Wikipedia:WikiProject on Adminship/a la carte

In Wikipedia's early days, there were few regular Wikipedians. Those who were there knew each other more or less intimately and could judge each other based on extensive interaction. And policy? Policy was vague. The few admins there were could have a quick chat on the mailing list if they weren't sure how to apply it, or they could just be bold and see if anyone objected.

Now things are different. As the community has grown, it has become more distant and anonymous, with relatively few people knowing each other or being able to judge each other. More importantly, policy has developed extensively; there are 48 official policies, 61 guidelines, 45 style guidelines, 50-odd naming conventions, and of course the 145 essays that everyone will expect you to have read.

This is necessary: consensus can no longer be gauged by any one person, or by a quick discussion on the mailing list. It must be carefully considered and then written down, so that everyone is working from the same set of principles and the same issues need not be rehashed over and over again. Over time, good decisions accumulate; there are excellent reasons for all of the instructions in our dozens of policies. But it's still difficult and time-consuming to learn them all.

Yet, an admin is affected by all of them, excepting only the rules for RFA and perhaps some miscellanea. Ordinary users are affected by most of them as well, but then, if an ordinary user violates policy it's quite easy for anyone at all to reverse the violation. If an admin violates policy (which, in this case, means disregarding carefully-crafted and well thought-out rules out of simple ignorance, rather than knowing them but making an intelligent decision to ignore them in the particular case), who will reverse it? There are only 846 others who could possibly do it, as opposed to the 48,380,708 who could reverse an ordinary user's actions. That's a ratio of 57,188 to 1, not even counting anonymous editors!

So while all users should know relevant policies, more policies are relevant for admins and it's more important that they know them (even if, knowing them, they may make an intelligent decision to ignore them in a particular instance). So how can we be confident that our admins will be familiar with the workings of all of the things their privileges give them the right to manage?

A solution

Most Wikipedians confine themselves to a limited number of pursuits within the project. Some of these pursuits might benefit from adminship: vandal-fighting, copyvio-elimination, AFD participation. But does a PUIist really have to know anything about AFD to get the tools he needs to handle copyvios more effectively? Well . . . yes, he does, because you see, he may decide at some point that he feels like closing some AFDs. And while he's presumably trustworthy, he still might not be very familiar with the process, or he might have radical views on inclusionism/deletionism, or who knows what. So there's need for caution — and that means people legitimately proficient in copyright don't get the tools they could use to benefit the encyclopedia.

It goes without saying that this isn't good. However much people may say "adminship is no big deal", or "if you're trustworthy enough for blocking you're trustworthy enough for deletion", it's simply not the case, as can be witnessed in many RFAs. Dedicated vandal-fighters routinely get judged on their AFD participation in deciding whether they should be allowed to block. Someone who pledges to keep down the backlog at CAT:PER will garner opposes for his lax stance on copyright. Useful talent is being wasted.

With the question posed this way, the solution is obvious. Simply make a number of groups complementary to adminship, each with more limited privileges than that rank. Tailor them all to common tasks, and restrict the candidates to those tasks. Then only knowledge of one or a few policies, and the candidate's attitude on only one issue, need be considered when judging whether to give someone tools to benefit the encyclopedia. This way, the interests of those who would perhaps be legitimately questionable at adminship as a whole can be channeled into one or a few issues, with their dedication and attitude to them more easily evaluable.

Following is a mock-up of a possible breakdown for these groups:

antivandal

Rights: block, rollback, protect

Comments: Revert vandalism, block vandals, and protect against vandalism.

copyright

Rights: delete, undelete

Comments: Handle PUI and other copyright-related admin issues.

deletion

Rights: delete, undelete

Comments: Handle AFD and its kin, as well as the non-copyright-related CSDs.

tech

Rights: protect, editinterface

Comments: Handle high-use templates and editing of the interface.

enforcer

Rights: block

Comments: Enforce 3RR and ArbCom violations, and handle sockpuppet accusations.

Collectively, these cover every admin task. Each has its boundaries; while occasional use of privileges outside one's assigned sphere could be tolerated, consistent or long-term infractions would result in summary removal of rank. This will assure those who consider the candidate that they need only focus on what's germane to the rank itself, without thought to other issues whose rights may overlap. It will keep the discussion focused on what's important: what the candidate can contribute to Wikipedia in a specific sphere.