RE: [vhdl-200x] 1076 & Entity balloting

Jayaram,

I think that works in some cases. It would not have worked for VHPI nor
likely to work vhdl-200x.

Besides the skill set needed there are time pressures for the volunteers.
In most cases, they are too busy creating the standard with their volunteer
time. Dennis made a good observation that perhaps DASC is not the best
place for advancing the state of the art involved in creating a leading-edge
standard ( which is an oxymoron ), but that is what is happening.

Finding a volunteer to do the work has not been possible in recent years.
Finding willing tech writers who have the skills and available time to edit
a large HDL LRM with significant new content is not easy. Funding them
helps set a timetable for progress of the work that is much more aggressive
than otherwise possible with volunteers. VHDL-200x is right to seek funding
for this work rather than rely on volunteers. It will turn out to be a
referendum on whether the industry and its users will get behind seeing
vhdl-200x rapidly progress or languish with respect to SystemVerilog. Maybe
the funding won't arrive but the donation of competent resources to do the
job will.

-John

-----Original Message-----
From: owner-vhdl-200x@eda.org [mailto:owner-vhdl-200x@eda.org] On Behalf Of
Jayaram Bhasker
Sent: Friday, June 25, 2004 10:25 AM
To: Bailey, Stephen; vhdl-200x@eda.org
Subject: RE: [vhdl-200x] 1076 & Entity balloting

There is a 5th choice (other than the 4 listed by Steve), which many of the
other WGs have
used:

5) A volunteer editor.

regards,
- bhasker

-----Original Message-----
From: Bailey, Stephen [mailto:SBailey@model.com]
Sent: Thursday, June 24, 2004 10:55 AM
To: vhdl-200x@eda.org
Subject: RE: [vhdl-200x] 1076 & Entity balloting

Hi John,

First, Accellera has provided funding for 1076, specifically VHPI work. (I
know you are aware of this, but I need to make sure all are aware.) It is
my understanding that the VHPI funding has come from contributions by EDA
vendors that are Accellera members.

However, the nature of Accellera membership is that it is currently biased
towards Verilog using companies on the end user side of their membership.
Therefore, some how, some way, we need to get VHDL using companies to step
up to what should be an activity they are interested in.

On the Verilog side, they have used Accellera effectively and self-taxed the
development of SystemVerilog. (I am told at least $150k has been spent on
getting SV to the 3.1a LRM level.) On the VHDL side, we have tried to do
everything within the IEEE and beg for money from Accellera primarily for
LRM editing work (the technical proposals being developed by volunteers).

So, our choices come down to:

1. Pull VHDL-200x work out of the IEEE domain for the time being. Move the
work under Accellera. Solicit new Accellera membership and funding of the
VHDL work under Accellera. Once we get the LRM written, hand it off to IEEE
DASC.

2. Keep VHDL-200x work in the IEEE. Ask Accellera to fund it. Which would
still require me (or someone) to solicit new Accellera membership and
funding of the VHDL work. (EDA vendor contributions are unlikely to be
sufficient to cover the cost.) Here, Accellera collects and provides the
funds. But the Accellera membership has very limited influence as long as
the WG is individual membership based: a binary decision to fund or not.

3. Keep VHDL-200x work in the IEEE. Solicit funding of the VHDL work and
channel that funding through the IEEE SA. With individual membership, the
companies have no ability to directly influence the work (except the binary
fund/not-fund decision). With organizational entity membership, they get a
vote in the WG and on the ballot.

(For me, the use of Accellera or IEEE as the funding vessel is a secondary
consideration. It doesn't matter as long as there are no funds.)

4. Hope some affluent angle will drop from the sky and provide us with the
$200k that is the estimated cost for LRM editing of 2 revisions under
VHDL-200x. (I'm personally not counting on this possibility. ;-)

It seems the common point of skepticism is whether a change in membership
will increase our likelihood of funding. Here, a couple of role-playing
exercises should be very illuminating:

First, I ask everyone to pretend like they are the decision maker at a
company like Nokia, ST, Rockwell Collins, Xilinx, etc. Would you be more
likely or less likely to provide funding to either Accellera or IEEE if you
have no vote (control/influence) on what that money is used to produce or if
you do have a vote?

Now, pretend you are me and you are making a presentation to these
organizations whose PATRONAGE IS REQUIRED FOR THE SUCCESS of VHDL-200x. Do
you think you are more likely to be successful in gaining the funding you
require with or without the ability to offer them the benefit of a strong
influence on what their money is buying?

We are selling something. How do we ensure it is packaged as attractively
as possible?

I can start soliciting companies at any time. Since I don't like wasting my
time, I prefer to do so from the strongest position. That's why I want
organizational entity membership approved first.

No one wants to give up a benefit that they currently have. I understand
and sympathize. Peter and I have pointed out ways we can keep strong
individual participation and influence (where it really counts) not just to
help ease the transition, but because we want to keep and encourage
individual participation.

I'm hoping that everyone is keeping their focus on the priority: The future
of VHDL. Also, I hope that everyone recognizes that without $'s there is no
future. Hundreds of pages of new or revised LRM text does not grow on
trees!

-Steve Bailey

> -----Original Message-----
> From: John J. Shields [mailto:jshields@ieee.org]
> Sent: Wednesday, June 23, 2004 6:34 PM
> To: Bailey, Stephen; vhdl-200x@eda.org
> Subject: RE: [vhdl-200x] 1076 & Entity balloting
>
>
> Steve,
>
> I've read through this discussion. I am more open to the
> idea of entity membership, but remain skeptical of its
> overall effect. There are serious concerns raised and the
> funding reality is compelling. I would be interested in using
> our liason to Accellera to get the viewpoint of the member
> corporations who are funding Accellera and presumably would
> provide significant financial support to the DASC.
>
> Is it important to Accellera members to have entity-based
> membership over vhdl-200x effort and presumably to extend
> this model to all future DASC efforts?
>
> Is it likely that Accellera member companies will scale up
> their investment to meet DASC annual needs in addition to
> Accellera's own needs or tradeoff their standards $ (e.g.,
> earmark substantial Accellera budget for DASC, let DASC starve)?
>
> Would an entity controlled and funded DASC obviate much of
> the need for Accellera to exist? Scale it back in any way?
>
> As a follow-up to that question, what roles do the member
> companies wish to promote as appropriate for Accellera vs DASC?
>
> I don't want to see an inappropriate discussion. It is these
> entities and their money we seek and for which we are
> proposing controversial change.
> Can we be direct and ask the players to discuss what we are
> considering and give us some meaningful feedback?
>
> If pressed to vote now, I'd vote against and seek the
> political groundwork for improving our funding to be carried
> further first.
>
> Regards,
> John Shields
>
>
Received on Fri Jun 25 14:48:11 2004