|
|
 | | From: | Edward A. Lee | | Subject: | Re: Possible process domain bug | | Date: | Fri, 17 Dec 2004 00:09:51 -0800 |
|
|
 |
The problem with this interpretation is that SR inside PN would
end up being highly nondeterministic...
Edward
At 02:58 PM 12/16/2004 -0800, Christopher Brooks wrote:
> > The case is that I don't totally agree with the semantics of the
> > ProcessBased + SR. One of the tests that I have attached use the timed
> > PN director to show this.
> > Basically, the SR composite only fires when the transferInputs() method
> > returns,
> > this is, after reading exactly one token from each input, since the PN
> > receiver
> > is blocking. But by SR's nature, it should react to events and not
> > necessarily wait for
> > ALL inputs. Two possible alternatives are :
> >
> > (1) firing as soon as a new token is available for the SR composite.
> >
> > or (2) firing based on a input patterns of available data. The
> > reasoning behind
> > this is extending the concept of causality of the atomic actors inside
> > the SR composite and
> > its related dependencies graph, to the composite boundary. For each
> > output definition of the
> > SR composite, you should known the possible input patterns.
------------
Edward A. Lee, Professor
518 Cory Hall, UC Berkeley, Berkeley, CA 94720
phone: 510-642-0455, fax: 510-642-2718
eal@eecs.Berkeley.EDU, http://ptolemy.eecs.berkeley.edu/~eal
----------------------------------------------------------------------------
Posted to the ptolemy-hackers mailing list. Please send administrative
mail for this list to: ptolemy-hackers-request@ptolemy.eecs.berkeley.edu
|
|
|
