Subject: ECBV Function Semantics, revised
From: John Havlicek (havlicek@adttx.sps.mot.com)
Date: Sat Mar 30 2002 - 07:28:58 PST
Dear Committee:
The ECBV Function Semantics document I sent last night
has a restriction that I had intended to relax. The
attached document has the relaxation.
The substance of this change is the following. The original
version said that function instances within task definitions
should be statically evaluated as they appear in the task
definitions. This means that constant parameters passed in
a task instance cannot be used in trying to resolve the
function instance. I had intended not to be so restrictive.
The new version says that function instances in task definitions
should be statically evaluated as the task instances are
unrolled in forming the dep() digraphs (described in the
ECBV Statement Semantics document).
The static evaluation of the function instances within task
definitions is still a part of the compilation process, but it
is no longer reasonably characterized as a "pre-processing" step.
Therefore, I also changed the phrase "pre-processing step" to
"compilation step" in describing the way function instances are
statically evaluated and replaced by plain expressions.
I apologize for any confusion caused by this oversight.
I am currently working to make final revisions and corrections
to the previously submitted statement and matching semantics
documents.
Best regards,
John Havlicek
This archive was generated by hypermail 2b28 : Sat Mar 30 2002 - 07:32:28 PST