Tue Jul 21 11:43:02 CEST 2009
semantics and static analysis
The more I see the structure of what I'm trying to do, the more
"normal" it all becomes. This is good. Looks like I'm about to find
a bit more connections to existing approaches.
1. a well-defined operational semantics (relative wrt a single
abstract point, not the target's concrete machine language) will
help for "interpretation" based analysis: either full simulation or
simplified abstract interpretation.
2. this can be used to assist (by verification and test) in building
locally-specified higher level semantics for sublanguages,
consistent with the operational semantics.
The idea is to start from the machine, and simply increase the
abstraction level at certain points + provide a set of tools to do