sorry...

but i don't think translating assembly code to C code that basically emulates the assembly code and doesn't look anything like "real world" C is really what we're looking for here..

loop detection is one very very small piece of the problem. please refer to my earlier post.

it's obvious that after 12 yrs of work martin ward has done a really good job of building a tool that's extremely useful to migrate from assembly to C that is functionally equivalent and obviously really makes a huge difference to his business and enables him to really do a good job in his market. i'm not criticizing him here, i'm basically saying the people pointing to this work and saying it proves anything about automated binary analysis and 'real' decompilation are jumping the gun as badly as the people who hand wave when asked about the indirection problem.

this whole thing is a great example of the classic 'oh that's easy.. it's just a small matter of programming' that basically lasts as long as it takes for people to stop theorizing and start trying to implement. the fact it's taken martin ward TWELVE YEARS to get as far as he did says a lot.

i'm still trying to think of something cool to do with dean's IDARuby plugin.. and still running into the useful vs possible conflict :)