Minutes of conference call, Thursday March 4, 1:00PM - 2:30PM

Attendees:

Mike Andrews		Mentor Graphics
Arun Balakrishnan	NEC
Tim Ehrler		VLSI Technology
Ted Elkind		Cadence
Darshan Rauniyar	Mentor Graphics
Wolfgang Roethig	NEC


* Discussion on future direction with physical library modeling

The ALF group is partially aware on the activities launched by OVI
for physical library source standardization and by SI2 for physical
library API standardization. We want to determine our level of interest
in being involved.

A.I. Gather opinions and discuss the matter in the ALF meeting March 29. -- all


* Procedure of finalizing and submitting spec.

The plan is to incorporate remaining incremental changes to the
current version by Wednesday, March 10, to stage that version as
ALF 1.0.11 and to submit it to the PS-TSC chairman of the OVI board.
We expect feedback from OVI by March 29. Between March 10 and March 29
we can still review ALF 1.0.11, but keep the document unchanged.
If the OVI board has requests for changes and approves our requests
for changes as well, all changes can be made in one shot before the
HLD conference April 6.
It is up to OVI to approve the resulting version 1.0.12 officially
as ALF 1.1 or to request more work before approval.
OVI has the opportunity to announce the status of ALF at the HLD
conference.


* Review of ALF 1.0.11

Chapter 1.3	more about RTL design planning

Design planning is an important topic, and ALF has related features.
This should be mentioned in the introduction. Since design planning
itself covers many possible applications, the specifics of ALF support
have to be stated carefully.  A formulation is already proposed in
the spec.

A.I. come up with appropriate statements -- all

A.I. incorporate statements in spec. -- Wolfgang

Chapter 2.3	new subchapter 2.3.3 on signal integrity

Chapter 3.4	changes according to minutes of March 1st

Chapter 3.6	changes according to minutes of March 1st

Chapter 3.12:
The introductory statements about vector expressions should be amended
to avoid the misunderstanding that application tools would always need
to postprocess an event report from simulation in order to evaluate
vector expressions.

A.I. Incorporate statement about tools reading event reports versus
simulator generating event reports. Functional models must respect causality.
-- Wolfgang

A.I. Fix issues with 3.12 reported by Arun -- Wolfgang

A.I. Fix issues from Tim Ehrler's email, related to MUX primitive -- Wolfgang

A.I. Fix the following details -- Wolfgang

	3.12.1	1st sentence change "context" to "concept"

	table 3-3 lexical "token"

	3.12.1 mention case of zero-delay
		v ariable

	3.12.3	scope = variables of event report

	3.12.4 change <?> to <&> 
		binding priority of -> and & is the same only
		for vector_event_sequence, not for general vector_expression

figure 1-1
	change "model compiler" to "model generator"
	function model goes also to timing tool,
	timing model goes also into simulator
	connect scan and synthesis tool
	point characterization clearly to timing and power
	add signal integrity
	add universal constraints

	2.3.3 change "has a lot to do" to "is closely related to"

	3.6.3.17 change "ITAL" to "VITAL"

	3.12.8	typos:	patttern
			expliciely