I don't see a problem with the system. It seems to me that any abuse of the system is by far the exception, not the rule.
We need to be as encouraging as possible of the asking of good questions. A case has been made (see here for example) that many people, for example graduate students or assistant professors, find the atmosphere at MathOverflow intimidating as it is, and thus withhold what might be nice questions. Some people even feel that awarding 5 points per upvote isn't enough (it used to be 10 points in MO.1).
On the other side, it has occasionally been recommended (without a view to enforcement) that people try first to get a sense of the culture by answering a few questions before asking a question themselves. This came up in discussion here: http://mathoverflow.tqft.net/discussion/1438/mochizuki-proof-of-abc/. (I think this is too high a barrier to set as a policy, but the sentiment that people should look around a little first to get a sense of the place is, I think, a sound one.)
Overall, it's pretty hard to get away with asking many bad questions (bad for MO, I mean). There are not a few users (including the OP, Stefan Kohl) who invest a good deal of time maintaining quality control. The occasional controversies that erupt are not currently, in my opinion, cause for serious alarm (in fact, they usually have a silver lining, in that those fine mathematicians who line up on one or the other side of the controversy show they really care about fostering the health of MO).
I don't think the questions of any one user has quite approached flood level. We do see now and then some (arguably) overly energetic people who eventually become a nuisance to others; moderation here can be tricky and calls for patience while the evidence accumulates. I am sure that the official moderators move too slowly for some people's taste and at the same time too quickly for others, but that's the way it sometimes goes!
