Subject: RE: Gate connection
From: Prinz, Andreas (prinz#DResearch.de)
Date: Wed Apr 03 2002 - 10:01:40 GMT
Hi Thomas,
> could you please elaborate on this issue? How do you, in the abstract
> syntax, distinguish between a gate on y1 and a gate on y2?
> Note that both y1
> and y2 are instance sets of the same agent type.
Since Martin seems to be unable to answer this, I will try
to find a solution.
Lets first start with an example
(according to your description):
1 system s;
2 signal sig;
3 block type b;
4 gate g in with sig out with sig;
5 endblock type;
6 block b1: b;
7 block b2: b;
8 channel from b1 via g to b2 via g with sig;
9 endsystem;
As far as I understand, we have in SDL the concept of a
full path qualifier, which can be used to unambiguously
identify any entity within a specification (at least it
used to be like this until SDL'96). From my point of view
this is still an important concept, because the semantics
relies on many places on this property of full qualifiers.
The question is now, if we have broken this property by
introducing the agent types as core concepts.
To put it differently, the question is what is the full
qualifier of the identifiers g on line 8 above.
So what about the following expanded line 8?
channel from <<system s>>b1 via <<system s/block b1>>g
to <<system s>>b2 via <<system s/block b2>>g
with <<system s>>sig;
In fact this means that the full path of an entity is the
path of the *instance* as opposed to the path of the
*definition*, which would be <<system s/block type b>>g.
Best regards,
Andreas
P.S.: I think this is a general problem with the introduction
of the agent types into the abstract syntax, which has also
problems on other places. We have to assure that the full
qualifiers are unambiguous, or we have to change the structure
of the SDL semantics.
--
Dr. Andreas Prinz
DResearch Digital Media Systems GmbH
Otto-Schmirgal-Str. 3, D-10319 Berlin, Germany
Fax: +49 (30) 515932 299 http://www.dresearch.de
Tel.: +49 30 515 932 256 | +49 30 386 26018
mailto:prinz#dresearch.de
This archive was generated by hypermail 2a23 : Mon May 05 2008 - 20:30:55 GMT