Subject: Accellera FV -- October 2000 Meeting Minutes (Working Group Rules)
From: Harry Foster (foster@rsn.hp.com)
Date: Wed Dec 06 2000 - 07:04:54 PST
I'm re-sending our October and November meeting minutes (separately)
to the new mail reflector so that they are maintained as permanent
records in the archives. All messages are archived under a 'hypermail'
archive at http://www.eda.org/vfv/hm and are threaded by
date, author, subject, and thread.
I plan on summarizing our December meeting over the weekend and send it
out the first part of next week.
Best regards,
-Harry
================== NOVEMBER 2000 Accellera Meeting Minutes =================
OVI Formal Verification Working Group Meeting - 2 October 2000, 9-11am PDT
Chair: Harry Foster Co-Chair: Carl Pixley
MeetingPlace Telephone Number: 404-774-4103
Scheduled Meeting ID: 9999
Meeting date (mm/dd/yy): 10/02/00
Start time (hh:mm) : 09:00 AM(US/Pacific Time)
Meeting name: OVI Formal Group
Attending (either in person or via representative):
Carl Pixley (Motorola) <pixley@adttx.sps.mot.com>
Claudionor Coelho (Verplex) <coelho@verplex.com>
Danny Geist (IBM) <Danny_Geist@il.ibm.com>
David Giramma (Synopsys) <davidg@synopsys.com>
Ed Clarke (CMU) <Edmund.Clarke@cs.cmu.edu>
Erich Marschner (Cadence) <erichm@cadence.com>
Harry Foster (HP) <foster@rsn.hp.com>
K.C.Chen (Verplex) <kchen@verplex.com>, represented by Bwolen(Verplex)
<bwolen@verplex.com>
Peter Flake (Co-Design) <flake@co-design.com>
Hillel Miller (Motorola) <hillelm@msil.sps.mot.com>
Tom Fitzpatrick (Cadence) <tfitz@cadence.com>
Tom Anderson (0-in) <tla@0-in.com>
Swami Venkat (Synopsys) <swamiv@synopsys.com>, represented by Arif Samad
(Synopsys) <arif@synopsys.com>
Vassilios Gerousis (Infineon) <Vassilios.Gerousis@infineon.com>
Yaron Kashai (Verisity) <yaron@verisity.com>, represented by Mike
McNamara (Verisity) <mac@verisity.com>
On the email list, but not attending:
Jayaram Bhasker (Cadence) <jbhasker@cadence.com>
Sudhir Kadkade (Silicon Forest Research) <Sudhir_Kadkade@sifr.com>
Yaron Wolfstal (IBM) <wolfstal@il.ibm.com>
Ghassan Khoory (Synopsys) <gkhoory@synopsys.com>
This meeting was in some ways both a continuation of the previous series of
meetings and a new beginning for this working group. In this meeting, the
ground rules for operating the committee were revisited. Decisions were
made first by consensus and later by vote based on attendees at this
meeting, in part because difficulties in attending previous meetings by
telephone made it difficult to determine who should be considered a voting
member based upon the previous ground rules (in which attendance at the
previous 3 consecutive meetings was required for a participant to be
considered a voting member).
Agenda:
A. Recordkeeping
A.1 Minutes
Harry proposed that we select a volunteer to take minutes of each meeting.
Proposal accepted. Erich to take minutes of this meeting.
A.2 Web Page
Harry proposed that he would ask his son to maintain the web page for the
group. No disagreement. Harry mentioned that he would need details from
Vassilios regarding how to update the web page.
B. Operational Ground Rules
B.1 Scheduling Meetings
Harry proposed that we schedule two meetings in advance. General agreement.
Discussion narrowed down the requirements to:
- hold meetings during Monday-Thursday, not on Friday
- hold meetings (to complete by) no later than 3pm PDT, to facilitate
participation from the east coast, Israel, Munich
- stick to monthly meetings (too difficult to schedule more often), but
encourage email discussion between meetings
- hold meetings at company sites where meeting rooms, teleconferencing
will be available
Arif volunteered to have Synopsys sponsor the next meeting (at ICCAD) by
providing a meeting room. Carl volunteered to have Motorola sponsor a
meeting at FMCAD, but it was later decided not to have a meeting during
FMCAD.
The second meeting will fall in the first week of December. Harry will work
out exact dates and times for both meetings and propose them by email.
B.2 Definition of Quorum
Harry proposed that a quorum be defined as 2/3 of the active (voting)
members. Vassilios explained that a voting member is one who has attended
the last 3 meetings, either in person or by telephone, or who has had a
substitute attend in his/her place. Mike McNamara suggested amending this
to "3 of the last 4 meetings", which is the definition used in the 1364
committee, and there was general agreement. Questions arose about (a) the
status of substitutes who continue attending on their own, and (b) whether a
company should be allowed to have more than 1 vote. Someone objected to the
"1 vote per company" idea, because different groups within a company may
have different positions. Mike pointed out that the "3 out of 4" rule
accomodates having each group in a company send its own representative, yet
it discourages a company from "stacking the committee" to influence the
standard because of the significant resource commitment required. The "3
out of 4" rule was generally accepted as a good definition for determining
who is a voting member.
Erich pointed out that we'll need to start each meeting with a roll call,
including a statement of the role of each attendee in the meeting (i.e.,
representing him/herself, or representing someone else), so that we can
collect the information necessary to determine (a) who are the voting
members in a given meeting, and therefore (b) how many voting members
constitutes a quorum for that meeting.
