Re: Prioritization on SVCC's Charter, API issues

Subject: Re: Prioritization on SVCC's Charter, API issues
From: Bassam Tabbara (bassam@novas.com)
Date: Mon Oct 07 2002 - 16:31:08 PDT

Yeah Let's discuss. Although we'll definitely need to reframe your questions, they remind me of how poll questions are posed as in a) do you like this nasty thing, b) or this good thing :-)! Just kidding ....  (and don't ask me to propose an alternate, I'll know it when I see it :-)!).

-Bassam.
P.S. The web site says we are not meeting tomorrow ?? My calendar says we are, who's right ?

"Warmke, Doug" wrote:

 

SVCC Team,

**** Prioritization Issues ****

I agree with Stuart Sutherland and John Stickley that it
would be a good idea to prioritize our work to make sure
we get something valuable done by our SV 3.1 deadline.

In my discussions with users, I find that there is high
demand to have a direct name-based binding between
the C and HDL language domains (aka DFLI).  I hear
about this so often that I think this topic is the
most important technical issue for the SV-CC committee
to cover by the deadline.

Note: I had composed this mail last Thursday but didn't
send it yet.  It looks like Michael R. agrees with
this assessment.

Also, my Mentor Emulation Division colleagues agree
with this prioritizationy.   Does anyone disagree?
If so, let's get the emails out and then discuss
in the meeting tomorrow.

**** Thoughts on API for SV ****

As several people have pointed out, it is difficult to
write API's for language features that are still in flux.
I understand what Kevin is trying to propose, but I can
see the following problems with his suggestions:

    a) There is a large body of legacy PLI/VPI that must
       continue to be supported in SV.  (anyone disagree?)
    b) Introducing the C/C++ elf/stabs-based debugging tool
       approach is an elegant concept, but if we introduced
       that and continued with PLI/VPI, the system would have
       redundant parts and thus be overly complex.
    c) I don't think that there is a large body of elf/stabs
       expert programmers in the EDA industry or the
       semiconductor industry which we serve.  There are
       a few real experts, but not a large body.

How will we put this issue to rest and decide
"Yes we are going to put VPI into SV (and possibly extend it)"
or
"No we are rejecting all API access to SV and will enhance
the language to do all the jobs that the old C code did"
?

Let's discuss this at the meeting tomorrow.

Thanks and regards,
Doug Warmke
Model Technology, Inc.

-- 
Dr. Bassam Tabbara
Technical Manager, R&D

Novas Software, Inc.
bassam@novas.com
(408) 467-7893
 

This archive was generated by hypermail 2b28 : Mon Oct 07 2002 - 16:34:44 PDT