I'm currently reading a few papers based on STAIRS, whereby it's discussed that there is a distinction in refinement treatments of sequence diagrams between alt blocks that permit underspecification (eg, |~|) and xalt blocks that require all alternatives to be reachable (eg, []). This is a note to myself that I need to consider this distinction and possibly add something similar to robocert.
One interesting thing that I haven't quite got into yet is how this interacts with guards on the legs of the alt and xalt. Maybe these just legitimately become &-guards at the CSP level, with all the semantic ramifications one would expect.
I'm currently reading a few papers based on
STAIRS
, whereby it's discussed that there is a distinction in refinement treatments of sequence diagrams betweenalt
blocks that permit underspecification (eg,|~|
) andxalt
blocks that require all alternatives to be reachable (eg,[]
). This is a note to myself that I need to consider this distinction and possibly add something similar to robocert.One interesting thing that I haven't quite got into yet is how this interacts with guards on the legs of the
alt
andxalt
. Maybe these just legitimately become&
-guards at the CSP level, with all the semantic ramifications one would expect.