Jump to content

Wikipedia talk:Tool apprenticeship/Archive 2

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1Archive 2Archive 3

Not so sure about this...

A couple of issues from my standpoint:

  • When it comes to abuse, every admin is already a temporary admin — period. Some of the more jaded of the community might think of "temporary adminship" as a way to escape from the bureaucracy of de-adminning someone, but it won't be. It will have the same amount of drama associated with giving the flag, and it'll have at least thrice that when yanking it. After all, nobody wants the scarlet letter/kiss-of-death associated with permission revocation in their user rights log, so they'll fight to the death to keep that from happening. We already see that happening from time to time on even the seemingly trivial user rights (like rollbacker).
  • Unneeded bureaucracy/hoop-jumping/stress. In the past, users really only felt obligated to gain "x" number of edits, make good contribs, and demonstrate civilility/trustworthiness/policy-adeptness in their contribs. These were things easy for any editor to do without even knowing it, so it was easy for RFA nominations to come out of nowhere and with no real intent to run. That's how I got nominated a long time ago—a random admin happened upon me. Nowadays, the landscape seems a bit different. Users feel obligated to first obtain:
  1. rollbacker,
  2. reviewer,
  3. accountcreator,
  4. file mover,
  5. bump their edit count, and finally
  6. edit filter manager (which they're extremely unlikely to get anyway, as ineptitude in editing filters can wreak more havoc than +sysop).
Personally, I don't think we need to add yet another trophy like trial adminship for users to get before they can pass RFA. Not only that, but each and every one of those is at least partially stressful—if not extremely stressful—for a large portion of users.
  • It would de facto be required before RFA — So not only would they have to jump through the above hoops, they get an added bonus: they have to do it all over again. As if one time wasn't enough, a whole new gamut of people will be grilling them for another week of their life about anything and everything under the sun.
  • They'll be on their best behavior or simply do very little at all, and as such, the point of the "trial" is defeated.

Taken in concert, if I was faced with dealing with all of that, I'm really not sure if I'd have even cared enough to do so. There are so many places you can screw up, and so many people just waiting for you to do so, plus it feels like it demands even more of your personal time. That one's big for me. At times even I have a hard enough time finding time outside of work, and if I was a new user having to deal with the implicit obligation of "leveling up" constantly, I'm not sure Wikipedia would be as high as it is on my list of priorities.

Long story short, if the community is willing to trust someone with adminship for a week—with all of the problems they can cause in just minutes—then I say let's just let them have it as long as they don't abuse it—that's how things have always been.

--slakrtalk / 04:03, 10 December 2011 (UTC)

I appreciate your feedback, Slakr. I'll respond to your points individually.
every admin is already a temporary admin [...] It will have the same amount of drama associated with giving the flag, and it'll have at least thrice that when yanking it
Unlike with admins, the proposal gives bureaucrats the right to unilaterally remove flags from misbehaving apprentices (generally in response to a complaint on WP:BN). Although many have reasonably expressed concerns that discussions will turn into mini-RfAs, I believe the mechanisms set up to avert this (the restrictions on apprentices and discussion management) may be successful. A trial could help determine which will actually occur.
nobody wants the scarlet letter/kiss-of-death associated with permission revocation in their user rights log
Everybody makes mistakes, and I think a candidate with a proven history of growing and learning from their mistakes is all the stronger for it.
I don't think we need to add yet another trophy like trial adminship for users to get before they can pass RFA [...] It would de facto be required before RFA
As mentioned above, this is not at all the intention of the process, and I would oppose the institution of any such requirement. I realise it may be out of my hands, but I think the community does embrace good candidates even if they haven't done everything. [edit: I've added a counterargument to this at Wikipedia:Tool_apprenticeship#RfArequirement]
They'll be on their best behavior or simply do very little at all
The requirements are designed to attract people interested in getting real work done, not people angling for admin. Moreover, re-requests may not be granted if the tool was not used substantially, so they have a little encouragement to take chances.
A final point: I want to avoid referring to this as "temporary adminship," since that implies a scheme where all tools are borrowed, rather than just one or two, and implies that apprentices are admins (they are not). Let me know what you think, and thanks. Dcoetzee 04:55, 10 December 2011 (UTC)
The "they'll be on their best behaviour" objection rather reminds me of this. Tom Morris (talk) 16:56, 12 December 2011 (UTC)
I have to say I find the idea that it will become a defacto requirement for RFA quite likely and the thought is troubling. I'm sure when the criteria for speedy deletion were first proposed those proposing them did not believe they would become the final exam for administrative candidates, but that is in fact what has happened, intended or not. I can see it now "Oppose: get the block right and use it for three months, then come back and ask again. random jerk 06:31, 13 December 2011 (UTC)." I don't like the idea of that one bit. Beeblebrox (talk) 06:31, 13 December 2011 (UTC)
I agree that this is a troubling possibility. However, there are many unnecessary de facto requirements at RfA, and I don't think dismantling the useful processes behind these (such as CSD policy in the case you mention) is the right way to address the problem. I trust that others will seek to address unnecessary RfA requirements in a separate proposal. Dcoetzee 10:27, 13 December 2011 (UTC)
So, you agree it is potentially a serious problem but you are confident somebody else will fix it? Yeah, sureit is so very easy to fix behavioral problems at RFA, a real walk in the park... Beeblebrox (talk) 19:53, 13 December 2011 (UTC)
To clarify, I don't think this process would make that particular problem any better or worse. It would be just yet another of many unnecessary requirements, if this happened. We don't eliminate the featured article process because some people think RfA candidates should raise however many articles to featured status. Likewise, this process is designed to get important stuff done that has nothing to do with RfA. Dcoetzee 10:18, 14 December 2011 (UTC)
What important stuff? Show me the urgent backlogs that could be cleared by these partial admins. Where are they? Go to WP:AIV, go to Category:Speedy deletion, go to WP:AN3. What do you see? Backlogs? Nope. Hardly anything needing done. The areas where we still have backlogs are backwater areas that nobody cares much about. Guess what? Clearing out those backlogs will be as unappealing to partial admins as it is to real admins. This solution will not solve the backlog problem, because there isn't one in any area of critical importance. Beeblebrox (talk) 19:07, 17 December 2011 (UTC)
Ekhm, Media in category "All Wikipedia files with the same name on Wikimedia Commons" The following 200 files are in this category, out of 1,454 total., The following 200 files are in this category, out of 18,047 total.. See Category:Administrative backlog Bulwersator (talk) 11:03, 29 December 2011 (UTC)
That's the point exactly though, that the backlogs are in areas that "real admins", to use your words, would not be interested in. There's plenty of gnomes who would use the bit in one esoteric area but never pass RfA for want of vandalfighting experience etc. It's frustrating that this is being regarded as "partial adminship" or "temporary adminship", because the premise is not really to add another layer of bureaucracy to "before a RfA", but to open up tools for those who do not need the full set and would not be able to acquire it anyway. sonia20:02, 18 December 2011 (UTC)
I will readily admit that there are some non-urgent backlogs that as an admin I'm not interested in working on. One backlog that needs attention is AfD; those old ones are the ones that are generally so complex that many admins don't appear to want to close them. This is obvious, because a wrong close, although there is always DelRev and the admin is just doing his/her job, often incurs bitter complaints even including calls for summary removal of the tools. What this current RfC tends to forget is that adminship involves more than the physical use of a set of tools. As in the example of AfD (which can be closed by NAC), there are specific areas where good judgement is required, and is only expected from editors who are admins. Hence this "partial adminship" or "temporary adminship" (which it is) is flawed because it does not and cannot test those qualities in a way that RfA can and generally does. As I have previously mentioned, those who should pass at RfA generally do, and those who shouldn't generally don't, and close calls are fairly rare. Thus, RfA as a process is not flawed per se, however it has been broken by some voters who oppose with a vengeance, some whose vote is really a vote against adminship in general, some who simply troll and/or seek to cause drama, and others who oppose in good faith but get it it terribly wrong. The support section is not perfect either - there are plenty of clearly unresearched votes ('Support - an awesome editor'), and some that are the result of off-Wiki canvassing, and others are from users who are so new they don't even understand what adminship is all about. Cleaning up the current process would address these problems while "partial adminship" or "temporary adminship" does not, and whatever candidate selection process is involved, it will also partly be an (un)popularity contest and will expose the candidates to similar iniquities. What they get wrong will be taken down and used in the future, and for what they get right , there will be no suitable metrics. --Kudpung กุดผึ้ง (talk) 02:52, 19 December 2011 (UTC)

RfC: Should we begin a trial?

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


Should the proposed tool apprenticeship process, whereby experienced users with a need for an administrator tool may be given the tool on a trial basis, conduct a six-month three-month trial? Dcoetzee 05:32, 28 November 2011 (UTC)

