (talk) 11:10, 17 November 2011 (UTC)[reply]

I have proposed an AfD-style consensus discussion. I think AfD is a great model here because it invites community input, takes load off decision makers (here, our small number of bureaucrats), but most discussions are short and straightforward. Dcoetzee 17:19, 17 November 2011 (UTC)[reply]

• Comments Assuming that current admins would remain as they are we have a mildly hypocritical situation: "Old money" verses "Nouveau riche".
  • Admins come and go. If no more can be made then admins only go (retirement, real life reasons, drama etc.). It is possible that we could end up with no admins. With no admins, who decides who gets what tools?
    • There is a possibility that too few or even no users could have use of any particular tool or set of tools at any one time for needed work to be done in a timely fashion. It is further possible that no users at all have use of some tools. Who would fix it and how?
  • At present admins have a position of responsibility that they are publicly answerable about. Their position is not to be taken lightly. The potential for abuse is thus greatly reduced as the rights are special and they want to keep them. If tools could be aquired by anyone at any time they could feel a little disposable. The family silverware and crockery may be cherished by generations but, why cherish paper plates and plastic cutlery?
  • The workload of admins with the power of oversight (those who dish out the tools) would be potentially too great for the requests to be dealt with in a timely fashion (assuming there are any pseudo admins left with the power to dish them out). We might find that little administration is being done since the few who can, spend all their time answering requests for tools.
  • I know this is covered in the previous posts but it's worth another mention: this would allow a whole new level of vandalism that could only be fixed by select individuals. If all vandalism could only be cleaned up by some particular kind of admin, we could kiss this project goodbye.
  • Finding admin to get a job done is at present very simple. It could be a great deal more difficult to find the right kind of admin to get things done.
    • A user could find a pseudo-admin to do some task and then the pseudo-admin may need to find another with different tools who may in turn need to find another. The leg-work would be time consuming and wasteful.
• At present admins are peer reviewed by a dedicated team who don't take the position lightly. Those few who get through are regarded as being pretty much unquestionable worthy. If we reduce adminship to a set of temporary add-ons, we could find that fewer and fewer take seriously the guardianship of the encyclopaedia.
• Summary: Some tools could be dished out in this fashion without causing too much trouble but not all. However, I think this idea should only be considered to run along side the present system so as to not degrade or trivialize positions of authority. It might prove useful to real-admin to have a record of use of certain tools dished out to users if they were to apply for real-adminship (just as use of rollback would be) later.
  • Analogy: I wouldn't want the police force to be created and maintained this way. We would end up with vigilante-ism on a grand and uncontrollable scale. I like my police to have gone through thorough training, to have been tested and vetted by qualified uber-officers, and to be constantly monitored by their peers and the public. I don't want my police to be a rag-tag disperate group of wannabes with truncheons looking to acquire guns. fg 13:23, 17 November 2011 (UTC)[reply]

Hi Fred, thanks for your feedback. First of all, based on feedback by yourself and others I am currently proposing this only as a supplement to RfA. Other comments:

With no admins, who decides who gets what tools? [...] The workload of [...] (those who dish out the tools) would be potentially too great for the requests to be dealt with in a timely fashion

Anyone can participate in request discussions. They are closed by bureaucrats. If the load becomes too high for bureaucrats, a recursive approach can be taken, in which the right to grant rights for tool apprenticeship is granted via the tool apprenticeship process to existing admins who are experienced in tool apprenticeship discussions.

• Potentially disastrous, blind leading the blind. Also the only re to my re that almost touches on the main problem, how many have what tools and when? We could hypothetically run out of "pot scrubbers" or "hammer wielders". Although I do now see (from your comment above) that you are now not proposing we replace administrators with have-a-go-heroes. fg
  • Yes, I've reframed this as a supplementary process. While I hope that we would always have enough users interested in each right/privilege, it is possible a small number of them would be very unpopular for whatever reason (e.g. only useful in specialized situations). This is one reason to keep the full sysop package around. Dcoetzee 18:58, 17 November 2011 (UTC)[reply]

If tools could be aquired by anyone at any time they could feel a little disposable [...] I don't want my police to be a rag-tag disperate group of wannabes with truncheons looking to acquire guns.

