looking at archives, arsdigita seemed to be a good and useful balance of explaining what problem is being solved and why it is being solved that way -- from any particular viewpoint in the system. why? probably because they had someone to enforce the rules and salaries depending on it.
i still think getting a bunch of people to agree on a required best solution is a sinkhole. there needs to be some kind of motivation to define and follow the standard. it needs to be loose enough to fluctuate based on the ideas of the community as a whole.
1. define the group that will care about this stuff.
2. get everyone to agree on a set of recommended standards for docs, code, and processes. it doesn't have to be perfect, just something everyone can agree on at that time. you can follow these, but you don't have to.
3. if you choose not to follow the standards, you are an evil "unsupported outsider" unless you can build a case to the standards group for why your method should replace or be incorporated into the current standards.
4. those that do follow the standards get community priority in support, placement, recomendations, and distribution. this is heroic.
5. the group requires periodic checking of existing code to make sure it fits current standards. if not, you get bumped to evil-unsupported-outsider status.
6. make the whole program very visible to the community.