  • Support as proposer. I believe tool apprenticeship will effectively allow users to gain experience with tools and related policy and combat our many backlogs, while keeping the level of scrutiny required low and so avoiding long discussions. Instead of speculating about future user performance, we can let them demonstrate it in a restricted setting. Dcoetzee 06:25, 28 November 2011 (UTC)
  • Support: RPP and AIV backlogs are becoming a daily occurence.Jasper Deng (talk) 05:48, 28 November 2011 (UTC)
  • Oppose. No way for community to vet potential candidates before granting tools, too open for abuse, too many convoluted "rules" to qualify for tools. Effort would be better spent on fixing RFA than this clusterfuck. --Jayron32 05:54, 28 November 2011 (UTC)
    • RFA is a little too stressful, because the user must get vetted for all, not just one of, the tools. There are only three rules, no more than the 3 parameters of WP:SOCK. WP:RFA2011 has not produced anything the community can agree on so far (and it will soon be RFA2012 if it doesn't do so soon).Jasper Deng (talk) 05:57, 28 November 2011 (UTC)
      • Still not convinced. Anyone who should qualify for any one of the tools should qualify for all, and if RFA is broken, this is not the solution. I fail to see how any user who is trustworthy enough to block another user is somehow not trustworthy enough to delete a page. --Jayron32 06:01, 28 November 2011 (UTC)
        • I probably know enough to block LTAs, socks, and vandals, but not enough to evaluate NFCCs or CSDs (A7), for instance.Jasper Deng (talk) 06:04, 28 November 2011 (UTC)
          • Knowledge is not my concern. Trust of the community is. If a person has the trust of the community to not abuse the block tool, then they also should have the trust of the community to not abuse the delete tool, and that includes not using the tool when they don't understand the underlying issues. Adminship is about having the faith of the community to make good decisions, and nothing more. This process does not (IMHO) have an effective method to guage that, and a user is either trustworthy or not. Period. If I would not trust them to use any one of the tools, I wouldn't trust them to use any. --Jayron32 06:12, 28 November 2011 (UTC)
            • I don't see how the trust must be determined by 1 single tool (I've seen lots of RfAs fail due to one minor flaw).Jasper Deng (talk) 06:18, 28 November 2011 (UTC)
            • Hi Jayron32, thanks for your feedback. I give my opinion on trust under Wikipedia:Tool_apprenticeship#Responses_to_anticipated_objections (trust does not always transfer between areas; trust should be based on experience). The many limitations placed on apprentices are also designed to reduce the amount of trust needed to approve their trial (probation, time-limited trial, other optional restrictions). Consensus discussion at Wikipedia:Requests for tool apprenticeship gives the community the opportunity to vet candidates, but my expectation is that not as much scrutiny would be required as for a full RfA. Dcoetzee 06:21, 28 November 2011 (UTC)
              • I think we're going to have to agree to disagree on trustworthiness; for me trustworthiness applies to the person and not the task: either I believe you are going to work in good faith and have the competency to produce good results, or you don't. A trustworthy person also has the self-awareness of their own competency: I wouldn't trust a person with ANY admin tool if I didn't believe they would have the wherewithal to avoid using the tool in situations where they didn't have a good understanding. For example: I don't have the edit-filter user right, even though I am an admin and could grant it to myself. Why haven't I? Because I don't have a clue how to operate the black magic that is the edit filter. I would expect any admin to have the self-awareness to use any tool similarly; if they didn't have it, I wouldn't trust them with ANY of the tools. The other issue, the "many limitations placed on apprentices" is what makes the process so convoluted and unworkable. You've replaced "trust of the community" with a complex, convoluted, and messy network of requirements, conditions, and rules which don't make a workable process, IMHO. There is no solid substitute for the trust of the community, and the only way to guage it is via a community discussion, not a set of unworkable "checks and balances". As I admited, RFA is broken; this process doesn't fix it, it merely takes the broken pieces of RFA and smashes them even smaller. --Jayron32 06:34, 28 November 2011 (UTC)
I understand what you mean, but I think that this feature would be helpful for current process as well, this would help to people who would be good sysops, but aren't established enough (I know many people who are good candidates for sysop but they don't believe that they would pass since there are too many people who could oppose, whatever the reason would be), and this would allow them to prove that they are good candidates for rfa later, so that it can improve the current rfa procedure. It's not about separating sysop bit to many "smaller" groups just to create some security issue, but rather to give the opportunity to the people who would like to be sysops in the future to prove that they would be good in that. Petrb (talk) 09:43, 28 November 2011 (UTC)
@Jayron32: I was thinking about your "edit filter" argument and here's my counterargument: if admins are people we "trust to know their own limitations," why do we not simply give all of them bureaucrat + checkuser + steward? The answer is that we don't trust them with everything, and with especially dangerous tools extra caution is warranted. I think a similar analogy applies here. I don't think the procedure is excessively complex (RfA, in fact, has more complex requirements). Dcoetzee 04:08, 29 November 2011 (UTC)
Dcoetzee, I don't appreciate this blanket claim of 'distrust', as I have attempted to explain without success on your talk page through several metaphors and analogies. In real life we have policemen, but this does not automatically infer that every citizen who is not a cop is a potential criminal. And a lack of trust is not the reason either why every traffic cop can't be a chief constable. --Kudpung กุดผึ้ง (talk) 06:05, 2 December 2011 (UTC)
I think we'll just have to agree to disagree here. I've got into quite a few arguments on this page, and I don't think either side will get the other on trust.Jasper Deng (talk) 06:09, 2 December 2011 (UTC)
Hi Kudpung, I explained this misunderstanding in some detail on my talk page, and reworded the proposal page accordingly. I never meant to imply that non-admins (or non-bureaucrats) are not trustworthy people, or should be treated with suspicion. That's what WP:AGF is all about. But we quite reasonably take care in evaluating candidates before giving them access to dangerous tools, just as the police do a background check on job applicants. In real life, police also train new staff in the use of their weapon according to the law and department policy, and tool apprenticeship is a natural analogy to this. Dcoetzee 02:25, 3 December 2011 (UTC)
An extension to that analogy is the paradox that while in the USA citizens have a constitutional right to bear arms, in the UK they have to provide a very good reason for wanting one, and 'not trusting my workmates' or 'just in case' are not on the list. In the UK even the normal patrolman is not armed, and neither are our vandal patrollers who actually do a pretty good job without weapons. Kudpung กุดผึ้ง (talk) 02:37, 3 December 2011 (UTC)
  • Support - I don't remember the last time CAT:CSD was empty. →Στc. 05:59, 28 November 2011 (UTC)
  • Somewhat opposed: vetting is important. As far as I can see, this proposal would only create a new class of user rights (or pseudo-rights, whatever). We cannot bypass the community. Sysop actions affect new (and old) Wikipedians daily, and the buttons should be granted only after a thorough process. I do believe RfA needs significant improvement, but I don't believe handing out semi-adminship or so-called "conditional"-adminship is the way to go forward. CharlieEchoTango (talk) 06:02, 28 November 2011 (UTC)
  • Support - some editors (like myself) just don't want the full admin bit (so are highly unlikely to put themselves through the stress of RfA for it!), but could make good use of a limited subset of tools. Personally I could make good use of delete (& undelete) when I'm on NPP, but that would be about all I;d want. Pesky (talkstalk!) 06:16, 28 November 2011 (UTC)
  • Neutral Still have my own doubts about the efficacy of such a system, but we'll see. I still see potential for abuse, though. —James (TalkContribs) • 4:23pm 06:23, 28 November 2011 (UTC)
    • Changed to oppose Oppose It is my belief that if a user is competent enough to handle a subset of tools, they should have the competence to handle the full bundle, bearing in mind that the RfA Reforms proposal is ongoing and would improve the battlefield nature of RfA. To me, this would encourage second-best. At this time, I cannot support the proposal. —James (TalkContribs) • 11:50am 01:50, 7 December 2011 (UTC)
      • Changed to support Weak Support I'm still skeptical about how the trial would work, but most of my concerns and doubts have in large been allayed. Though, I'm still of the opinion that a user who can be trusted with a small subset of tools should be of the same level of competence and knowledge to handle the full bundle, we should be able to trust users 100%, no less. —James (TalkContribs) • 12:11am 14:11, 11 December 2011 (UTC)
  • Oppose. Anyone who can use the sysop tools competently, with a mature sense of judgement and civility would pass RfA with flying colours, and an appreticeship to admininship would possibly only be considered by many as shortcut to what they consider to be an award for good work. While it is undeniable that RfA is most certainly an extremely unpleasant experience in many cases, it nevertheless continues to demonstrate that generally among those who are prepared to be subjected to it, those who should pass do, and those who are not ready for adminship don't.
Of even greater concern is the danger that a 7-day consensus discussion at Wikipedia:Requests for tool apprenticeship would be just as unpleasant, and would simply multiply the number of occasions for some individuals to be disparaging to the candidates and each other with impunity.
Current backlogs are not (yet) a deep cause for concern, and while reform of the RfA process may seem slow to some, so is the discussion and implementation of all major reforms and new policy, such as for example improvements to NPP that are being discussed since August 2010 that may or may not become a user right, that might or might not include some very limited powers of deletion. WP:RFA2011 is still in the process of gathering statistics and feedback in order to prioritise the many suggestions that have been made for RfA reform; there is no question that the task force is unable to agree on anything. --Kudpung กุดผึ้ง (talk) 07:44, 28 November 2011 (UTC)
    • No, people look (unnecessarily) for content too, and I think RFA2011 needs to try one of its radical alternatives.Jasper Deng (talk) 23:20, 28 November 2011 (UTC)
    • Hi Kudpung, thanks for your feedback. Some responses: this proposal is not actually about "an apprenticeship to adminship," exactly - it's not even possible during the trial for the user to acquire more than a few tools this way. It serves two groups of users: one, users that need just a couple tools to do work in a particular limited area, and two, users that want to develop experience with and show responsibility using tools prior to their RfA. Even if these users would in fact pass RfA, many feel more comfortable running if they have an opportunity to develop more experience first. Users can still pursue RfA directly without using it. I don't know whether the consensus discussion would be unpleasant, but the goal was to limit discussion by imposing restrictions on apprentices, so that less scrutiny is needed. Dcoetzee 11:33, 28 November 2011 (UTC)
      • @Jasper: Radical alternatives are not part of the manifesto of RFA2011. The project is concerned specifically with improvements - namely the level of civility and the quality, accuracy, and honesty of voting - to the current system. Contributors to the project were ignoring this primary goal, hence the sub-section was created to provide them with a forum for their ideas and suggestions for other systems, and at the same time enabling them to benefit from the enormous collection of resources that have been created for the project and are quickly available from the project's navigation.
      • @Dcoetzee: Tool apprenticeship is a euphemism for 'tool unbundling'- for however long, and is a solution for the wrong problems: there is no proof that the general nastiness of the current system is in any way related to fears of the tools being granted for life. To substantiate such claims, one must consider the broader picture, some numerical evidence should be presented that demonstrates the actual tenure and actual number of tool related operations of admins who have been appointed, together with graphs of their performance after the euphoria of being 'promoted'; of the 53 withdrawn rights, around 50% were desyoped by arbcom in the last 3 years, while 277 (to date) have been procedurally desyoped for inactivity, while of the 142 resignations, around 33% were during the last two years. The actual criterion (30 edits in the last 2 months - about 50% of the round 1,500) of what is considered an 'active' admin, is ridiculously low, nevertheless, in spite of the concern that that fewer and fewer candidates are now presenting, there are still no serious backlogs and the point at which the present proposal is needed - or any other radical solution - is a long way off being reached, while suggested measures to clean up the kind of voting that is keeping candidates away may soon be gaining more thrust.
Furthermore, an in-depth, neutral and objective study of New Page Patrol is shortly to be published by the WMF) that may or may not (depending on the way it will be interpreted by the community) lead to proposals for some of the rights discussed here being accorded to patrollers of the right calibre. Kudpung กุดผึ้ง (talk) 13:10, 29 November 2011 (UTC)
Hi Kudpung, I'll start out by saying that I absolutely think the less radical RfA reform proposals you mention also have merit independent of this proposal, and I would be happy to examine them as they arise. I also agree that, despite my efforts to prevent it, it is a real possibility that the discussions in this proposal may devolve into discussions resembling the current RfA. Although I could take a data-driven approach as you suggest to predict whether this would occur, in light of the relatively low risk of the outlined trial, I think this is better tested by experiment. I also plan to be operating "in the trenches" to help moderate discussion and do my best to prevent this outcome. I understand if you remain skeptical that this is possible. Dcoetzee 03:45, 30 November 2011 (UTC)
  • Oppose at this time Supporting since parameters changed As careful as Dcoetzee has been putting this proposal together, I still have the feeling this will create far more mess than it aims to solve. Very oddly I must also oppose on grounds laid out by one of the proposals proponents. See above "...we have to reduce that standard..." followed by my suggestion that we could instead "...rise to meet it" followed astonishingly but pointedly by "...we have to lower the standard...". I can't support lowering standards, and I feel this is exactly what is being proposed. Basically a shortcut around thankfully tight guardianship of powerful tools. fredgandt 08:01, 28 November 2011 (UTC)
    • No-one's rising to meet it, so we have to drop our part of it.Jasper Deng (talk) 23:20, 28 November 2011 (UTC)
      • I've worked in aerospace engineering and electronics and now in web development (noob). In these fields we have standards or tolerances. I've built circuitry so small it's components are fixed under microscopes and even made nuts used in part to hold the wings on the Eurofighter. If my colleagues and I were to have lowered our standards because we couldn't meet them, planes would be falling out of the sky and circuitry would be fizzing to death as we speak. Relaxing standards to fit the unfit is frankly retarded. I expressed these concerns earlier (near the top of this page), so don't think I'm picking on you Jasper. You have just reinforced my concerns. fredgandt 00:01, 29 November 2011 (UTC)
        • This is not quite the same as aerospace engineering. It isn't like we could lower the pass vote by 5 percentage points and have rogue admins horsing around within a few months.Jasper Deng (talk) 03:04, 29 November 2011 (UTC)
          • I want to clarify that I don't support Jasper's ideas that we need to "lower our standards." I think our standards for adminship should stay right where they are. Rather, I'm pushing the idea that when less tools are given with more restrictions (time limit, probation, other optional restrictions), naturally standards don't have to be as high, because the potential for abuse is much lower. Dcoetzee 05:17, 29 November 2011 (UTC)
  • Support Excellent idea + if you needed any help with mediawiki modifications in order to make this possible, I would be very happy to help you as dev. Petrb (talk) 08:26, 28 November 2011 (UTC)
  • Support. There are many users who can be trusted with administrative tools, but who lack the broad experience in many areas which is required at RFA, or simply may not want to become a full admin for various reasons. This process is perfect for this sort of user.
    From the perspective of others who may oppose this on grounds of lack of trust - that is why there is a built-in probation period: if someone is found to be abusing tools under this process during a trial period, the tools can be instantly revoked. The term "high standards" is vague, and I believe this process requires high standards of tool apprentices, in quite a different way to the way in which RFA requires high standards. — This, that, and the other (talk) 08:40, 28 November 2011 (UTC)
  • Oppose. Mostly per Jayron32, and because I just don't subscribe to the "RfA is broken" myth. This isn't needed. -- œ 09:11, 28 November 2011 (UTC)
    • Neutral. Still skeptical, but after discussing it and looking at from another's perspective, I (grudgingly) agree that change can be a good thing and there's no harm in giving it a try. I'll sit back and watch how the trial goes, if it turns out to be a kind of paradigm shift in Wikipedia administration then great, if it crashes and burns, then that's great too. -- œ 02:52, 29 November 2011 (UTC)
    • It's not RfA, it's the way it's being carried out here.Jasper Deng (talk) 23:20, 28 November 2011 (UTC)
  • Support test round to see how it goes. Pinetalk 09:44, 28 November 2011 (UTC)
