In this project I was helping a grad student run some tests for his master's project. He needed to switch a few hundred volts across a gas-filled radiation chamber, so the obvious choice was an h-bridge. Since this was just a test of the chamber to get some measurements, the hbridge never really left the breadboard stage, but it still performed nicely.
I really dislike bootstrap systems, so I chose to use optoisolators and a floating supply to drive the high side gates. Quick and dirty, but it works. I also added optoisolators to the low side gates to make sure they have the same turn on delay as the high ones.