As I think it was not yet mentioned explicitly in relation to your post: There is a site http://math.stackexchange.com/ which is on the one hand very similar to the present one yet on the other hand more open regarding questions. It seems to me that site might be what you think MO should be. (There are many peole that contribute on both sites).
]]>I think there should be a community gathering, called by the moderators, to review and reinforce policy. There is enough of an active user base and enough common issues on which there is not community consensus, that such a gathering makes sense to me. I feel that lack of such a redefining of MathOverflow will result in many discussions about what is proper, with no resolution, and with a possible fragmenting of and erosion of the community membership.
Alternatively, a poll could be conducted so that people are aware of the strength of each other's desires on various issues, e.g. research vs undergrauate level, resource for academics vs resource for significant fraction of the scientifically minded public, big list questions or no, J.R. Martinez or Rob Kardashian (you know, the important stuff). If even just the voting-to-close-or-open population were aware of the results of such a poll, I think it would add strength to MathOverflow the forum. (Plus, we might help J.R. get the top spot.)
Gerhard "Does Not Watch Dancing Shows" Paseman, 2011.11.03
]]>