And to follow that up: there are quite a few nuclear-blast calculators on the web, mostly based on Glasstone and Dolan (you can even do it in your mobile phone). Using this one:

http://meyerweb.com/eric/tools/gmap/hydesim.html

I find that the 15psi overpressure distance for a .2KT blast is .09 miles. That implies total destruction of even the strongest buildings. The Fukushima reactors are not that far apart, and even the building around the reactor that exploded wasn't totally destroyed. So Donuel's guess was way off. The explosive force of a buildingful of hydrogen-oxygen mixture sounds about right to explain what happened. Less hydrogen than the Hindenburg fire, but better mixed with air, hence more explosive.