I don't think WiMax was meant to solve the problems to which people have applied it. If you look at WiMax actual range, it's quite good, but the advertised range is at least double (hence all of the big ClearWire lawsuits).
I don't think setting up city wide wifi is unreasonable with present technology. I think it's an entirely attainable goal, the things that hold it back are bureaucracy and the same individuals who hold back things like project lightspeed.
But my broad point in response to your comment is that the benefit of saturating san francisco with AP's are manifold, whereas the costs and risks are minimal.
Sf has a population of over 6 and a half million and covers 600 square Km. I woudl love to see your estimate for how much that woudl cost to build and maintain.
Less than 100M to implement and under 10M to maintain per year.
I definitely don't think it's expensive compared to the benefits. Personally, if I'm pipe dreaming, I'd rather have ubiquitous 1gigabit fiber, which google pegged at 11Billion for almost all of the US to be covered.
