I'm sympathetic to some of your complaints about SE, but I think this is a mistake: > SE's problem is that accretion wins over refinement. IMO it should have b

I'm sympathetic to some of your complaints about SE, but I think this is a mistake: > SE's problem is that accretion wins over refinement. IMO it should have been more like a Wiki, with a much clearer distinction between content - i.e. stock answers, with comments - and debate about content. You can see the effect of this approach in many cases where the community has tried to establish a "canonical" Q&A about some recurring topic, and it's frequently not pretty; every tangentially relevant detail that somebody thought to tell the world about gets edited in, and then never gets edited out. Unless somebody is willing to purge lots of users' contributions entirely (which they generally aren't), this one great beast of an answer just continues to grow and grow in scope, gradually losing any sort of coherent narrative or direct relevance to the question. The failed Stack Overflow Documentation project failed in a similar way. Wikipedia articles about programming topics also seem to me to generally be tedious, confusing and riddled with inaccuracies. The competition-based model that SE uses certainly has failure modes, but a more collaboration-based model also has failure modes - ones that I think are more serious and crippling.