But yeah, if the author wanted to make a part 3 it might be interesting to see his implementation for something like (a simplified subset of) Java bytecode.
A good thing about BF, IMHO, is that its syntax is trivial to the point of almost non-existence, and that its memory model is C-like.
I agree that the language itself is very likeable for the reasons you stated, but I'm partial to things that I can easily check. It's called Brainfuck for a reason. ;)