Subject: [sv-ac] SVA clock flow proposal
From: John Havlicek (john.havlicek@motorola.com)
Date: Thu Dec 11 2003 - 11:36:54 PST
All:
In today's meeting of the PSL/SVA Alignment Subcommittee,
my clock flow proposal for SVA 3.1a was discussed, and it
was pointed out that the treatment of the clock as a
declaration with scope rules in this proposal is a divergence
from PSL, which treats the clock as an operator with
high precedence.
I have been conflicted on this issue because I have been working
hard both for alignment and for the SVA 3.1a extensions, and in
this case I have come to the conclusion that a PSL-aligned solution
really is not good for SVA.
The two key reasons that I have come to this conclusion are:
1. I believe the clock scoping in SVA should be intuitively
consistent with the flow of local variables. PSL does not
have local variables, so this is not a language design goal
for PSL.
2. SVA has always forbidden assertions that require checking
simultaneity of distinct clock events. Therefore, I think it is
a bad language design decision for SVA to require users to write
clocks where the clocks cannot change. PSL does not restrict
where the clocks can be placed, so this is not a language design
constraint for PSL.
Best regards,
John H.
This archive was generated by hypermail 2b28 : Thu Dec 11 2003 - 11:37:55 PST