It was an academic project with no almost outside participation, very limited documentation and out of touch with the community that could really use it. I am not surprised they didn't get anywhere. Good for the founders though, it looks like at least they got interesting jobs now.
I think one of our developers has heavily been using Boomerang on C++ code with some local patches... I doubt I could ever convince him to commit them to the tree and/or become a project evangelist, even if he was extremly excited and a heavy user of it. A really good decompiler won't come until we've perfected machine logic/AI a lot more, until then, the closest you can get to is raw pattern recognition/statistical analysis.