  • Support this small test - I think there is very little danger from the proposed trial, as it will be closely watched; I think it extremely unlikely it would cause any significant problems. And at the end of that small test, if nothing else, I think we will have learned a lot - which may lead to further ideas. Plus the test subjects might have a better shot at RfA (if they want). Thus, the test itself seems to be a net benefit, whether it proves successful or not. RfA is broken, totally.  Chzz  ►  10:05, 28 November 2011 (UTC)
  • Query What's to stop an editor who is in good standing - but crucially has already failed an RfA - from requesting all the tools in a procedural manner? The editor clearly does not have the trust of the community, but this less high profile discussion may mean it is not noticed. I do like the idea and the limited period offsets the risk, but I do think it could be abused to the extent that it snubs current community discussion. WormTT · (talk) 11:33, 28 November 2011 (UTC)
    • Hi Wtt. If you're asking whether a user could make a request for all tools, such a request would almost certainly be rejected, since a single request has to be aligned with a single task, and no task involves all tools (and the more tools a request involves, the more scrutiny is applied). If you're asking whether a user could accumulate many tools slowly over time: not under the process trial, which won't grant tools permanently. With the full process this might be a possibility, but it would take years and involve more effort (and community discussion) than an RfA ever would. Dcoetzee 11:48, 28 November 2011 (UTC)
      • Well, that does allay at least one of my fears - I'll Tentatively Support a trial. I'd like to see that due care is being taken over the community discussion. For example, I don't want to see someone with a history of being uncivil being given access to the delete button as deletions often have to be discussed with new users who do not understand why they should not have created that article. I'd also like to see some sort of throttle during the period of the trial, so they can be checked. All in all though, I think this would be a positive - it worked for rollback (ish). WormTT · (talk) 11:55, 28 November 2011 (UTC)
        • Thank you for your support - I'm not sure if a throttle is currently technically possible, but that could certainly be made a condition of a user's trial. I agree re: due care in discussions, and moreover whenever possible I like to see concrete experience - in your example, I'd like to see them have a history of not only civility, but actually tagging articles for speedy deletion and subsequently discussing their taggings with affected users. It's important to remember the social skills. Dcoetzee 12:12, 28 November 2011 (UTC)
    • @Worm: Content work is regarded as important at RfA too, and it is often unrelated to the judgemental qualities of a user.Jasper Deng (talk) 23:20, 28 November 2011 (UTC)
    Content work is indeed regarded at RfA. I think people compare the candidate to themselves and point out areas the candidate seems lacking. Now, if the majority of work they do is content work, it's hardly surprising that it's checked. This then means the editor gives a measure of trust to the candidate and supports based upon that. I like this proposal for thinking "outside the box" of RfA, only giving relevant tools. I worry slightly as all the repercussions of the tools may not have been seen - and I do think people who have never written content should not be deleting it - so I still tentatively support. WormTT · (talk) 07:39, 29 November 2011 (UTC)
  • I think this is a decent idea, and don't see the major harm conducting a trial will do. Don't count this as a support or oppose, more an observation. Steven Zhang The clock is ticking.... 11:41, 28 November 2011 (UTC)
  • Oppose for now I don't know whether this will be a good thing in the long term or not, but that is exactly what a trial is for. One of the proposed restrictions on the trail was for there to be a maximum number of concurrent test subjects (for want of a much better term), and that is something I strongly support. My oppose is based on the lack of explicit criteria by which the trial will be judged a success or failure, which was one of the reasons the flagged revisions trial ended with the drama it did. Once the criteria are agreed, then I'll have no problem supporting a trial. Thryduulf (talk) 12:29, 28 November 2011 (UTC)
  • Support the trial, nothing more. After the trial there should be a "study" how successful the trial was and what changed. (that means we need something like a monitoring of the related backlogs) Similar to other supported I think that although many non-admins will success a RFA, they simply don't have the courage to do the RFA, or the need of the full sysop-tools. They shouldn't affected in improving this encyclopedia by limits of their rights. Also the argument that the community should trust in their admins and thus giving them the full tools might not be correct: as correctly stated above: edit filters and other areas need knowledge and the trust is normally given in gaining respect. This can't be transfered (in my eyes) to other areas. mabdul 12:56, 28 November 2011 (UTC)
  • I like the idea of unbundling the block button, at least to deal with vandalism only accounts. I'm less comfortable with a generic unbundling as some things need to be used in combination, for example you shouldn't appoint an Autopatroller without checking their deleted contributions. I'm uncomfortable with the delete button being unbundled as this is an area where I fear mistakes are already made. Also the most pressing urgency is at AIV most other admin areas can handle having hours when there is no-one on call. At the very least an RFC should have the option of unbundling specific tools rather than all of them. WereSpielChequers (talk · contribs) 13:27, 28 November 2011 (UTC)
    • Hi WereSpielChequers, thanks for your comment. I want to clarify that users can request multiple tools, if they are needed for a particular task. A request that does not include all necessary tools for performing the task would not be approved. The full unbundling is intended to offer maximum flexibility in the types of tasks that could be served. Re deletion: I am aware that administrators make errors while deleting, but some non-admins have exceedingly good judgement in this limited area, and the tool would be revoked if they misuse it. I agree that AIV is a great place for apprentices to contribute. Dcoetzee 06:52, 29 November 2011 (UTC)
      • Thanks for the reassurance re combinations, I've struck that part as resolved. In my view non-admins who are ready for the deletion button should run at RFA, as I'm uncomfortable handing out the delete button lightly. I'm more comfortable unbundling some other tools, and I see our most pressing need as the gaps at AIV. So I'm keen that we do this for the block/unblock buttons - though perhaps with the proviso that if you aren't a full admin your block button should only work on IPs and accounts with fewer than 100 edits. I also think that any unbundling needs to include the ability to fix ones mistakes, so the unblock tool goes with the block tool and the ability to remove Autopatroller flags with the ability to set them. ϢereSpielChequers 09:58, 1 December 2011 (UTC)
        • This worries me: I want to clarify that users can request multiple tools, if they are needed for a particular task; the actual number of tools is not many, so there is no basic difference here than applying for the full set through the regular RfA process. At least RfA takes into account the performance that is not measured by the potential use of the tools themselves, such as maturity, civility, and a keen sense of judgement. Kudpung กุดผึ้ง (talk) 05:50, 2 December 2011 (UTC)
          • This is an understandable concern. When I say "multiple" the case I have in mind is vandalism patrol, which requires heavy use of both protect and block. I don't expect requests for more than two tools to ever be necessary or get approved - and even requests for two tools would be treated with more scrutiny. Just to make this more obvious, I'm restricting it to at most two as a condition of the trial. Dcoetzee 02:36, 3 December 2011 (UTC)
  • Support as a trial only. It seems to me that, although the concept is admittedly a work in progress, any harm that may result during the trial as proposed will be readily correctable, with no irreversible harm, and only a low risk of a lot of hassle in reversing anything. I think that it addresses a genuine need, and is the kind of improvement we should explore, rather than automatically rejecting as generally happens with administrator-related reform. But it should only be a trial at this point, subject to scrapping if it doesn't satisfy the community at the end of the trial period. About that period, I also note that the proposal seems to say that users would be given tools for a maximum of one month at a time, but then goes on to describe a scenario in which the term is three months. I'd like it to be clear that there is a one-month limit. --Tryptofish (talk) 15:21, 28 November 2011 (UTC)
  • Oppose the trial until it is made clear what happens at the end of the trial. I'm not fundamentally opposed to a trial, but after the disaster of the end of flagged revision trial, it needs to be clear that all trials should stop at the end of the 6 months and that afterwards a clear consensus is needed for either a further trial, or for rolling out generally. The limit of 10 at a time should be enforced strictly until the end of the trial. Also I know the trial does not include giving any right permamently, but when/if the trial happens and is successful, there must be a way to remove the right(s) from users who have been granted them "permamently" without having to go to Arbcom. I will not support establishing any more rights that can only be removed by Arbcom. Davewild (talk) 16:02, 28 November 2011 (UTC)
  • Oppose - I see no reason to over-complicate an already imperfect RfA process. The issue of trust has not been sufficiently answered for me. Trust is not about whether we think a potential admin would be able avoid accidentally/maliciously deleting the main page, it is about whether, as a community, we trust them to have the required discernment and judgement to properly use the tools given to them and to not use tools which they are not confident in using. RfA is more than "will this user break anything?" (which is how other requests for permission, such as rollback, work); RfA is asking "does this admin have the right skills to undertake the role of adminship, both technical ability and their reasoning, judgement and discernment?". I'm not opposed to releasing more powers from adminship (such as rollback) per se, but this should not be party of the RfA process. Whereas requests for permissions is a simple question of whether someone will use a tool properly, RfA carries a lot more responsibility. Admins are given the right to close AfD discussions, for example, and (in most cases) users cannot do this. The reason that I, as a non-admin, cannot close an AfD discussion as keep (without near-unanimous consensus) is not to do with the technical abilities I have - technically, I could close any AfD as keep/redirect - it is because the community has not elected me as a trusted individual. Therefore, adminship is about community trust; tool apprenticeship is about a few tools that some people might find useful - these are very distinct things. ItsZippy (talkcontributions) 17:06, 28 November 2011 (UTC)
    • This is not part of RfA, and is not intended to be. It's about training people to gain trust.Jasper Deng (talk) 23:20, 28 November 2011 (UTC)
    • Hi ItsZippy, thanks for your feedback. As Jasper noted, tool apprenticeship is not part of RfA but a supplementary and independent process. You are quite correct about AfD discussions, and that is why we would not give the delete tool to apprentices for closing AfD discussions (unless the community later chose to allow apprentices to close AfD discussions). In general, any place that policy says "an administrator must do this" now, apprentices will not be eligible to do those things without a change in policy, because they are not administrators. I have amended the proposal to clarify this point. Dcoetzee 05:37, 29 November 2011 (UTC)
  • Oppose. Getting the admin tools has always been a trial-and-error process, and that's what made it good. Adding this in would not only complicate things and add more bureaucracy, but would probably end adminship. Wizardman Operation Big Bear 17:26, 28 November 2011 (UTC)
    • It wouldn't end adminship, as that is a permanent right, while these are temporary and limited rights.Jasper Deng (talk) 23:20, 28 November 2011 (UTC)
    • Hi Wizardman, I understand your concerns about apprentices displacing administrators. My expectation is that, to the contrary, apprentices who have more experience with tools and related policy will feel more confident running for administrator later, increasing the number of admins. The trial will help to evaluate this impression in practice. Dcoetzee 06:11, 29 November 2011 (UTC)
  • Support trial. This seems to me to be an excellent way to put the tools into the hands of people who are willing to do the work. We need more people doing the work. I'm hopeful that something like this would be more of a feeder to adminship than a challenge to it. Spending some time doing similar tasks can give people experience and confidence to undergo RFA. :) Before a trial would begin, I would like to see clear consensus, though, on how tools are to be "revoked for blatant misuse". Is there a discussion, or is this a single administrator's opinion and action? If the latter, does discussion follow? Is removal permanent or just until the misunderstanding that led to blatant misuse (assuming it's not "blatant abuse") is worked out? I like that the permission follows a discussion period; I have seen users granted "autopatrolled" status, for instance, in spite of multiple copyright warnings on their page, and I would be uneasy with this if it were quite as lightweight as that. I also wonder if there would be a possible way of tracking usage of the tools. Could we create, for example, a Special:Log/delete for people who have this tool so that admins might do spot-checks? --Moonriddengirl (talk) 19:08, 28 November 2011 (UTC)
    • Hi MRG, thanks so much for your feedback. I've added a section Wikipedia:Tool_apprenticeship#Probation to address your questions that I think people should be able to agree on. I think it's okay to have removal be rapid and without discussion, since the user can issue a new request if it was made in error (I think it goes without saying that discussing the issue with the user either before or after removal is important). The idea of tools to help audit apprentices is an excellent one and something I could work on myself if the trial is approved. Dcoetzee 05:03, 29 November 2011 (UTC)
  • Support as a trial only, with a one month limit. I'd be happy for the proposed limit of ten simultaneous apprentices to be raised at least a little bit.-gadfium 19:56, 28 November 2011 (UTC)
  • weak oppose though a shorter trial (1-2 months) would move me to neutral. I think, per Wizardman, that this will have significant long-term impacts on the wikipedia culture. Not sure if those would be good or bad though. Hobit (talk) 20:21, 28 November 2011 (UTC)
    • Hi Hobit, thanks for your feedback. I'm certainly willing to consider a shorter trial, but I want to make sure that users have the opportunity to request renewal/extension of a trial at least once, since this is one of the primary feedback mechanisms of the system. A 3-month trial would be long enough for this. Would you consider this acceptable? Alternatively I could consider making tool trials shorter, perhaps 2-3 weeks. Dcoetzee 06:22, 29 November 2011 (UTC)
      • Either would be acceptable to me. After the pending changes trial I have serious issues with trials. Throw in a fairly clear statement that after the trial is scheduled to end the trial _will_ end unless there is consensus otherwise and I'll move weak support. I'm not sure it's needed, I'm not sure it's wise, but I'm willing to give it a shot as long as the trial duration is short and well defined. Hobit (talk) 04:31, 30 November 2011 (UTC)
        • Hi Hobit, I agree that a shorter trial is a better starting point. I've made it 3 months, and added all manner of language making it as clear as I possibly could that the trial will really, really end after 3 months. Hopefully this will ease the concerns of those who still remember the botched Pending Changes trial. :-) Dcoetzee 06:46, 30 November 2011 (UTC)
          • Changed to weak support. I think this is probably a bad idea, but I see little harm in testing the waters if the trial is shorter and now that it has a well-defined ending. Hobit (talk) 18:29, 30 November 2011 (UTC)
  • Support: this is a great idea as one user who would love to help out in page protecting where there is always a backlog. A trial will not cause harm to the community and no irreparable damage can be done by this. Edinburgh Wanderer 20:59, 28 November 2011 (UTC)
  • Support. It sounds like a useful trial, seeing whether some tasks can be done without being a "superuser" capable of passing the increasingly high bar of RfA is a great idea. If it doesn't work, don't extend it beyond the trial. Fences&Windows 22:03, 28 November 2011 (UTC)
  • Oppose We don't need "apprenticeships" on Wikipedia. That's yet another bureaucratic sublevel for eager newbies to get promoted to. There is no real shortage of admins that this proposal will resolve. /ƒETCHCOMMS/ 22:05, 28 November 2011 (UTC)
    • No, this is not intended for newbies at all. The gist is that there are many experienced users who are afraid to get into the other stocks at RfA to get the tools, and giving the tools gradually would remedy that.Jasper Deng (talk) 23:02, 28 November 2011 (UTC)
      • I didn't mean it was for users when they were newbies. I said newbies would see it as some "level-up" achievement. Either way, admin tasks are not mutually exclusive; thus, while some users may be competent enough to block a vandal or sockpuppet, the fact that they are not confident in their CSD abilities or file policy knowledge does not instill in me any level of trust in their competence to wield any sort of administrative tool. /ƒETCHCOMMS/ 00:58, 29 November 2011 (UTC)
        • Hi Fetchcomms, thanks for your feedback. Regarding the status-seekers, User:M.O.X also raised this concern earlier. A user who is clearly not ready would not be approved; while a user who is just pretending to be active, experienced, and responsible in an area, would ultimately "become the mask" and actually accomplish things. Regarding your final point: I really believe a user can confidently and responsibly wield tools in a limited area without a complete understanding of all areas (indeed many admins fit this description). I believe a trial would help to demonstrate this. Dcoetzee 07:20, 29 November 2011 (UTC)
          • I disagree with the last part of what you wrote. I think it's very concerning that some admins still do not understand the non-free content policy or other important policies/guidelines, yet these are the people that new users are going to look up to for advice. /ƒETCHCOMMS/ 15:21, 29 November 2011 (UTC)
