>The problem with this interpretation is that SR inside PN would
>end up being highly nondeterministic...
>Edward
Yes, this is true. But it is also possible to write nondeterministic SR
models because of improper use of NonStrict actors. I have myself done
this sometimes. More experimentation can reveal if a more flexible
interpretation of SR inside PN is justified.
One trivial case is when you have topologically distinct group of
actors that don't share any boundary input port within the SR
composite. In this case, you can separate each group in different
composite actors, thus exploiting more concurrency.
I have being working on a more restricted version of the SR model.
Under this scenario,I think it is possible to have a more relaxed PN
-> SR interface. If anyone is interested in hearing, let me know.
Ivan
----------------------------------------------------------------------------
Posted to the ptolemy-hackers mailing list. Please send administrative
mail for this list to: [EMAIL PROTECTED]
