Attendees:
Mark Hahn, Cadence (Chair)
Jin-sheng Shyr, Toshiba (Co-Chair)
David Defourque, Fujitsu
Vassilios Gerousis, Motorola
Greg Schulte, Ambit
Andres Teene, Symbios
Next Meeting:
The next meeting will be a teleconference on Wednesday, 5/27/98,
from 9-11 PDT. Based on feedback from the group, we will switch
the on-going day and time to every other Tuesday at 9-11 PDT,
starting on June 9th.
Action Items:
Who When What
---------- -------------- --------
1. Jin 5/22 Send an updated version of the
charter/scope document to the reflector
2. Mark 5/27 Set up a link from the top level web page
to the attachments directory.
3. Mark 5/20 Do an e-mail poll on changing the meeting
time change
4. Mark 5/27 Work the legal issues of the Ambit donation
with Vassilios
5. Greg 5/22 Send the Ambit document to Mark for posting
on the web site.
Details:
1. Charter/scope discussion
We reviewed the revised document from Jin
- "implementation cycle" should be changed to "design cycle"
- The overlaps with the SLDL and RAIL efforts should be
handled and described in the same way as PDEF and P1500.
- Board level design constraints should be deleted from the
list of explicitly excluded areas
- Vassilios suggested that we should be explicit about whether
we're going to include a constraint exchange format within our
scope, and if not, what we expect will be done instead.
3. Constraint dictionary discussion
Mark suggested changing terminology from "constraint dictionary"
to "constraint taxonomy".
We discussed approaches to ensure consistency and avoid
duplication in the documentation for the constraint taxonomy,
a syntax-independent description of the full constraint semantics,
and the constraint description language.
The syntax-independent description of the semantics is tricky
because we will need to talk about each of the parameters for
a constraint, but there won't be a keyword to refer to like in
the syntax document. We agreed to use the attribute names from
the EXPRESS model for this.
The document for the constraint description language will need to
cover both syntax and semantics, to make it usable for designers.
The problem is that we need to ensure consistency between the
semantics in the syntax-independent and in the syntax-dependent
documents. We agreed that we should have a single semantics
description, and use conditional text to manage a single source
document for both.
4. Status update on Ambit constraint language
donation
Greg said that the Ambit document was ready, and that he would
send it to Mark for posting on the web site. We talked about
legal questions related to the donation, which didn't seem to
be a problem, but we need to make sure that right process is used
for the donation and that the terms and conditions for using it
are well defined.
Greg thought that we would be free to use it as-is, build on it,
use parts of it, or not use it at all. Ambit's main concern is
that we shouldn't use the same name but different semantics for
constraints, which would cause backward-compatibility problems.
This shouldn't be an issue, given that we will probably prefix
all commands with a unique id like "gcl_" to avoid name collisions
when interspersing constraint commands and other tool-specific
commands in a tcl script.
Thanks,
Mark
-- Mark Hahn phone: (408) 428-5399 Architect, Deep Submicron Business Unit fax: (408) 428-5959 Cadence Design Systems email: mhahn@cadence.com