            • I won't discuss here the degree to which admins should understand policy comprehensively, but the terminology of apprentice was chosen to emphasize that these users are "in training" and not as generally experienced as full administrators, a message that I think new users will understand, especially in combination with other factors. I also hope that by acting as a stepping stone to RfA that in the long term, new users will have a greater number of admins to choose from when seeking advice. Dcoetzee 04:03, 30 November 2011 (UTC)
    • The "apprentice" part concerns me especially because I generally have less trust for apprentices. There's plenty of connections to be made with the simple viewdeleted right, and that's at least one privilege I don't feel comfortable letting less-experienced users have. /ƒETCHCOMMS/ 21:13, 30 November 2011 (UTC)
      • The viewdeleted privilege would not be available in this trial, and I don't think it's really that important to warrant apprenticeship.Jasper Deng (talk) 00:37, 1 December 2011 (UTC)
        • So these apprentices would be able to delete pages, but not be able to view the pages they've deleted? That seems ludicrous. Jenks24 (talk) 01:12, 1 December 2011 (UTC)
          • I'm sure the apprentices would remember why they CSD'd a page. If something's wrong, the page creator can be forwarded to a real admin. They have no other reason to view deleted stuff (I wouldn't want to be able to view RD2'd material (revdeleted)).Jasper Deng (talk) 01:16, 1 December 2011 (UTC)
          • Yes, this is related to Wikipedia:Tool_apprenticeship#undo. Apprentices, like any other user, can seek help when they need tools they don't have. If they give clear deletion reasons, as they should be doing anyway, this will largely obviate the need to view the deleted material. Dcoetzee 02:59, 1 December 2011 (UTC)
        • Doesn't that seem a little ... you know ... inefficient? A regular admin can undo a deletion themselves, whereas an apprentice would need to find someone else every time they made a mistake. That's especially annoying for an article's creator, Jasper Deng, and could very well make them think Wikipedia is too difficult to be worth their time. What sort of website has workers that can delete things but make you ask someone else if you want it undeleted? /ƒETCHCOMMS/ 22:17, 1 December 2011 (UTC)
          • Re "make you ask someone else" - if an apprentice realised their deletion was unjustified, they would not send the hapless article creator to requests for undeletion. They would simply ask an admin on their talk page to undo the action. The admin would review the situation and do so if it was appropriate. The burden on article authors would not be increased, and the overhead would be quite low. It is an annoyance, but they shouldn't be making mistakes too often (if they are, they should really get their tool revoked). Dcoetzee 22:23, 1 December 2011 (UTC)
  • Question How is this supposed to work? I thought that it was technically impossible to grant some admin rights but not others, and (unless I missed it) this proposal doesn't address the technical side of things at all. Nyttend (talk) 22:19, 28 November 2011 (UTC)
  • To those crying "bureaucracy, no!!": The cliff one has to scale to get admin tools is too big. It's like trying to become a black belt without any other belts, at a time when people with black belts are urgently needed for the military. True, some people will do it in one fell swoop, but no, most people will have to use an intervening level. Since this will be open to only a limited number of people at a time, there's no real "bureaucracy" here.Jasper Deng (talk) 23:09, 28 November 2011 (UTC)
    • Your metaphor is flawed in that anyone can "earn" other "belts" at any time—e.g., writing articles, participating at AfDs, being a civil and rational voice in discussions. The bureaucracy is in the implementation but also the concept, because it further complicates a very simple idea. /ƒETCHCOMMS/ 05:16, 29 November 2011 (UTC)
      • No-one decides to go that far (my point). If we were a simple website, we would have only one or two privileged user groups, instead of the four or five we have right now. And, we need to encourage adminship, not just have an open door to it. Wikipedia did not become the world's 5th-most-visited site without some seed articles to attract users to edit.Jasper Deng (talk) 05:24, 29 November 2011 (UTC)
    • That's your opinion, and I disagree. I think RfA is not difficult if a user puts their mind to it, and encouragement isn't necessary for qualified users who want to become admins. If someone's afraid of being publicly criticized, then being an admin is definitely not going to help. /ƒETCHCOMMS/ 22:21, 29 November 2011 (UTC)
  • Support the trial. It may be worthwhile to see how it works, since there are pros and cons, both anticipated and unanticipated, which will be fully visible once we understand the mechanics of this solution. Pundit|utter 00:39, 29 November 2011 (UTC)
  • Strongly oppose This would simply allow inexperienced, eager, power-hungry users to get buttons that they're sure to abuse. See comment below. HurricaneFan25 01:10, 29 November 2011 (UTC)
  • Do you have no faith in the judgement of the bureaucrats who are to close the proposed requests? →Στc. 02:57, 29 November 2011 (UTC)
  • Hi Hurricanefan25. I understand your concern about abuse. See the new section on probation. Besides the fact that tools would only be given to users with sufficient experience in the area in the first place (that is one of the listed requirement) following a 7-day consensus discussion, any abuse would lead to rapid revocation, and they wouldn't be approved again any time soon. Are there other measures you'd like to be seen taken to address this issue? Dcoetzee 05:11, 29 November 2011 (UTC)
  • Support trial. I think the goal of this is basically being a proving ground for candidates to grow into becoming RfA candidates. There's two models that this kind of proposal could potentially take: the Police Community Support Officer model, where we create essentially junior admins on a semi-permanent basis, or there is the work experience model, where we give a potential adminship candidate a chance to prove him or herself, or potentially to rule themselves out from having to go through RfA as unsuitable. The only thing I think would be good would be if this could be tied into admin coaching. This way, a candidate for tool apprenticeship has an admin who is responsible for his or her use of the tools. As RfA reform goes, this is a nice, easy practical proposal that I think we actually can try and see if there's any merit to it. If it doesn't work out, the RfA reform crowd can go back to the drawing board and try again. —Tom Morris (talk) 01:17, 29 November 2011 (UTC)
    • Hi Tom, thanks for your support. I agree that coaching by admins would be great, and would strongly encourage it, but as a practical matter I don't think every apprentice would be able to find a coach/mentor, so I've avoided making this mandatory. Reviews at request time are a lightweight substitute. Dcoetzee 09:04, 29 November 2011 (UTC)
  • Support I want to see how this works out. It could work. One thing that worries me is people mass opposing anyone that runs though it because they don't like the process, irregardless of the person. That would be sad, there's little doubt in my mind that it will happen. After all, the number of opposes for reconfirmation RfAs, just because they were reconfirmations. Sven Manguard Wha? 02:54, 29 November 2011 (UTC)
  • Oppose: this will create a class of parallel pseudo-admins whose privileges and responsibilities are unclear. There may or may not be consensus that RFA is broken, but there certainly is no consensus that adminship itself needs to be greatly altered. One of the most important things about it is that the community has always had and still has a clear sense of what an admin is, who is one and who isn't, and what it means for adminship to be taken away. I don't see muddying that category as helpful. Furthermore, it is my experience as a long-term admin that the tools are inseparable from each other: for example, deletion and page protection are often part of the same process (whether protecting a title as deleted or protecting a just-undeleted page), and similarly, sorting out a complex conflict my involve both protection and blocking. The only tool I could see being separated out would be the userrights tool. Chick Bowen 03:43, 29 November 2011 (UTC)
    • Beside the point. That's for people who want to do that. There are also processes that require only one of the tools. You'd have much more time for dispute resolution if vandals and socks could be blocked by others. We are trying to generate consensus here.Jasper Deng (talk) 03:45, 29 November 2011 (UTC)
      • As I say, not in my experience. I did not realize how much the tools depended on each other until I had used them for a while. There is no way to know, if you haven't tried to address one of the various admin responsibilities, which tools you will need to do so. Chick Bowen 03:47, 29 November 2011 (UTC)
        • This would be for only one or two of the responsibilities, not all of them.Jasper Deng (talk) 03:49, 29 November 2011 (UTC)
          • Precisely, and that's why it's not a good idea. Just "one or two of the responsibilities" is not often feasible; usually, different administrative areas are linked together, e.g., someone creating hoax articles must be blocked and the articles deleted; or with a vandal, pages should be protected and/or deleted, and the user blocked. An admin isn't useful if he/she can only perform a certain task or two, and leave the rest of the job unfinished for someone else to clean up. /ƒETCHCOMMS/ 05:19, 29 November 2011 (UTC)
    • No, that's simply inefficient and unnecessary. /ƒETCHCOMMS/ 15:23, 29 November 2011 (UTC)
    • Hi Chick Bowen, thanks for your feedback. I've given some thought before to the issue that some tasks require many tools (see "What about overspecialization?" under Wikipedia:Tool_apprenticeship#Responses_to_anticipated_objections). As Jasper noted, apprentices can collaborate, or can ask admins for help through the same channels as normal users. Although this incurs communication overhead, I think the work done by the apprentices would far outweigh this (whether this is the case is one question a process trial could help answer). I've endeavored to clarify that apprentices are not administrators, and are not entitled to act as representatives of the community. Dcoetzee 06:34, 29 November 2011 (UTC)
      • Actually, I am arguing that all admin tasks require multiple tools, because the admin needs experience with the full spectrum of the tools to know how to deal with the situations that arise. In this sense, the definition of adminship is the crux of my argument with you, and it doesn't matter how clearly you define these apprentices as non-admins, because you are still asserting that admin experience is not necessary in order to carry out specific admin functions, an assertion I strongly disagree with. Chick Bowen 20:49, 29 November 2011 (UTC)
        • This is an interesting argument, but one could argue similarly that experience with checkuser/oversight rights provides valuable experience for admins that informs how they deal with situations like handling socks, revision-deleting personal information, etc. More experience with more tools does help, but is not absolutely necessary, and our existing system recognizes that - the important thing is that they recognize when a situation requires help from others, and apprentices could be evaluated based in part on this. Dcoetzee 04:24, 30 November 2011 (UTC)
  • Comment RFA is broken, because less and less users are willing to stick their neck out and participate in it, simple as that. However, tool apprenticeship is probably not the answer. The answer is the fix RFA by making it less confrontational so more users will be willing to apply for adminship the traditional way. However, fixing RFA requires a lot of editors who don't see it as broken to admit there is a problem, but some aren't willing to take that step yet despite the many flaws that are presented over and over again. — Moe ε 05:36, 29 November 2011 (UTC)
    • Hi Moe, thanks for your feedback. While I do think there's room for improvement in RfA, it faces many barriers to change. In light of that, I think tool apprenticeship will help users who are intimidated by RfA to develop confidence in their use of tools and related policy so that they can later feel confident applying, while simultaneously getting more work done on the project. Would like to know your thoughts on this. Dcoetzee 06:08, 29 November 2011 (UTC)
  • Comment Wouldn't this be an open invitation for a would-be rogue admin to gain the tools under lesser scrutiny, build up some trust and receive them permanently at RFA, and then go and block Jimbo and delete the main page (or whatever mischief he intended to do? Seems like this would make a would-be User:Wonderfool's path to adminship easier--of course they would wait until after the apprenticeship period to go rogue. 169.231.52.208 (talk) 07:43, 29 November 2011 (UTC)
    • I don't believe blocking Jimbo and deleting the Main Page are actual concerns of editors. If you block Jimbo and delete the Main Page, those can easily be fixed, and your administrator tools would be taken from you almost immediately. — Moe ε 08:16, 29 November 2011 (UTC)
    • Although experience as an apprentice would be one factor taken into account at RfA, it does not reduce a user's burden to meet other RfA requirements. Red flags pointing to potential rogue admins would carry just as much weight, and such a user would likely not be promoted to admin (or if they are, it's unlikely their evil plans would've been detected regardless). Users who go rogue during their apprenticeship would be stripped of tools almost immediately as a probation violation. Dcoetzee 08:22, 29 November 2011 (UTC)
  • Support This will make administrative tasks less bureaucratic and it will also lighten the admins' workload. PaoloNapolitano 08:23, 29 November 2011 (UTC)
  • Support a brief trial. We've got to do something differently.—S Marshall T/C 16:47, 29 November 2011 (UTC)
  • Support. Nothing ventured, nothing gained. Ironholds (talk) 17:22, 29 November 2011 (UTC)
  • Oppose per Chick Bowen. Ruslik_Zero 19:33, 29 November 2011 (UTC)
  • Strong Oppose The current system is crappy, but works. This proposal would only add an extra layer of unneeded bureaucracy, create endless drama, and serve no beneficial purpose to the encyclopedia. With much respect to the proposers, this is a terrible idea. -FASTILY (TALK) 20:02, 29 November 2011 (UTC)
    • Hi Fastily, I appreciate your feedback. I would like to emphasize that one of the main goals of the proposal is to reduce discussion (and drama) around requests as compared to RfA by placing restrictions on apprentices. The benefit to the encyclopedia is, I think, obvious - the apprentices would accomplish work (like blocking vandals, speedy deleting new pages, handling protect requests) that otherwise would either not get done or take up time that an administrator could be spending doing something else. Dcoetzee 04:34, 30 November 2011 (UTC)
  • Support. I have long argued for a process permitting routine use of the tools (blocking people who go around inserting penises into BLPs, clear cut CSDs etc) with something like RfA remaning to determine those who are deemed competent enough to make tough judgement calls (difficult blocks, contentious XfDs and so forth). I do recognise the risks and drawbacks of this, indeed vehemently disagree with the idea that someone should be allowed to make difficult blocks of close contentious XfDs based on a lightweight process. However, the fact that this appears likely to happen is a relatively small price to pay for what is otherwise a giant leap in the right direction, and I'm confident that after the trial the appropriate tweaks will be made to this one part of the solution. —WFC03:16, 30 November 2011 (UTC)
    • Hi WFC, thanks for your support. These are good points, and I definitely recognize in areas like non-admin closure that less experienced users are not always trusted with tough decisions, even where they have the technical ability to perform the action. To some extent I think probation will discourage risky actions, but to further decrease risk it might be profitable, at least for certain tasks like the ones you mention, to require that apprentices restrict their use to "obvious cases." In the long term, I think the best way to proceed here is to delegate to the various areas the responsibility of determining what they allow apprentices to do and not do. Dcoetzee 04:12, 30 November 2011 (UTC)
  • Suggest (again) I might support this proposal if it was better understood how the process would work and how it might effect Wikipedia. The changed proposed is fairly massive. It is proposing a new breed of admin. I suggested weeks ago (near the top of this page), "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.", and would still like to see that sort of trial before a trial with live ammo. I think it would be beneficial to all concerned and possibly help alleviate a great many concerns. fredgandt 03:34, 30 November 2011 (UTC)
    • Hi Fred, I apologize that I didn't respond to your proposal the first time you mentioned it. Although demonstrating that discussions can be straightforward as I claim would be invaluable, as comments by others makes clear, I was concerned that a mock trial would attract few participants and carry little weight because the participants would be aware that nothing is really on the line. There might be ways to partially correct for this e.g. by suggesting that the discussions would "carry over" into the real trial - I'm open to ideas here. Dcoetzee 04:40, 30 November 2011 (UTC)
      • Acceptable. If the trial was to run as described but that no tools were handed out. The community then gets to evaluate the process and how they feel it went, whether their concerns mattered, if users who they felt shouldn't get tools were or not granted them etc. etc. and we can all get together for another round of discussion and (cough) voting to decide finally if those who were successful in bidding for tools get them. The trial then continues to see how the process deals with the use of the tools and any misuse of them, more discussion and (not) voting.... It may seem long winded but if something is worth doing, it's worth doing properly. I feel it is unlikely anyone would find this course disagreeable. Trial > More discussion > If agreed, tools get handed out > Trial > More discussion > Who knows? > Get some sleep. fredgandt 05:01, 30 November 2011 (UTC)
      • More simply. A chance to slam the brakes on just prior to tools being handed out would sway my vote (oops I mean erm bold text comment). fredgandt
        • Comment: Everybody is able to get the full sysop rights at the testwiki, the huggle testwiki and/or testwiki.org... So if you/anyone want to play admin to learn the tools, test it there. mabdul 13:13, 30 November 2011 (UTC)
          • Excuse me? Point that finger with care. fredgandt 15:51, 30 November 2011 (UTC)
