How are policy questions revisted?

Question

What is the protocol for revisiting general policy decisions on meta? Specifically, I'm not talking about debating whether or not a particular situation warrants an exception to a rule. Rather, I'm asking about situations where the rule itself needs to be revised or replaced in terms of being a canonical answer.

As context:

• This question recently came up on meta, citing a much older meta question with what appears an out of date accepted answer.
• I proposed these guidelines regarding homework questions which were marked as an accepted answer. Later, after attempting to apply them, I feel that they omit important criteria. Making any significant changes to it after the fact doesn't seem like a good idea as votes may no longer reflect the extent to which others agreed (or not) with it.

In general, how does the Stack Exchange model handle this type of situation? As a potential solution, could community wiki answers be useful here? They would allow the community at large to update answers to policy question to reflect evolving norms.

Answer 1

Generally I'd expect that in order for us to adopt some change to an existing policy, the answer with the proposed change should at least outscore the answer that was originally established as policy (and maintain that for some period of time; I usually wait at least three days to a week to assume that votes on tag blacklisting and such are settled, for example, and those are fairly minor).

---

Having a place where all the current policy decisions are listed, as we do with all the blacklisted tags, is definitely something worth exploring. There's a policy tag here that may help with that. Although I don't think it's necessarily a great way to deal with identifying what is and isn't current policy, because it's easy to edit out "maliciously" (and hard to see when somebody would have done so), and also easy to forget to go back and remove it from abandoned policies.

I think a big meta-meta post like our tag blacklist post is probably better.