Skipping Refinement · HackerPapers discussions

Skipping Refinement — mentioned by mrefj

Thu, 22 Dec 2016 03:50:39 GMT

> My problem is that some reductions require moving certain transitions backwards or forwards in time. This is definitely a more involved problem. A cleaner solution, I believe, requires a new notion of correctness than trace containment/simulation up to stuttering. This is because neither of these notions allow one to move forward or backward observable transitions. The key here is to differentiate between non-observable and observable non-determinism. While the former can be accounted by using prophecy variable in linear time framework or directly in the branching time framework. I will…

Paper and discussion history. A mention is not an endorsement.

Skipping Refinement — mentioned by mrefj

Mon, 12 Dec 2016 14:35:28 GMT

> But reasoning with functions is a specific level of abstraction. Excellent point. The level of abstraction is a fundamental concern, not just in modeling but also in reasoning about systems. And state machines (as used in TLA or otherwise) add temporal abstractions to the mix of abstraction mechanism. To add to its power, it does not lead to any loss of generality; since any program in a language with well-defined operational semantics can be modeled using state machine (transition systems with infinite state space to be little more precise). > In TLA, bisimulation and trace equivalence are…

Paper and discussion history. A mention is not an endorsement.