          • @Mabdul: The purpose of apprenticeship is not primarily to learn how to technically operate the tools, but to learn how to responsibly use them according to policy while also getting real work done.
          • @Fred: At the moment I still favor a live trial because discussions will be taken more seriously, but I will definitely revisit your idea if there isn't support for a live trial right now. Dcoetzee 02:43, 1 December 2011 (UTC)
  • Oppose *We do not have a shortage of admins. We have perhaps a shortage of careful admins, but I think that the degree of idiosyncrasy and carelessness among admin is considerably less than when I joined 5 years ago. We should continue the course of increasing the degree of selectivity. There are sufficient quasi-admin tasks for prospective admins already to judge the quality of candidates. DGG ( talk ) 05:11, 30 November 2011 (UTC)
    • We are losing quite a lot of active admins each year to retirement/voluntary desysopping.Jasper Deng (talk) 05:13, 30 November 2011 (UTC)
    • Hi DGG, thanks for your feedback. I want to note that I'm not advocating a less selective RfA process, and this process is in some ways more selective (it requires clear experience in the area the tool will be used in). The main difference is that administrators can perform tasks in a variety of areas, while apprentices are limited to a single area with regular reviews. In fact, the ability to create apprentices to fill out workloads might facilitate more selectivity in the choice of full administrators. Dcoetzee 06:29, 30 November 2011 (UTC)
  • Support Opposers raise valid concerns, so probably a small scale trial would be best. FeydHuxtable (talk) 15:22, 30 November 2011 (UTC)
  • Oppose No way is this an acceptable solution. If you can't gain the communities trust you shouldn'y have access to the toolset. Spartaz Humbug! 18:37, 30 November 2011 (UTC)
  • Oppose per Jayron, Kudpung, DGG. If I trust someone with some of the tools, I trust him with all the tools. If RFA is broken, this is not a good way to fix it. Many admin tasks I do require a variety of tools, and not just a small subset, and not the same ones each time. Edison (talk) 19:59, 30 November 2011 (UTC)
    • See Wikipedia:Tool_apprenticeship#trust on the trust claim. Although some tasks do require a variety of tools, many tasks don't, and those are the tasks apprentices would contribute on. I also would not characterize this mainly as an RfA reform measure, but as a supplementary process that serves other purposes like getting users experience with using tools according to policy, and getting more work done on the project. Dcoetzee 02:31, 1 December 2011 (UTC)
  • Oppose, per virtually all of the above opposers. My views on RfA reform have been very well advertised for quite a long time, and I'm a coordinator of a reform project myself. So as much as it pains me to oppose an attempt at change, this is simply not the type of reform that I feel is needed. Swarm X 22:38, 30 November 2011 (UTC)
  • Comment there seems to be a lot of 'oppose, because, my idea is much better'. Well, hell, people have had better ideas for the past several years; none have made any difference whatsoever. We all have ideas about what is "best" - but to just oppose because 'this doesn't QUITE match my own ideal' isn't helping. So, can the opposers please try to indicate what would work, not just what wouldn't - I mean this in the best possible faith.  Chzz  ►  23:16, 30 November 2011 (UTC)
  • Support. Clear start, clear end, clear parameters during the Trial and clear parameters for running the Trial. --Shearonink (talk) 23:30, 30 November 2011 (UTC)
  • Support the trial. I see no terribly concerning issues with at least trying a new system for user tools. Tyrol5 [Talk] 23:33, 30 November 2011 (UTC)
  • Oppose - Not really much to add; if I trust someone with a few of the tools, I trust them with them all. Going through a process to determine trust in them by the community as a whole seems preferable to me. ~~ Hi878 (Come shout at me!) 02:08, 1 December 2011 (UTC)
  • Support one week tool use instalments for 3 month trial Dcoetzee has bent over backward to work this proposal out to suit our opinions. Not only for that reason, I am choosing to support this but for that ethos. The proposal is to try something new (and daring) and see what comes of it. It is being so well considered I trust it to be run well. One important point though: Just because the tools will only be handed out for use in one week instalments, this should not mean the requests should be treated less officiously. I very rarely agree with bureaucracy, but where weaponry is concerned, I think more is better. I'd be interested now to know what admin are on-side to watch over the wannabes trainees. fredgandt 02:50, 1 December 2011 (UTC)
    • Hi Fred, I appreciate your support, I know this wasn't an easy decision. I agree that requests even for a short period should be taken seriously. Monitoring of apprentices during the trial would be relatively easy since there are only 10 at a time - I'm definitely going to contribute, and I would make sure before the trial begins that we line up a sufficient number. Dcoetzee 03:08, 1 December 2011 (UTC)
  • support trial per Chzz above; many of the opposes seem to be "change is needed, but not this change." At the very least, what this will lead to is a "no, we actually gave that a go, not just armchair speculation, but it didn't work out and this is what we've learnt" at WP:PEREN and a way forward for the future. Honestly speaking, I was never very keen on the idea of unbundling (y'know, the whole "well why can't they just be trusted with the whole gamut), but I can see how it would benefit some (why else would it be so frequently proposed?) and am willing to give it a go. I guess part of it is that I see now that the concept is not so much "we would rather you not have those tools, only this one", as "would you prefer to only have what's relevant"-- not that big an ask, and it removes the Very Big Weight from the shoulders of our (declining in proportion to activity on this project) active admin corps, and from the bit itself. Still cynical, but nothing ventured, nothing gained. sonia05:57, 1 December 2011 (UTC)
    • I see it as a solution looking for the wrong problem(s). It's still not clear whether it is intended to address the diminishing pool of admins, the diminishing number of candidates, or the unpleasantness and flippant and unresearched voting that has become commonplace at RfA. In my opinion it does neither, and is simply a noble attempt to implement a stand-alone experiment for the perennial issue of unbundling of the tools. We should really be looking at ways to improve the climate of the existing system and encourage participants to vote objectively and with civility. Do that, and we'll have all the admins we need, and who can be trusted with all the tools. Kudpung กุดผึ้ง (talk) 05:31, 2 December 2011 (UTC)
      • I don't necessarily disagree. However, one could argue that the article feedback tool etc are likewise a solution looking for a problem. However, it is clear that the one thing Wikipedia doesn't need is stagnancy. That's what the AFT is for, at least to be a "release" for reader dissatisfaction at the quality of what is becoming more widely used by the day, and provide a little constructive criticism for the editors. That's what the new NPP system is for, to radically change the way patrollers approach their task, and see if this improves conditions in that role. That's what the GEP is for, to bring Wikipedia if not credibility as a resource, at least a reputation as a modern teaching aid. That's what the Athena skin is for, to rebrand the project for an increasingly large share of our audience. The one major flaw of the wiki model is that, while it does scale to a project this large, new things that aren't led from above don't get tried very often because consensus is hard to come by. The Foundation isn't going to reform RfA unilaterally. Thus while I'm cynical about this particular idea, I think that any steps the community can take toward not just going around in circles disagreeing with itself or factionalizing into groups whose actions actively disrupt each other's goals are absolutely fantastic. For me, this proposal will provide closure to an oft-floated idea if it fails, and open the way for more dialogue if it succeeds. Either is good for progress of some sort, and that works for me. (It's also far, far more achievable than the vaguely pretty goal of "let's make people nicer to each other", however much I'd love that to be the case.) sonia09:48, 2 December 2011 (UTC)
        • All proposals for change should have a clearly defined objective. This one doesn't - it's simply an experiment for the sake of experiment and doesn't address any specific issues. All it offers is a fast-track to adminship for wannabes who can't /won't get the tools the regular way and is a veiled attempt to reduce the standards that are about right. There are plenty of people who meet those standards, and the only way to get them to come forward is to impress upon the voters who only go to RfA to be vindictive and get away with it, to either clean up their act or stay away. And this proposal is not designed to address that, so what is it for? As someone else inferred here, any mistakes the sorcerer's apprentices make, will be written down and taken in evidence - without the accompaniment by Dukas or the help of Goethe's gentle magician. Kudpung กุดผึ้ง (talk) 12:45, 2 December 2011 (UTC)
  • Oppose WP:CREEP, and a solution without a problem. So if you say RFA is broken, then fix that; don't try to go around the problem. --Rschen7754 06:53, 1 December 2011 (UTC)
  • The problem with trying to reform RfA comes in many parts, one of which is the number of people who say "RfA isn't broken", one of which is the small group of people referring to our attempts to clean up RfA by tagging us as "RfA Deform", one of which is the fact that a loud-voiced group of people strenuously object to any attempt at enforcement (and frequently attack people who attempt even a reminder) of civility and NPA at RfA, and so on. The other point, again, is that there are probably quite a few people (including myself) who don't want the full adminship hat and will therefore never ask for it, but who could make good use of a small subset of tools. (Like delete/undelete). Pesky (talkstalk!) 10:20, 1 December 2011 (UTC)
  • Hi Rschen7754, thanks for your feedback. As noted in my support above, the main problems intended to be solved by this proposal are: 1. helping to complete a lot of work around the project that requires admin tools, like CAT:CSD, vandal patrol, protection requests, etc.; 2. allowing users to gain hands-on experience with responsible use of tools according to policy with continual feedback, permitting a clearer assessment at any later RfA and increasing their confidence to run. Dcoetzee 20:05, 1 December 2011 (UTC)
  • So the solution to the backlogs is... adding another backlog? This sounds like a lot of overhead. Also, we allow users to gain a lot of hands-on experience already; RC patrol is not restricted to admins, for example. Sending an AIV report illustrates the judgment that is necessary to determine if a user should be blocked. --Rschen7754 21:53, 3 December 2011 (UTC)
  • Question: One part of me if quite glad to see that there is still some community vetting in the form of a 7 day "mini-rfa" if i may call it that, rather then just shoving this under WP:PERM where just one editor decides. However, one of the main drawbacks of a regular RFA (In my point of view) is that the requirements to be promoted tend to increase more and more. Is there anything that prevents this from happening here as well? As some editors mention: "If i trust them with one tool i trust them with all". However, if i read that the other way around it would state "If i don't trust them with all tools, i also won't trust them with one". What ensures that the current RFA standards won't simply flow over this light version of that process? Excirial (Contact me,Contribs) 11:22, 1 December 2011 (UTC)
    • Hi Excirial. In addition to the various restrictions on apprentices, such as short trial length and probation, there will be effective moderation of discussion including removal of discussion irrelevant to the tool/task area under request. Encouraging conditional supports in place of opposes will let concerned users set their own terms. Also, because it's a consensus discussion and not a vote, the closing bureaucrat can exercise discretion and assign less weight to less reasonable concerns. Dcoetzee 19:42, 1 December 2011 (UTC)
      • To be honest i doubt it will be that simple seeing some of the RFA's. Still, i will give this a Support as a trial only provided that implementing this change doesn't require a large share of development time to implement. I can see quite a few caveats and i do wonder if this will work out in the long run, but a trial won't hurt, and by now it might actually be a lot better to start working on a trial-and-error basis rather then discussing what should be done another few years without result. Excirial (Contact me,Contribs) 20:32, 1 December 2011 (UTC)
        • Thanks for your support! The dev time to implement the basic groups is literally seconds. Automatic "timeout" of rights would require more effort, but that can wait since bureaucrats can implement it manually in the meantime. Dcoetzee 20:47, 1 December 2011 (UTC)
  • Oppose for various reasons already outlined, including bureaucratic overhead, propensity for drama etc., but above all because the notion of an applicant (for want of a better word) having an "immediate need" as set out in the justification for this proposal. How can someone have an immediate need but have to wait for a process to determine whether they are suitable to be given delete or block rights? It makes no logical sense. The "immediate need" would have to be scrutinised, why then an Admin. not just deal with it immediately? Leaky Caldron 18:29, 1 December 2011 (UTC)
    • Hi Leaky caldron, thanks for your feedback. It's clear that my intention wasn't accurately conveyed here. This is not intended for one-time cases, or for urgent situations, but for users who have an ongoing need for a tool and can start using it right away after receiving it (see the Example section for a typical example). I've revised the proposal to clarify this. Please let me know if this makes more sense to you. Dcoetzee 19:50, 1 December 2011 (UTC)
      • Someone having an "immediate need" for some of the tools smacks of their wanting them to "win" some content dispute or edit war. Edison (talk) 20:53, 1 December 2011 (UTC)
        • I think "Immediate need" should be read as "Would have an immediate use for that tool in their current line of work". Think about long term new page patrollers who might do well with the ability to delete pages themselves, or vandalism patrols who are capable of deciding if a user should be blocked. I presume the immediate part mostly means that this isn't some form of permission that you can request but never really use, such as people who only patrol a handful of new pages once a month or so. Excirial (Contact me,Contribs) 21:07, 1 December 2011 (UTC)
          • Yes, Excirial has the intended interpretation. I've rewritten that point to say "can start using it right away" rather than "immediate need" try to make this clearer. I don't want people requesting tools based on speculation about things they might use it for in the future, but only for things they're definitely going to use it for right after they get it. I'm open to suggestions on how to make this clearer. Dcoetzee 22:17, 1 December 2011 (UTC)
  • Support seems sensible - and along the lines of granting bot access without the ordeal of 'placing candles under the feet' that is RFA. Just keep the block button out of the tools set and we'll all breathe easy. --Joopercoopers (talk) 01:32, 2 December 2011 (UTC)
  • Oppose We all know full well that these apprenticeships (including on a trial basis) will be gamed by editors who blast away making totally uncontroversial blocks, deletions and protections until they get the big prize. Of course, that's not to say RfA isn't gameable as well, but we shouldn't make the process of becoming an admin even more gameable than it already is. And also because any marginal benefit brought about by this proposal, even on a trial basis, is outweighed by the expense of community time in administering yet another layer of bureaucracy. --Mkativerata (talk) 07:10, 2 December 2011 (UTC)
    • No, users will have to be evaluated before getting an apprenticeship. Please read my comment above about bureaucracy.Jasper Deng (talk) 07:14, 2 December 2011 (UTC)
      • If you're going to spam every comment on this page, at least understand what is being said. I don't care that users will be evaluated before getting an apprenticeship, except to say that it's more red tape. My point is the apprenticeship will be meaningless as any sensibly cynical editor will game it to smooth the passage through their future RfA. And yes, I did see your intervention regarding bureaucracy above, and I am most unconvinced, thank you. --Mkativerata (talk) 07:18, 2 December 2011 (UTC)
        • RfA is RfA, not this. Our purpose is to make RfA less stressful. Apprenticeship!=free pass at RfA, because commenters at RfA will still scrutinize the actions to determine readiness for all tools. This is only training, not admin coaching. When we train someone to drive, the test is made easier but whether the test is passed or not is up to the assesser.Jasper Deng (talk) 07:23, 2 December 2011 (UTC)
          • Your first sentence continues to show you haven't read what I said. Your second and third sentences contradict each other. Your fourth sentence is irrelevant. Your fifth sentence is an absurd analogy. Why can't people come here, make a comment, and not have to keep coming back when their comments get misconstrued or just plain spammed? --Mkativerata (talk) 07:28, 2 December 2011 (UTC)
            • I perfectly understand your comment: gaming the system and bureaucracy will increase. Scrutinizing still will be done for apprentices, but it will be more likely to pass based on work as an apprentice, with less criticism of things like "mislabeled this CSD." If you don't want me to reply to your comments, don't make them. (and while you're at it, please don't make comments on my comments ("...is an absurd analogy."))Jasper Deng (talk) 07:32, 2 December 2011 (UTC)
              • There will still be RfA which is always partly a (un)popularity contest. This system will not deprive the wolf-pack of their favourite hunting ground when applications for the tools reach normal RfA stage. And if you believe it will, I'd like to hear how. Kudpung กุดผึ้ง (talk) 07:42, 2 December 2011 (UTC)
                • In fact if you believe in the wolf-pack theory (lets pretend for a moment that I do), this would only make things worse. Any errors during the trial would be blown out of proportion. And if the trialist did well, the wolves could say "you're just a badge collector who used the tools uncontroversially for a week or so to boost your RfA prospects". --Mkativerata (talk) 07:49, 2 December 2011 (UTC)