As is now emphasized, people on their tool trials are under probation. Although they are easier to obtain, tools can be rapidly revoked for misuse. I consider this acceptable because all tool actions (unlike police actions) are easily reversible.

this would allow a whole new level of vandalism that could only be fixed by select individuals

For most types of vandalism only a couple tools are required (protection, blocking, rollback). Some types of vandalism may require other tools (deletion for attack pages and move vandalism), but the person user the vandal can merely tag these pages for speedy deletion, and have another user delete them. Deletion is not an urgent matter.

• Scenario: User acquires tool to apply protection to pages, either with the intent of misuse or because of mental breakdown, personal tragedy, Wikidrama (whatever reason) etc. suddenly takes to abuse. They (by use of a simple bot) replace the text on 10000 pages with "F@#K WIKIPEDIA" in 500px text and then protect the pages. The only people who can clear it up are admin-like. This is fine if an admin-of-sorts or a real admin is around who actually has the tools needed to nuke the edits. fg 17:59, 17 November 2011 (UTC)[reply]
  • In this scenario actually any user with the protect right and access to a suitable bot could reverse the damage. The bot could be supplied by any trusted user. Nuke isn't so helpful for users who happen to also have legit edits that we want to preserve. But yes, admins would still be around. Dcoetzee 18:58, 17 November 2011 (UTC)[reply]

It might prove useful to real-admin to have a record of use of certain tools dished out to users if they were to apply for real-adminship

This is indeed an advantage of running the tool apprenticeship along RfA, which I have noted in the proposal.

A user could find a pseudo-admin to do some task and then the pseudo-admin may need to find another with different tools who may in turn need to find another

I do expect higher communication costs in some situations. I expect overall efficiency to increase due to having a much larger number of partially-empowered users available to help with the workload.

I hope this addresses your concerns, but please let me know if you have other feedback. Thanks! Dcoetzee 17:35, 17 November 2011 (UTC)[reply]

I've been in total support of the idea of unbundling admin tools for quite some time, so I like the idea of tool apprenticeship. Would this (excuse me if I've missed it somewhere!) run alongside some kind of tool-mentoring? In real life, apprentices work alongside / under the supervision of a journeyman or master in the same trade, and I think this might be a good idea. Pesky (talk …stalk!) 17:39, 17 November 2011 (UTC)[reply]

Hi Pesky, thanks for your comments. I've been reluctant to propose mandatory tool mentoring, although it is a splendid idea and one that I had in mind, because I fear the overhead of an assigned mentor responsible for continuous monitoring of every apprentice may be substantial. I've been going for the more ad hoc notions of probation (if you mess up and someone notices, your tool can get revoked) and review (when you come back for your next request, your logs are reviewed at that time). However there's no reason not to have voluntary mentors, and I would support mandatory mentors for problematic users. I haven't written any of this up yet but will give it further thought. Dcoetzee 18:01, 17 November 2011 (UTC)[reply]

Getting together a list of people who'd be willing to be mentors in specific areas would be good. Then apprentices could hand-pick a mentor from the list. Pesky (talk …stalk!) 18:20, 17 November 2011 (UTC)[reply]

Support but I don't really understand why tool should be automatically removed after trial, however it looks like role based security model which is rather useful for this site, explanation: having more permission sets would make it easier to give one access to the tool they exactly need while there wouldn't be need for such restrictions as now for whole sysop bit, I myself needed a single permissions once (global confirmed) unfortunately "lowest" permission including this is a global rollback, which I wasn't considered trust-worth enough for, other example are people working on AFC who need to perform tasks like deletion or edit of some protected template very often. Petrb (talk) 20:15, 17 November 2011 (UTC)[reply]