Further discussion resulted in the proposal that a quorum be defined
differently for "more significant" and "less significant" issues. For more
significant issues, 80% of voting members must cast a vote for the decision
to be valid, and to achieve this high percentage, such ballots will be done
by email. For less significant issues, such as operational decisions made
during meetings, only 50% of voting members must cast a vote for the
decision to be valid. However, any voting member may request that a less
significant question be reballoted as a
more significant question, under the 80%, email procedure. This was
generally agreed to, and later (see below) was approved by ballot.
B.3 How Motions are Made
Harry proposed that motions be made during meetings, by any participating
member. In the discussion, this was expanded to say that advance notice, by
email, of a motion to be made is acceptable - even encouraged, to put the
issue on the agenda, so that people concerned with the motion can make a
point of attending that meeting - but that there should be at least one
meeting where the motion is discussed before any vote is taken. To enforce
this, it was suggested that "a motion can only be seconded in person, at a
meeting, by a voting member", and this was generally accepted. Vassilios
also pointed out that, for major issues, we cannot both introduce a motion
and vote on it in the same meeting, in order to ensure that any member
concerned with the issue have an opportunity to participate in its
discussion and in the vote.
B.4 How Resolutions are Made
This topic led to three more-or-less separate discussions:
- how the Board approves standards, with what criteria, and with what
goals
- how this group will ballot issues, and how votes will be counted to
decide an issue
- what Yaron meant by "resolution", and why the question came up
Discussion 1: the Board. Vassilios made it clear that the Board ratifies a
draft standard among member companies by seeking technical advice to confirm
that it meets the requirements of all member companies (including those who
are not represented in the working group). Carl asked if the final Board
vote on a standard is based on a simple majority; Vassilios said no, he
thought that more than a simple majority is required, but he did not say
what percentage approval is required to approve a standard, nor what the
basis is for that percentage (i.e., how many board members must be involved
in the vote). In any case, it is clear that the Board's role is not like
that of the IEEE Standards Board, which is chartered to evaluate the process
by which a standard is developed rather than the technical details of the
standard.
Discussion 2: How we ballot. We noted that the previous discussion to
define what a quorum would be for deciding lesser and greater issues (50%
and 80% respectively) is independent of the percentage of positive votes
cast by the voting members (constituting the required quorum) on a given
issue. Erich suggested that simple majority would be sufficient for votes
on lesser issues, implicitly assuming that a more significant majority would
be required for more significant issues. Whether this implicit assumption
is reasonable, or valid, was discussed. Vassilios noted that OVI TCC rules
state that, in a deadlock situation, the chair casts the deciding vote.
Discussion 3: What Yaron meant by "resolution", and why. It was suggested
that this question came up as a result of the committee's decision to accept
additional submissions of proposed language definitions - so the question
becomes "what does it mean to make a decision, if we are going to come back
later and change the decision". Mike (?) noted that the decision to revisit
the question was made by committee vote, so it was handled correctly.
Someone suggested that, if reaching consensus requires revisiting some
decisions, it is reasonable to do so. Vassilios argued that our goal is not
to reach consensus, but rather to deliver what we promised (a proposed
standard) on time, and therefore we cannot continually go back and revisit
decisions. Someone countered that our goal must be to promulgate a standard
that people will accept, and that we must not put forth a proposed standard
that is nonsense. Harry suggested that while putting together a good
standard is important, it is also important to have realistic and achievable
milestones - and as Vassilios was saying, to achieve them.
To close out this discussion, we considered the following motion (made by
Erich, seconded by ?) that sums up the discussions above:
motion -
1. A "voting member" is anyone who has attended 3 of the last 4 meetings.
2. A ballot regarding a "less significant" decision requires a quorum
consisting of at least 50% of the voting members (at the time of the vote).
3. A ballot regarding a "more significant" decision must be taken by email
and requires a quorum consisting of at least 80% of the voting members (at
the time of the vote).
4. Except for final approval of the standard, each ballot will be decided
by simple majority.
5. Final approval of the standard will require a quorum of 75% of the
voting members, and a 75%positive vote.
A roll call vote was taken, and this motion passed - 1 no, all others
on-line at the time (approx. 10) - yes.
C. Language Donation Status/Procedure
A brief discussion ensued regarding the challenge of dealing with the legal
issues associated with submitting a proprietary language definition for
consderation as a base for the standard. [This is difficult to summarize,
so I'll simply present the dialogue as I wrote it down. -EM]
Danny - we don't have a legal rep. to deal with our submission - how do we
proceed?
Carl - the board has taken Moto submission to lawyer for review - looking
for resources to enable other contributions
Vassilios - board support will take too long - they only meet every 2
months - just sign the release letter
Carl - but the release letter cannot be signed by most companies
Vassilios - then donating companies should provide funds for legal review
Carl - raise it at the board meeting
Vassilios - done
Vassilios - this committee must agree first on terms to look at a language
submission, before the issue is raised to the board - Carl should send out
submission letter to the committee
Carl - this is a new requirement - we submitted letter to OVI before DAC,
but it has not been brought up in any meeting
Vassilios - any time you go outside the normal procedure, you'll delay the
process
Carl - you are not representing the board's position - I'm on the board, the
board recognizes it needs to change its process, obtain legal representation
??? - doesn't this mean that we, the committee, can propose to the board
that we accept looser restrictions - i.e., that the original letter is too
onerous?
??? - this should be a topic of discussion for the next meeting
??? - some people who have accepted original rules may have objections to
relaxing rules now
Arif - what about deadline?
Harry - We should only look at these four (IBM, Synopsys, Verisity,
Motorola), because those companies complied with deadline.
This archive was generated by hypermail 2b28 : Wed Dec 06 2000 - 07:09:13 PST