    • @Mkativerata: If I'm understanding your objection, the idea seems to be that voters at RfA will assign too much weight to success and failure at apprenticeship, allowing the unready to pass and unfairly rejecting those who make a few small errors while learning. While this is possible, at least at first, I think it far more likely that the voters at RfA will assign appropriate weight to apprenticeship experience in their decision, well aware of these possibilities (and I would be the first to chastise voters who do otherwise). Your argument would apply to anything a user does that could increase their chances of passing RfA, such as working on featured articles, mentoring newbies, tagging articles for speedy deletion, discussing policy etc. We haven't banned these activities for at least two reasons: 1. RfA voters weigh this experience appropriately; 2. it's important work to get done. The same applies here. Dcoetzee 01:49, 3 December 2011 (UTC)
  • Oppose Last I checked, there were 1500+ admins; I've never found a case where action was needed where one wasn't available. The key element of administrators is judgement, not technical prowess in a particular area. Either an editor can be trusted not to use an admin capability in an area they weren't sufficiently familiar with to use wisely, or they're not ready to be an admin. Have we all forgotten WP:NOBIGDEAL? Gerardw (talk) 03:43, 2 December 2011 (UTC)
    • Hi Gerardw, thanks for your feedback. You're right that merely finding an admin is not hard - but at any instant there is a lot of administrative work waiting to be done around the project, and getting more of it done faster is a good thing. Re "technical prowess": this isn't about learning how to use a tool technically (which could be learned on Test Wiki) but about learning to use tools in compliance with policy, which is much more difficult. Re your final comment: while we do trust the judgement of admins to be careful, admins are not in fact given access to all tools (checkuser, oversight, bureaucrat), nor are they guaranteed to pass a request for these tools. Likewise, some apprentices who can capably use a single tool might not be prepared to pass a full RfA (or even if they are, they might feel more comfortable developing confidence with compliant use of the tools first). Dcoetzee 02:52, 3 December 2011 (UTC)
      • You are not a checkuser or oversighter because the community does not trust you with the tools. You're not a checkuser or oversighter because you haven't asked for the additional tools. You'll find out if the community 'trusts' you with them if and when you apply for them, and not before. Perhaps your proposal should be the paradigm for admins to have checkuser and oversighter on a trial basis. Kudpung กุดผึ้ง (talk) 15:15, 3 December 2011 (UTC)
        • I did not claim that the community doesn't trust them with these tools. I merely said that they are not given access to these tools, and are not guaranteed to pass a request for these tools, which is true. Dcoetzee 23:50, 3 December 2011 (UTC)
    • Hi Gerardw, yes there are 1500 or so admins, but less than 750 are in the slightest bit active, and that's a big drop from the peak of 1021 active admins in early 2009. We are likely to appoint about fifty this year, which is two thirds of the number we appointed last year which was in itself only two thirds of the year before. So when we consider potential reforms like this we need to think not just whether we need this reform with today's number of admins, but can we get this or some other reform agreed, tested and working before the number of active admins falls to levels that damage the wiki. ϢereSpielChequers 01:08, 8 December 2011 (UTC)
  • Support trial for now. This should narrow the permissions gap. Marcus Qwertyus 13:33, 3 December 2011 (UTC)
  • Strong oppose. Getting the sysop flag is not and should not be a big deal. We don't need 172 permissions. If the community trusts the user, he'll pass an RfA. If he wouldn't pass an RfA, we have no business devising a scheme for him to get the rights anyway. If the candidate's good and should have the tools, he should have all of them, thru a proper RfA. If not, he has no business having some. Snowolf How can I help? 21:26, 3 December 2011 (UTC)
    • Thanks for your feedback, Snowolf. Please see Wikipedia:Tool_apprenticeship#trust. Dcoetzee 23:50, 3 December 2011 (UTC)
      • I agree with that section but it doesn't address the issue of, If we don't trust a user enough to give them all the tools, why should we trust them with one tool when it's inefficient to do so? For example, consider a user who creates a page that is a copyvio and under a title that has already been deleted before. In this case, the user needs to be blocked, the title needs to be deleted, and it also should be salted. That's three different tools: blocking, deletion, and protection. Why are we giving someone the delete button and then making him/her have to find someone to push the protect button, and another person to hit the block button? Why can't we just stick with the simple and easy system of "you get all three buttons or you get none"? Furthermore, the comparisons in the "trust" section linked above are not applicable to administratorship; rollback exists because there is such a high level of vandalism and also because it is not difficult (regardless, I think our standards for that are too low because I constantly see users misusing it and driving away newbies!). Bureaucrats can hand out (and take away) admin rights, so it would be unreasonable to expect that anyone capable of adminship is capable of bureaucratship—yet no one is suggesting we divide up the crat abilities! /ƒETCHCOMMS/ 00:12, 4 December 2011 (UTC)
        • As for bureaucrats, we don't have a shortage of them, so we have no need for division of their tools.Jasper Deng (talk) 00:20, 4 December 2011 (UTC)
        • @Fetchcomms: Wikipedia:Tool_apprenticeship#overspecialization. Some tasks do require multiple tools to deal with, but apprentices would be focusing on tasks that don't (which are plentiful), and admins would still be around to work on those that do. I've removed bureaucrats from the response, since checkuser and oversight are better examples anyway - the point is these are tools that for one reason or another we don't give to all administrators automatically. Nevertheless administrators remain productive in tasks that don't require access to these tools, and by analogy apprentices can also be productive with their limited tool set. Dcoetzee 00:31, 4 December 2011 (UTC)
      • I've read the section. I don't agree with your reasoning however. It's a fallacy to compare rollback, a right that was granted years after everybody and their mother had thru tools such as Vandalproof, lupin's popups and various js that offered limited rollbacks, with the rights you propose to give to other users. Rollback was never admin-exclusive, the technical ability to do so thru mediawiki itself was. To me, if a user doesn't pass an RfA for any reason, he should be blocking users or protecting pages. Sysop is no more of a big deal then any of the single user rights. It's no big deal, let's stop treating it like it is. If a user fails an RfA, he has no business getting the tools, if he wants to block user, no harm in a RfA. I fail to understand what's wrong with RfA that needs to be fixed in this way... Snowolf How can I help? 05:41, 4 December 2011 (UTC)
        • The way RfA works makes adminship quite a big deal.Jasper Deng (talk) 05:57, 4 December 2011 (UTC)
          • Adminship is no big deal, as any sysop knows. Most of the time it's even more trouble than it's worth going through RfA for. Those troubles will be encountered by anyone who is given the tools through whatever process. Getting through RfA is the only big deal - like getting through boot camp with a bullying Sergent Major. Nevertheless, it is human internet nature for many people to regard any level of moderatorship on any run-of-the-mill web forum as a big deal - again, only until they actually get the tools. — Preceding unsigned comment added by Kudpung (talkcontribs) 06:57, 4 December 2011 (UTC)
            • Adminship is being regarded as a big deal, which is what RfA reflects. Wikipedia, at least the English-language project, is not a "run-of-the-mill web forum", and adminship has greater significance than moderating a forum. (I should know; I've modded a few in my time.) That Wikipedia is so widely used that hoaxes started here can perpetuate throughout the web means that we have increasingly started to take ourselves, as curators and creators of content, and thus the admin position, more seriously. This proposal limits the damage that an individual who has not stood for the full gamut can do, by restricting the number of buttons that can be pushed (no inadvertent editing of the MediaWiki space or viewing of compromising deleted content, for example), and also through removing the virtually "for life" nature of adminship in favour of a speedy removal process, both of which should relieve some of the community-created pressure. In exchange we get more efficiency since the gnomes who maintain an area or two are able to do more independently without the need to go off and create content or report vandalism or whatever is currently needed at RfA. sonia09:07, 4 December 2011 (UTC)
              • I'm so glad you agree with me on that Sonia. of course it's a big deal - for all those who want it for the wrong reasons, and they need to understand that not only is Wikipedia not a run-of-the -mill forum, it's nether a forum, nor the socal media site some editors try to turn it into.FWIW, I 'moderate' 27 forums that are hosted on my server, because my clients who own them can't figure out the maze of granular permissions that come with the phpBB admin control panel, and they are swamped with too many people wanting to be moderators. So where we do disagree, is that is exactly the situation we want to avoid here: having a whole priethood of policemen (and women) where nobody will know who has what rights and what they can do with them. As I pointed out lower down, in a country like the UK where we have no gun culture, if you were to give every citizen the right to bear a weapon, there would be enormous problems, some of them don't know which end the bullet comes out of, not to mention they would be shooting the wrong people, even by accident. No, we need policemen who are policemen, and no half measures. Although it's true that a hoax or an attack page soemtimes sits for an hour before it gets noticed, it's rare, and this is more often than not the fault of the patroller who has tagged them with A7, A1, or A3. There are rarely 'any' serious backlogs at new page level - most backlogs are at XfD closures and files. Kudpung กุดผึ้ง (talk)
  • Support, per User:OlEnglish: "if it turns out to be a kind of paradigm shift in Wikipedia administration then great, if it crashes and burns, then that's great too." Change is good. --Conti| 10:31, 4 December 2011 (UTC)
  • Weak oppose - As well intentioned as this proposal is, I suspect that introducing a two-tiered system of trustworthy users will, with time, only evolve into another barrier which can only even further reduce the pool of candidates and thus the relative number of successful applicants at RfA. Further, if a user is truly (1) in good standing with the community, (2) has an immediate and ongoing need for administrative tools, and (3) is active and has sufficient experience in a relevant area, then they should generally be justified in requesting full adminiship. If we cannot adequately facilitate such legitimate requests, then the underlying problems need to be confronted and resolved. On a side note, I find the negative characterization of arguments based on generalized trust somewhat amiss; this assessment is itself fallacious, firstly because it assumes that we are solely debating the merits of granular permission control, when in reality the proposal also purports to devise an administrative apprenticeship model that is arguably incongruous with its parent process, and secondly because the banker-brother analogy's bipolar examples also fail to consider that there are commonly domains of trust within a respective role. Financial institutions aren't trusted with children, but they're generally presumed trustworthy with all things finance: checking, savings, investments, lines of credit, safety deposits, counsel, retirement planning, and last testaments. The comparison of two extraneous extremes is not an accurate depiction of what is being discussed.   — C M B J   22:01, 4 December 2011 (UTC)
    • Hi CMBJ, thanks for your feedback. I understand your concern - the system isn't intended to serve as a prerequisite to RfA, and I would oppose the institution of such a prerequisite, so I would hope the RfA candidate pool would not be diminished. On your second point, although I think some would agree that people meeting the apprentice criteria should be passing RfA, this isn't the state of things, and others would disagree for a variety of reasons (e.g. that admins should be familiar with policy broadly; that admins should handle disputes gracefully; that admins should have substantial experience in many areas including content contribution; and so on). This division makes the concept of reform in this area no small matter. You're right that my analogies in favor of granular permissions suffer from certain limitations, but the main thrust is that a user who is experienced in a limited area but unable to pass RfA, for whatever reason, might still reasonably be granted one or two tools (subject to restrictions) in order to accomplish more in their area. Dcoetzee 05:18, 6 December 2011 (UTC)
      • Even if you would personally oppose the institution of apprenticeship as a prerequisite to RfA, that does not mean other contributors will share in that view; particularly as time passes. The way I see it, creating an optional means by which administrator candidates may prove themselves cannot logically benefit those who decline to do so; however, it can and will be used as ammunition against them by detractors. Consequential effects could also be rather chilling; it is not a stretch to imagine that observing habitual attacks on the premise of bypassing apprenticeship against others at RfA would deter many would-be administrator applicants, or that the pool of viable and willing applicants would decline because of complacency with apprenticeship, or that consensus would move in the direction of impelling applicants to honorably withdraw and instead pursue apprenticeship in borderline or controversial administrator nominations, or that apprenticeship would eventually be perceived by some as—or systematically develop into—a carrot on a stick trap. The only foreseeable way around these sorts of complications would be to sunder the functional relationship between the two processes, to the extent that they are so disparate as to be totally disassociated from one another, which may or may not be a practical goal. On the second point, a literal interpretation of the apprentice criteria, as it currently reads, still does not allow for sidestepping of the same general expectations. Few active contributors can be considered in good standing if they're unable to handle disputes gracefully, and it's not as if anyone with sufficient experience in the areas of blocking users or protecting pages can be designated as such without a broad grasp of policy; in order to diligently make and execute good decisions in these areas, a contributor must be able to perform contextual discernments based upon numerous discrete policies. This leads us back to the notion that there are domains of trust, and more importantly, the question of why someone experienced and reputable enough to expunge users cannot also be trusted to moderate them, for example.   — C M B J   09:07, 6 December 2011 (UTC)
        • I agree with the first sentence of your initial post CMBJ, but ask whether that would be such a bad thing? My analysis is that Wikipedia would benefit from recruiting more janitors, but better judges. You get more janitors by lowering the hurdles, better judges by improving scrutiny. This trial has completely missed the point in that respect. But it will demonstrate to those who may or may not self-identify as sheep that small-scale experiments will not destroy the whole project. —WFC18:38, 6 December 2011 (UTC)
        • Hi CMBJ, based on your concern about this process becoming an RfA requirement, I've added a counterargument at Wikipedia:Tool_apprenticeship#RfArequirement. Dcoetzee 03:33, 11 December 2011 (UTC)
  • Strong Oppose I agree with the idea that you need more then one tool to do a seemingly simple task. CSD work often used the protect, move without redirect, delete, and block buttons. If I can trust a person with one, I should be able to trust you with the whole package. --Guerillero | My Talk 22:34, 4 December 2011 (UTC)
    • Thanks for your feedback, Guerillero. While it is true that a CSD worker is more efficient with access to multiple tools, I see it like this: an apprentice working in CSD, before being given the deletion tool, must request admin assistance for all of these actions, whereas after being given the deletion tool they only have to request assistance for a small portion of these actions, so it's a clear net gain to overall efficiency. It may also be reasonable for them to continue to request assistance for e.g. blocks, since they might be very familiar with WP:CSD, but not be familiar enough with blocking policy to safely use that tool unaided. On your last point see Wikipedia:Tool_apprenticeship#trust. Dcoetzee 05:34, 6 December 2011 (UTC)
  • Support – For the sake of gathering data and to put some issues to rest. People will be asking, "What if…?" until a trial that could help answer that question occurs. We should base arguments for or against the "Tool apprenticeship" proposal based on past experience and evidence. This trial should create that experience and evidence. Without a trial and the resulting data, arguments for or against the "Tool apprenticeship' proposal will only be based on hopes and fears. Let this trial create something more tangible than hopes and fears. --Michaeldsuarez (talk) 00:27, 5 December 2011 (UTC)
  • Very Strong Support — Speaking only for myself, even though I'm not sure if I'd be eligible for it, I would love to have access to the delete button so I could be of greater assistance at AfD and NPP. However, I'm not so sure I'd be interested in receiving the whole admin toolset, and I'm too scared of what would happen if I were to submit an RfA (even if I did have more than 3000 edits, which I don't). Aside from my own preferences, unbundling the administrative tools could help to eliminate the hierarchal structure that Wikipedia has unfortunately developed over the years. I'd even support ending the actual technical position of administrator while allowing the people who have already been granted the mop to retain all priviledges that are associated with the role. Master&Expert (Talk) 04:07, 6 December 2011 (UTC)
  • Support, interesting idea. --Teukros (talk) 19:04, 6 December 2011 (UTC)