Hi Petrb, thanks for the feedback. The reason for removing the tool after trial is that it forces the user to go through a review if they want to get the tool permanently (or for a second trial). Besides offering them direct feedback on their usage, this also makes it more likely they'll receive the tool the first time around, since people are more willing to give someone a chance for a limited time. I forgot about editprotected, I think I'll add that to phase 1 trial. Thanks! Dcoetzee 09:27, 18 November 2011 (UTC)[reply]

According to previous comments where users were concerned about damage possibly caused by this I think it would be necessary to review nearly everyone, thanks for explantion, it makes sense Petrb (talk) 10:24, 18 November 2011 (UTC)[reply]

Just my two cents: In short, I don't think it is a good idea. Many technical-related reasons have been given in past discussions, so I will only mention the social ones: growing into adminship requires learning about various policies. Granting just one tools subset to a user we encourage them to specialize in this particular category and not try to match criteria for adminship. Also, on a more general level, adminship should not be a big deal. It does not require high-level technical skills at all anyway. Therefore, clearly users who have not passed RfA are not yet approved by the community not for their lack of knowledge or technical skill, but rather for not being proven and trustworthy. Giving them the selective possibility to delete/protect/block/etc. is probably not such a great idea then. Finally, apprenticeship is about learning from somebody who already mastered some lore, and it works. Just giving tools may, paradoxically, lead to less supervision and less learning :) Pundit|utter 22:26, 18 November 2011 (UTC)[reply]

Hmmm. I've considered a mandatory mentoring component, do you think it would improve the proposal? I want to avoid giving the impression that this is just about learning how to use the tools on a technical level - it's also about learning related policies in a hands-on way, and about exhibiting responsibility given access to dangerous tools. I also think a system with less discussion would be an attractive interim option for users who are good RfA candidates but don't yet feel ready to run. Overspecialization is a concern, but I tend to think the degree of specialization of a user is driven primarily by their interests. I'm speculating at this point but that's why I hope for a trial to help answer this kind of question. Dcoetzee 22:50, 18 November 2011 (UTC)[reply]

You might want to consider a mock trial to work out and study how well requests for tools could be handled. Since no tools would actually be handed out everyone involved would be simply playing along like those paramedic training days with the volunteer corpses. fg 22:58, 18 November 2011 (UTC)[reply]

You're right that all this is just speculation. I'm just not necessarily persuaded if adminship is scalable in the sense that I, for instance, either trust a person to be a good admin or not (and this trust applies to all tools in the administrative arsenal). I agree, of course, that such apprenticeship would help in developing more confidence in good RfA candidates, and also probably provide them with useful cred for the RfA itself. Yet, deep down, I don't see the significant added value component over the regular mentorship model already in place. Pundit|utter 00:31, 19 November 2011 (UTC)[reply]

You say that "Granting just one tools subset to a user we encourage them to ... not try to match criteria for adminship." There is no reason why tool apprenticeship should lead to adminship in all cases. A deleter, for example, may have no desire for other tools, and might not be interested in becoming an admin. Keep in mind that this proposed process would attract other users (particularly those without a broad range of experience in many different areas), not just those who have applied for RfA and failed. — This, that, and the other (talk) 02:03, 19 November 2011 (UTC)[reply]

• Btw, instead of selecting tools for limited adminship, one COULD consider probationary periods for people (not)passing RfA marginally. E.g. in case of somebody with 75% support, and a group of admins vouching and guaranteeing supervision, this person could get a 3-month adminship extended to regular edminship without a vote, just by approval of the mentors. But I don't believe that this idea could go through in the community :) Pundit|utter 01:10, 19 November 2011 (UTC)[reply]

There are quite a lot of people (myself included) who just don't want the full admin bit, for one reason or another (or many reasons). If someone were to nominate me for adminship, I'd decline it; but if it weren't possible to decline it I'd oppose myself on the basis that I don't need the whole mop! However, if people could have unbundled tools to apply in their own specific areas of choice / expertise, it would certainly reduce the workload on the full-admin types. We may want to have more active full-admins, but specialists are certainly not to be sneezed at just because they are specialists. To go back to the police parallel, they have all kinds of specialists, too - computer experts, scene-of-crime experts, you-name-it experts. What results is teamwork. Nothing to stop tool-apprentices applying for full adminship if they want to, at a later stage.

