> We have named this optimization 'PartialExecuter', the key idea behind it being taking advantage of known function parameters to find inner code blocks that cannot ever be possibly executed.

I'm wondering, what exactly are the differences of this approach from "classical" partial evaluation? This seems to be a special case of it, unless I missed something.