Arbitrary break for convenience of editing

  • Neutral per fetchcomms. One thing that concerns me is that Jasper and others say that the viewdeleted right would not be one of the ones handed out. I can sorta see the reasoning behind this, but it makes it very hard for NPP's like me who, not through making mistakes, but because of newbs who want to know what they did wrong, need to be able to see articles we've tagged for CSD. I agree with fetchcomm's other comments about the inefficiency of needing three different users to, say, block a problematic editor, delete an article he's created and protect the page, but I won't oppose on that alone because hey, this is just a trial. Nolelover Talk·Contribs 19:21, 6 December 2011 (UTC)
    I believe that 'viewdeleted' has been left out of this trial deliberately, for a few reasons; it is not something that can be assessed or checked on, in any way - no log entry is created when people simply look at deleted pages. Also, historically, there has been concern (including from WMF) about allowing any but a few specially-selected users (viz. admins) to be able to view deleted content. Remember, this proposal is for a one-off, short-term trial, and consequently kept as simple as possible. After evaluating the results, future possibilities for including other user-rights are certainly an option. The idea here though is, to "keep it simple" and get some trial that can be accepted, which could help evaluate /propose how some form of system could possibly operate over a longer term.  Chzz  ►  20:07, 6 December 2011 (UTC)
Don't get me wrong; I do understand why the viewdeleted right won't be used (and that isn't the reason I went from support to neutral anyway), but from the POV of a CSD'er who would only be interested in the delete/undelete trial, it becomes almost useless/more of a hassle then its worth. I should add to the author of this page that nowhere on the proposal are the actual tools specified. It might be a good idea to spell out exactly which ones are gonna be used in the trial. Nolelover Talk·Contribs 21:12, 6 December 2011 (UTC)
A limited viewdeleted were you could view stuff that you deleted yourself would be a good compromise. -- WOSlinker (talk) 21:44, 6 December 2011 (UTC)
That would be great, if technically possible? Nolelover Talk·Contribs 21:46, 6 December 2011 (UTC)
This is possible but would require changes to the Mediawiki software. I think most of the time the use of a good deletion summary can avoid the need to see the deleted text, but you do have the option to either retain local copies of articles you delete or get assistance from an admin (although I realise these are awkward). Re: the tools available during the trial, see Wikipedia:Tool_apprenticeship#Proposed_process_trial (bullet 3). Dcoetzee 04:09, 11 December 2011 (UTC)
  • Support a trial. I think it would be basically harmless, and if it works well we may be able to torpedo some backlogs in the future. Mark Arsten (talk) 19:45, 6 December 2011 (UTC)
  • Support This has potential to benefit many areas of the Wiki, eliminating many serious backlogs. I believe it will give users experience in true administrative duties, and will offer constructive feedback to them and there experiences with said tools. Tarheel95 (Sprechen) 20:03, 6 December 2011 (UTC)
  • Question It says after the initial week of usage, the tool can be requested to be extended. It does not say how long it can be extended for. Is it just another week or is it a longer period? Will it be possible to keep on requesting extensions? This could make passing a RfA more difficult if it's possible to keep on requesting extensions as when someone got through the RfA process, they wouldn't be able to say that they needed some of the tools anymore as they would be able to get them through this other process instead and you could get people opposing saying that they already have the tools they need. -- WOSlinker (talk) 20:09, 6 December 2011 (UTC)
    • I think there should be some sort of limit on the extensions. Say 4 weeks at most. Then after that you would then need to apply for adminship or wait at least another 4 weeks before requesting the tool again on a temporary basis. -- WOSlinker (talk) 20:12, 6 December 2011 (UTC)
      • Hi WOSlinker, I've tried to clarify this. Extensions don't make much sense under the new terms of the trial (since discussions and trials are both 7 days), so I've changed this to simply say users can request a new trial (also for one week) after the end of their trial. Because discussions are 7 days, they won't have access to the tool more than half the time. Would this mitigate your concerns? Repeat trials are important as their requests are the main opportunity for feedback. Thanks! Dcoetzee 08:57, 7 December 2011 (UTC)
        • Yes, that seems ok for the trial. Since the the number of people in the trail is limited, it's good that no-one can keep one of the tools for the whole time as it allows more people to have a go, so Support from me. -- WOSlinker (talk) 10:08, 7 December 2011 (UTC)
  • Strong unconditional support - We NEED something like this to allow experienced users to reduce backlogs URGENTLY. Barts1a | Talk to me | Yell at me 22:42, 6 December 2011 (UTC)
  • Support I see absolutely no harm from a trial, and plenty of potential good — even if it's only to learn that it doesn't work (though I'm confident that something more positive will come of this). We need to be more bold and confident, and far less fearful, about proposals like this that are so well thought out. First Light (talk) 03:41, 7 December 2011 (UTC)
  • Support. Good idea. Once implemented, make sure that you document the process of how it works and start a "best practices" page to document lessons learned. Cla68 (talk) 05:15, 7 December 2011 (UTC)
    • Thanks for your support, Cla68. I'd like to get some clarification on your proposed "best practices" page and the kind of information it would contain. Are there analogous pages I can look at for other processes? Dcoetzee 03:18, 11 December 2011 (UTC)
      • I have yet to see anyone start a "best practices" or "lessons learned" page regarding any of WP's administrative processes. What it is is a page which documents what works best and what doesn't so that the things that work best become established habits and the things that don't work well aren't repeated. For example, if you notice that editors seem to do best when given a certain tool first, then write this down on the best practices page. The page doesn't need to be a certain format. It will organize itself as time goes by. Cla68 (talk) 07:35, 11 December 2011 (UTC)
        • Ah I understand now. The report to be completed after the process trial is primarily intended to serve this purpose, and the draft of the report can be started during the trial to document lessons as they happen. I also think it'd be great to develop a page to help advise good and bad arguments for tool request discussions, akin to WP:ATA. Dcoetzee 12:30, 12 December 2011 (UTC)
  • Support - worth a try. After all the talk over many years about unbundling the admin tools, it's high time we actually tried doing it to see what happens, and I'm encouraged to see the degree of support this proposal has received in contrast to previous ones. My only concern is the one raised by CMBJ above: that in the long term, this process may actually make it harder to pass RFA, as voters may say 'Oppose, user should prove his competence with all the tools at Tool Apprenticeship before being given them on a permanent basis'. But on the other hand, if it means users gain access to the tools who wouldn't have got through RFA, it would be more good than bad. It's hard to say what the overall impact on adminship would be, but that's why it's worth a trial. Robofish (talk) 12:39, 7 December 2011 (UTC)
  • Support - Provided when it is first asigned, editors are watched closely and have to file a report about evry time they use the tool for the first week. Oddbodz (talk) 17:36, 7 December 2011 (UTC)
Oddbodz, all the administrative actions included in the trial are logged, automatically. They are logged under the type of action, by user. For example (and with permission from that admin to use them as an example), here are deletions performed by GorillaWarfare (talk · contribs).
I am absolutely sure that lots of people will closely watch what happens with this trial. Furthermore, the proposal says that, serious misuse can lead to immediate tool revocation,ref which is above and beyond our regular procedures for dealing with problems. Chzz  ►  23:36, 7 December 2011 (UTC)
  • Support a trial. This will give us useful data about whether such a system could work. — Mr. Stradivarius 06:07, 8 December 2011 (UTC)
  • Support This would be helpful to prospective admins.--Hallows AG (talk) 07:41, 8 December 2011 (UTC)
  • Oppose per Jasper Deng's clamouring for it and per the many strong statements above. I've no doubt Dcoetzee means well, but this is dangerous. There are already too many inappropriate users with sysop-privileges; this would only make matters worse. I've seen the arguments that the request process is broken, and it is, as are the stillborn efforts at reform. The problem isn't the current RFA, it's the current community. Alarbus (talk) 11:05, 8 December 2011 (UTC)
    Hi Alarbus, thanks for your feedback. I would like to emphasize that I don't support all of Jasper's statements, and that apprentices (unlike the administrators you allude to) would be much more tightly controlled; their trials are time-limited and tools can be immediately revoked if misused. During the process trial, all apprentices will even be continuously monitored. Does this address your concerns or are you afraid the risk is still too high? Thanks. Dcoetzee 02:15, 9 December 2011 (UTC)
    Hello. I've read most of this and the related pages, and I do see that there would be short leashes. But I still see this as a camel's nose issue. Too many would leverage it to slip through a later RFA having faked it for a week or a month. Fix RFA by fixing the community. Alarbus (talk) 02:46, 9 December 2011 (UTC)
  • Weak support. Personally I don't think tool apprenticeship is a good idea, but I'm willing to give it a chance to prove itself. If it turns out to be a disaster, well, it will show, but if it works that would be fantastic. It must be made very clear that the trial must end and not be renewed indefinitely like the PC trial (nearly) was. (Even though I supported PC, I agree that prolonging the trial would be an abuse of process.) -- King of 01:10, 9 December 2011 (UTC)