@Dcoetzee - you say that "over-specialisation is a concern" - why on Earth should it be? Consider your local large hospital.. It's crammed with people with extremely narrow areas of specialisation; neurosurgeons, orthopaedics surgeons, dermatologists, internal medicine specialists, plastic surgeons, cardiac surgeons ... the list is extensive! And because they specialise, they have extensive practice, experience and expertise in their own field. And they can work as a team (and often do), as well. I have to say I'm far happier about the idea of my ace neurosurgeon operating on my neck than I would be about the idea of my local GP (general / family practitioner) doing it! Pesky (talk …stalk!) 09:36, 19 November 2011 (UTC)[reply]

I haven't thought that there may be a large part of people not aspiring to adminship because of too many tools that come with it. If it indeed is so, the proposal makes quite a lot of sense. The comment about over-specialization is applicable only if we assume that limited adminship is just a step towards full status. Using your metaphor, to become a surgeon you still need to get your education in all areas of medicine. Pundit|utter 14:54, 19 November 2011 (UTC)[reply]

Thanks for the feedback - I agree that some users will definitely want to specialize and have no interest in becoming an admin, and I think this proposal will be attractive to them, so I revised a bit to emphasize that possibility. The reason I say overspecialization is a concern (as opposed to specialization, which is good) is that too much specialization could lead to requiring more people with more overhead to get certain tasks done. However, I think there will always be enough "general practitioners" (admins) to work on those cross-cutting problems, and that creating specialists to target the enormous workload present in limited areas will increase overall efficiency. Dcoetzee 15:24, 19 November 2011 (UTC)[reply]

Well, admins often have narrow specializations, too, but it is just up to their choice :) Pundit|utter 15:38, 19 November 2011 (UTC)[reply]

The obvious answer is to do what we already have with OS and CU - have lists of specialists, and a list of GPs. :o) Pesky (talk …stalk!) 22:26, 19 November 2011 (UTC)[reply]

Could you please create a list of tools which would be allowed to users? Probably we could list there also recommended requirements / trial time for each of them. Thanks Petrb (talk) 12:15, 20 November 2011 (UTC)[reply]

Here is my proposal:

• Block other users from editing

User should be active in antivandalism activities

• Change protection levels and edit protected pages

Users should be established as editors and should prove need to for a tool, be active in dealing with semi-protected edits etc.

• Delete and undelete specific revisions of pages
• Delete pages

Both should be given together, probably to people active in afd, csd, patrollers etc.

• Edit other users' CSS files
• Edit other users' JavaScript files
• Edit the user interface

All together - to people with technical knowledge assisting on support boards / dealing with updates of things like blacklists, or scripts, even helping to people whom own scripts are broken or causing harm

• Mass delete pages

IMHO no need to give to anyone

• Not create redirects from source pages when moving pages

For people in afc project

• Undelete a page
• View deleted history entries, without their associated text
• View deleted text and changes between deleted revisions

For people dealing with undelete requests from others

maybe you could insert your ideas or regroup it, thanks Petrb (talk) 12:25, 20 November 2011 (UTC)[reply]

So I propose that we have some set criteria to obtain this proposed userright. So I propose:

• 5000 edits
• 25 log entries
• Got admin coaching

Optional, but prefered:

• Has sysop access on 1 or more wiki (including test-adminiship on incubator)

Also, I propose that it will not include "unblockself", since it may be nessesary to block the user, and that the rights expires after 1 month.
~~Ebe123~~ → report on my contribs. 12:53, 20 November 2011 (UTC)[reply]