Second arbitrary break

  • Support a trial. No reason not to, really. An informed discussion about a process cannot be had without knowing how the process will work in practice. T. Canens (talk) 10:19, 9 December 2011 (UTC)
  • Support a trial with immediate termination if abused multiple times, as we won't know until we do. ~~Ebe123~~ → report on my contribs. 11:36, 9 December 2011 (UTC)
    • Thanks for your support, Ebe123. To clarify, are you proposing to prematurely terminate the process trial if it is abused multiple times? Something like this seems reasonable but I'd like to get your thoughts on how bad things should get before it's terminated. Dcoetzee 04:10, 11 December 2011 (UTC)
  • Oppose. I would not support any editor getting potentially destructive tools, particularly deletion, who hasn't been given those tools by the community at RFA - the idea that an editor might "need" to be able to delete articles for a week just seems a strange one. There are plenty of editors who could be described as being "in good standing" who I would not trust with these tools. --Michig (talk) 21:33, 9 December 2011 (UTC)
    • Hi Michig, thanks for your feedback. I understand your concern about misuse, but this is the reason there are many controls on apprentices - the 7-day request discussion would offer an opportunity for any interested party to voice concerns about problematic editors like the ones you mention, and even if they pass any serious misuse would lead to immediate revocation. The week trial length is unusually short, but the point is after that they can make a new request in which their actions will be reviewed (originally trials were a month, but in this early stage short trials help to assuage some concerns about things getting out of hand). Dcoetzee 03:42, 10 December 2011 (UTC)
  • Oppose Generally per Kudpung. I believe that the solution to the very real problems with RfA is not in unbundling the tools.--Wehwalt (talk) 12:58, 10 December 2011 (UTC)
I also notice that almost all of the oppose !votes have attracted comments, whereas relatively few of the support votes. That is a practice common at RfA and it is not terribly popular there.--Wehwalt (talk) 13:05, 10 December 2011 (UTC)
  • Hi Wehwalt, I apologise for my inconsistent responses, I've tried to go back and respond to anyone I missed. Many supporters don't raise any concerns or points that I disagree with so I don't respond to them, although anyone else is welcome to. I agree that other reform is needed at RfA, and would be happy to consider other proposals, but I think this one has independent merit. Dcoetzee 03:40, 11 December 2011 (UTC)
  • Oppose. If you want the tools, go through RfA. There is no time limit to get the backlogs cleared. And revocation based on abuse/misuse should not be the only action possible. Try tell a newbie that his account accidentally got blocked because someone wanted to play a cruel joke on him. And all that guy loses is the ability to do the prank again for that account. Plus, it'll be prone to abuse by people with an agenda who create a sockpuppet just to obtain the tool and settle old affairs. Under this scheme, we're giving them another incentive to engage in sockpuppeting activites and will undoubtly increase the workload of checkusers. OhanaUnitedTalk page 14:20, 10 December 2011 (UTC)
    • Hi OhanaUnited, I understand your concerns that malicious users might seek to abuse the process. While a few might slip through and end up rapidly blocked, I believe most sockpuppets would either be dissuaded by the need to demonstrate sufficient experience in the area the tool would be used (which would typically take months), or would be detected during request discussions. Dcoetzee 03:44, 11 December 2011 (UTC)
  • Support Trial The proposal seems well thought out, it is trying to address the issue so lets see how it goes, not sure it is a long term solution; but the trail might help us decide. Mtking (edits) 01:34, 11 December 2011 (UTC)
  • Support - This is the first step towards the very reasonable goal of making tool-use much more sane than the current system. If it works then even better. RfA is indeed broken, but fixing it need not preclude taking this positive step forward. The results of this will also inform the results of future RfA developments. There's no need to be either/or about it. Shadowjams (talk) 23:40, 11 December 2011 (UTC)
  • Strong oppose I'd be happy to explain my reasoning upon request. HurricaneFan25 21:50, 11 December 2011 (UTC)
  • I for one, would be curious to hear your reasoning. (Also did you realize that this isn't in the main comment section?)Crazynas t 23:06, 11 December 2011 (UTC)
    • Moved to main comment section. @Hurricanefan25: during the earlier discussion that you removed from this page, you seemed to be content to support the process trial after it was modified to limit the term to one week instead of one month. That hasn't changed, so I'm also curious what other factor has led you to reverse your decision again. I'd like to ensure it's not a misunderstanding. Thanks! Dcoetzee 12:14, 12 December 2011 (UTC)
      • I re-thought it some; "block" is a very powerful tool and can easily be abused. "Delete" is an iffy — I'd say there's only two current non-admin NPPers out there who I'd trust with it, and there are none who I can trust to judge consensus well. "Protect" prevents many people from editing and just a slight lack of judgement can be a very bad thing. And I don't exactly trust anyone to give or take user rights. Simply put, I don't trust anyone except a select few to apply tool apprenticeship well. HurricaneFan25 15:05, 12 December 2011 (UTC)
        • I'm not sure block is what I'm most concerned about. If a tool apprentice screws up with block, protect or delete, it is undoable. You can bet that the tool apprentices are gonna have people watching them like hawks for screw-ups in much the same way that people do during the week of their RfAs. If a tool apprentice can bite a newbie with a block, they can do the same by deleting their pages or applying rev-del on specific pages (if indeed the delete bundles includes rev-del). —Tom Morris (talk) 17:01, 12 December 2011 (UTC)
  • Support a trial, per other supporters above. I don't see what the big deal is, at least it's a step forward from the status quo. —stay (sic)! 17:41, 12 December 2011 (UTC)
  • Strong Oppose per OhanaUnited and lots of other people above. --Dweller (talk) 22:09, 12 December 2011 (UTC)
  • Oppose I have never supported the various proposals to do something like this for the simple reason that admins need to be admins, not partial or fake admins. If you see a page that needs to be deleted, created by a user who should be blocked, that will now take two people instead of one. If the page needs to be salted as well, that's three partials to do the work that one real admin can do in a few seconds. Look around, despite the lower number of admins we do not have a significant problem with backlogs on crucial areas. This will create more work, not less, in many situations that real admins currently can and do handle just fine with just one person. Beeblebrox (talk) 06:22, 13 December 2011 (UTC)
    • The apprentices are not tasked as replacements for adminship, but trainees for them. I admit it can be a little inefficient, but usually just one of the tools (block, delete, protect) is urgently needed, while the others can wait.Jasper Deng (talk) 06:25, 13 December 2011 (UTC)
I'm not sure I agree with that premise, but if I did I would point out when your partial admin is sitting there holding the wrong tool, the less urgent one, they will have to go and find a real admin to deal with the more urgent problem. Even if the partial has the right tool, they will have to go and find a real admin to do the rest of the work. There is no significant problem with backlogs in critical areas, and therefore no need for an army of partial admins who can only do one admin task at a time. Beeblebrox (talk) 19:01, 17 December 2011 (UTC)
  • Strong oppose. This pointless and dangerous. If the community is going to discuss whether someone should have the tools for a period of seven days, they should be given the tools. It's called RFA. They should not be given the tools only so that a bureaucrat can snatch them back a week later. It's unnecessary busy work for bureaucrats, and doesn't gain much anyway. There are enough active administrators that if someone needs something adminny done, they can grab any one of us and ask. If they aren't comfortable asking a admin on their talk page, they leave one of the many "hey come look at this" templates to grab our attention (CSD tags, {{adminhelp}}, {{editrequested}}, etc.), or leave a note on a noticeboard. This way, someone the community actually trusts is doing the action, so there is some oversight over the actions being taken. If we give the tools to someone who we admit up front we do not fully trust to have the tools on a regular basis, they could easily wreak havoc before one of the very few bureaucrats can be called in to desysop them. The stewards have made it clear they're going to leave such actions to us now, except in dire emergencies. As Beeblebrox points out above, splitting the tools up doesn't help either, as now all these half-admins are going to be tripping over their own feet trying to get things done that we haven't decided to let them do. No. I cannot support this. Hersfold (t/a/c) 05:07, 13 December 2011 (UTC)
    • Hi Hersfold, thanks for your thoughts. While it is true that actions involving careful judgement should be reserved to those the community has judged as possessing that judgement, I think there are many tasks that are straightforward, like, say, deletion of obvious vandalism or attack pages, closing of snow delete AfDs, banning a vandal who received level 4 warning and persisted, etc. Would you consider supporting if apprentices were limited to perform these straightforward tasks? Dcoetzee 10:38, 14 December 2011 (UTC)
  • reluctant support I don't actually see the danger. RfA has passed sockpuppets and other nefarious characters in the past (sometimes with high margins and sometimes twice). And in many RfAs I don't see trust being tested. Rather I see checkboxes for avoiding wikipolitics (but not too much! Don't want to go to RfA without friends) and exercises in heroic patience in the face of stupid questions and nasty or dim-witted opposing comments. Literal use of the tools can create problems but all of them are reversible. And something has to give. However I strongly suspect that this will morph into a defacto requirement for RfA quite quickly if it is enacted. In which case it will have not only become useless but will have added a significant barrier to entry to a process which is already nearly completely defined by its barriers to entry. It could also just move RfA from where it is not to a smaller discussion on the 3 month trial. However in the absence of a magic wand to wave and make adminship no big deal again, I'll support this. Protonk (talk) 05:40, 14 December 2011 (UTC)
    • Thanks for your support, Protonk. I'm also concerned about the possibility of apprenticeship becoming a requirement, and have no way to prevent it other than loud protest - but sticking to my analogy, I don't think we'd eliminate the featured article process because some users want RfA candidates to have a ton of FAs. As in that case, I think this process could have a lot of benefit outside of informing RfA. Dcoetzee 10:47, 14 December 2011 (UTC)
  • Oppose - This is not going to help anybody for several reasons. (1) As soon as this goes live (if it does), we will have users begging for delete/block/protect to chalk up as their newest "trophies". (2) If I can trust somebody with the block tool, why shouldn't I trust him with delete and protect as well, since he should have the capability to understand when he is ready to become more active in using them? On the other hand, if I cannot trust somebody to know when to delete pages, how could I trust him with the block tool when he might block somebody for creating inappropriate pages? (3) There is the problem that an admin commonly needs multiple tools, as Beeblebrox mentioned. I regularly use 'protect' to stop a vandal attack on a page. I then use 'block' on the vandals involved. I then use 'revisiondelete' to hide the revisions of the page that contain grossly defamatory content. (4) I agree with Hersfold that it is "unnecessary busy work for bureaucrats and [won't] gain much anyway." (5) Lastly, and most concerning of all, people will likely start voting: "Oppose, go get experience with block then come back and request adminship." That's my opinion on taking apart the administrator toolbox. Reaper Eternal (talk) 18:23, 15 December 2011 (UTC)
  • Support. I have some serious misgivings about this proposal, and a number of editors I respect have made very cogent arguments against it... but as King of Hearts says, I think it's worth a try if only for the empirical data as to whether such a thing would be a complete trainwreck or a useful addition to our processes. These days the idea of 'rollbacker' being a separate privilege from the admin toolkit is completely uncontroversial; it's not inconceivable that one or more of the other admin tools would lend itself to a similar unbundling. Only one way to find out. 28bytes (talk) 00:56, 17 December 2011 (UTC)
  • Oppose. The bundle of administrator rights is a complicated and interlocking package that is simply technically impossible to cleanly break apart. 'delete' without 'undelete' is asking for trouble, but 'undelete' requires 'viewdeleted' which cannot (per Foundationd decree) be granted to non-admins. The most 'dangerous' permission in the bundle is not 'block' but 'editinterface', which allows the theft of other user accounts through cookie sniffing, including accounts with higher permissions; yet many editors here completely fail to properly evaluate this. This will serve only to create more work for everyone in picking up broken pieces either left behind by apprentices who do not have the tools needed to actually do things properly, or deliberately dropped by people exploiting the new security weaknesses this will create. Happymelon 13:03, 17 December 2011 (UTC)
    Hi Happy-meleon, thanks for your feedback. I want to note the trial does not include the undelete, viewdeleted, or editinterface rights, for precisely the reasons you describe (re delete without undelete, see Wikipedia:Tool apprenticeship#undo). I agree that it'll require some extra work to manage mistakes by apprentices, but I think the work they accomplish would be far more. Dcoetzee 21:42, 17 December 2011 (UTC)
    Hi Dcoetzee. I gathered that vague impression, although there is relatively little clarity on this page or the subject page about the parameters for the trial, something you should definitely consider addressing. But I oppose the trial in any degree because I think it is a useless exercise, given that the problems with the full system are in my opinion (as I described above) insurmountable. As such, running the limited trial does nothing to achieve its primary purpose of demonstrating the viability of the full system. Happymelon 12:44, 21 December 2011 (UTC)
    editinterface is already given as a global right on meta. ~~Ebe123~~ → report on my contribs. 23:37, 28 December 2011 (UTC)
  • Support: Janitors don't need machine guns. Someone with experience and strong evidence of sanity, and no blocks, flaming RfCs against them or "sentences" handed down from ArbCom, should be able to do basic stuff like move a misspelled article over an edited redirect, or add a feature to a template they maintain after someone protected it yesterday because it got popular. Yes, these tools have their risks - some editwarrior could go to a protected article and change it to their preferred POV version. So what? They'd get reported at WP:AN/I and the fun would be over, and they'd be unlikely to ever be trusted again, much less with full sysop bits. There is no reason that people should have to be able to convince the community that they can be trusted with the ability to block other editors in order to receive the ability to get really rather basic things done around here. There are numerous reasons that Wikipedia is losing good editors, and admins, faster than they can be replaced. This is a major one. — SMcCandlish Talk⇒ ʕ(Õلō Contribs. 14:22, 17 December 2011 (UTC)
  • Strong support - this seems to be the best serious proposal for adminship overhaul that has ever come up. RFA has become basically a breeding ground for edit stalking, running the risk of outing, and is more or less like an Internet equivalent of the Spanish Inquisition (and not the Monty Python version). I, for one, would never submit myself to it and know many sensible users feel the same way but who would still be able to help clear up the bureaucratic backlogs that drag down WP's quality. Kansan (talk) 07:47, 19 December 2011 (UTC)
  • Support It's a trial, a short one at that. It's scope is small with little chance for it harming Wikipedia, and a greater chance of helping it. Any theoretical problems with the process will be answered during the trial. Angryapathy (talk) 15:19, 19 December 2011 (UTC)
  • Support; I think apprenticeship could be a positive step forward for wikipedia but there are some valid concerns; a trial would be a good way to make sure, one way or the other. There seems to be a broad agreement that RfA is broken in some way but we never seem to get consensus on exactly what is broken, and any proposed RfA-fix is opposed by others who feel it breaks a vital part of RfA (and imho apprenticeship is not the best solution, but it's a solution that actually has a chance of being tried). So, we need better information... bobrayner (talk) 20:45, 27 December 2011 (UTC)
  • Now seeking closure, see #Seeking closure for RfC. Dcoetzee 08:16, 28 December 2011 (UTC)
  • Support trial - as this idea is worth testing Bulwersator (talk) 11:12, 29 December 2011 (UTC)

No assurance that a "trial" is not a Trojan Horse

EDIT: My opinion has changed based upon the responses to what I wrote below. See bottom of thread. --Guy Macon (talk) 02:56, 8 December 2011 (UTC)

We had a "trial" of Pending Changes and those who supported the "Trial" were promised that it would be "for a two month trial period that begins on June 15, 2010." It didn't end.
Then there was an RFC where the clear consensus was to give it an extension with a drop-dead date of December 2010. It didn't end.
Then there was another RFC with an overwhelming consensus that Pending Changes should be immediately removed from all articles with no exceptions. It wasn't removed.
Then an administrator, acting upon the clear consensus of the community, started removing Pending Changes from articles. He got blocked. The block was quickly removed after community outrage became evident, but there were zero negative consequences to the blocking administrator.
Eventually we managed to kill the undead beast, but to this day there is a clear refusal by those who ran the "trial" to learn any lesson from the fiasco or to acknowledge in any way that consensus was ignored and promises were broken.
We need a firm and clear published policy that promising to try something for a limited amount of time and then breaking that promise will never again be tolerated on Wikipedia. That is the bare minimum required to start to regain the editors' trust.
Because of the continued failure to create a policy that breaking such promises will never again be tolerated, there is zero reason for anyone to believe that any proposed "trial" will end when promised.
Because nobody in authority has stepped up to the plate and announced that allowing a trial to continue without consensus will no longer be allowed, I have to assume any vote for a "limited time trial" is actually a vote for a "trial" that never ends, and thus I oppose any proposal for a limited time trial as being potentially misleading to those voting on the proposal. --Guy Macon (talk) 21:00, 7 December 2011 (UTC)
Guy, there is "nobody in authority" in Wikipedia. That is why the project's administration is such a mess. No one has the authority to make sure standards are consistently enforced, promises are kept, and lessons are learned and institutionalized. A configuration control board or some other governance committee would be able to do what you suggest, but guess how that idea would fly if someone suggested it? Too many administrators and editors in Wikipedia like its current, chaotic, anarchic state, because it allows them to get away with doing what they want to do. Cla68 (talk) 23:06, 7 December 2011 (UTC)
We often get these messages from vandals and socks, and it should have some sort of truth. Wikimedia's sysops were responsible for PC, but here, it's easy to stop.Jasper Deng (talk) 23:11, 7 December 2011 (UTC)

Nobody knows better than I do, how incredibly frustrating "Pending Changes" was to all of us. However, we cannot stop all progress because one single trial was so terribly mismanaged. The PC trial was implemented by the Wikimedia Foundation; this trial is proposed by the community itself. The PC trial was vague from the outset, and support/opposition was obfuscated throughout. From the very inception, this trial says, very clearly, One week before the end of the trial, bureaucrats will immediately cease to close requests or give out any tools. Because anyone who received a tool during the trial will have the tool expire one week after they received it, all tools given during the trial will be revoked by the end. The trial will not last longer than 3 months, and will not be extended.1 If that needs further clarification, we can do so. Is there any possible way that you could be reassured that this is, genuinely, a time-limited trial which will not be extended?  Chzz  ►  23:26, 7 December 2011 (UTC)

I intend to personally ensure the trial ends at the specified time. The request pages will be plastered with warnings not to close after the end date. If bureaucrats for whatever reason fail to remove the permissions by the specified time, I will instruct the users to not use the tools or else be blocked until the tool is removed (now added to the proposal). If bureaucrats continue to grant permissions after the one-week-before-the-end point, I will also instruct those users to not use the tools or else be blocked until the tool is removed, and chastize the bureaucrats, reporting them if it continues. Nobody with any sense is going to block me for enforcing my own trial's duration. I take the end date very very seriously. Dcoetzee 02:15, 8 December 2011 (UTC)
The above comments satisfy my concerns completely. I am 100% assured that this won't be another debacle. Thank you all very much for the thoughtful responses --Guy Macon (talk) 02:56, 8 December 2011 (UTC)
I completely agree with the absolutely necessity for the end date of the trial to be adhered to, regardless of how successful it is, and I'm glad that other supporters including the proposer are stopping at nothing to ensure that this will be the case. —WFC10:25, 8 December 2011 (UTC)
If there is one thing I am sure we can all agree on it is that trial periods need a specific plan for what happens when they are over, and however this turns out I and I am sure many others appreciate the effort made towards ensuring that there is a clear plan for what happens after. Beeblebrox (talk) 06:26, 13 December 2011 (UTC)

Evaluation criteria

It is a bit confusing that the term "trial" is used both for the period (variously given as one month and one week) that a particular "apprentice" (individual user) is allowed to use a sysop tool, and for the period (now apparently three months) that we (collectively) are supposed to run this experiment. For either case, it is not clear to me how the results will be evaluated after the trial period, and what the consequences are.

  • Suppose an apprentice gets the tool to protect articles, and uses it with deft skill in an entirely appropriate manner, beneficial to the project. Now what after the trial period expires?
  • Suppose that after the three months, it is clear that some apprentices used their tools well, most did not do much good, but also little harm, and a few were bad apples. Now what?

 --Lambiam 11:35, 9 December 2011 (UTC)

Hi Lambiam, I apologize for the overloading of the term "trial" (I couldn't think of a suitable synonym) - I've been using "trial" versus "process trial" to distinguish them. Users will receive tools for one week during the process trial (one month was in an older version of the proposal - I missed a couple updates on Wikipedia:Requests for tool apprenticeship but got them now). Users always lose their tools after their trial period expires, but can then make new requests, where their past performance is the main factor in determining whether they'll receive a new trial.
I reasonably expect that the report giving feedback on the process trial will examine the most and least successful users, their requests, how problems were handled, etc. and propose revisions to the process in order to have more successful users in the future. Once these revisions are complete, it will be up to the community whether they think the next phase of the process will be worthwhile or not. I think it's logical to assert that the net benefit to the project should outweigh the overhead of managing the process, in the long run, so I would certainly hope for more than a zero net gain. Let me know if this answers your questions. Dcoetzee 12:59, 9 December 2011 (UTC)
If there can only be 10 apprentices at a time, then, assuming most use their tools productively and appropriately and request renewal each time, either most requests for renewal will not be honoured, or there will not be that many different apprentices in the full process-trial period. (For example, if 90% of the renewal requests are honoured, there will be at most 22 different apprentices in total.)
You mention "the next phase". Are there already ideas about this next phase, assuming the process trial by itself is a resounding success?
Is there a presumption that apprenticeship is a stepping stone to adminship? Personally I'd be interested in a (very small) subset of things that admins can do but ordinary users cannot, but I'm not interested in eventually requesting admin status.  --Lambiam 23:17, 9 December 2011 (UTC)
Hi Lambian, good questions. Yes, it's possible there will only be a small number of users involved in the process trial. I think this is acceptable, and should produce sufficient data. I don't have many ideas about the next phase yet, whether it'll be another trial phase or a full rollout, and what'll be different about it - that will all depend on how this one turns out (regardless, it would only proceed with community approval). And no, it is not presumed that all apprentices wish to go on to adminship, although it is intended to encourage that. People who really just want to help get work done and nothing else would absolutely be welcome to apply. Dcoetzee 03:20, 10 December 2011 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.