RE: Value Message from PCI-SIG Chairman

[Dave New <NewD@esi.com>]

Thanks, Duane for very nicely putting together
exactly what I was thinking as I read through
the really informative piece that Roger has
posted.

First of all, I'm glad to see that Roger has
been listening to the recent conversation
unfold, and has felt the need to lay out the
SIG's case for its current membership fee
structure.  Very well thought out, and
it does make a point, particularly for those
companies that can afford the manpower and
expense to send folks to training, conferences,
and plugfests.

Duane, however, makes some really good points
for 'the little guy'.  Even though he talks
mostly about small companies, my concept of
a small company runs about oh, two or three
folks, sometimes just one.  At this point,
what you have is an entrepreneur/cottage industry
person, someone who is likely to read Circuit
Cellar or similar journals for ideas and/or
tinkering.  Students would be included in this
group, as well.  Essentially, lots of enthusiasm,
but little or no funds for membership fees,
training or travel.

Because of the plethora of PCI chipsets and the
decreasing cost of software and hardware tools
for programming the parts and laying out the boards,
folks on shoestring budgets will soon be able to
turn out PCI-compliant boards.  These boards
will attack various niche markets (or no market
at all), like specialty data acquisition or
controller boards for laboratory robotics, etc.

With the commoditization of PCI we are about
where we were 15 years ago with the ISA bus --
accessible, and within reach of a reasonably
outfitted laboratory.

25 years ago, when the microcomputer industry
was very young, Motorola captured the imagination
(and eventually the mindshare and pocket books)
of a whole generation of college students, by
making databooks and sample chipsets and evaluation
boards free or almost free to colleges.  I  was
one of those folks that parlayed that initial
exposure to the MC6800 (and its peripheral chips
and evaluation board) into first a senior project,
and then a thesis topic (Microprocessor Control of
an Amateur Radio Repeater Station).  At the time,
this was considered quite advanced subject matter
for tinkerers and students.

See where we are now...

Cheers,

-- DaveN

-----Original Message-----
From: Duane Clark [mailto:dclark@akamail.com]
Sent: Saturday, October 28, 2000 1:33 PM
To: Roger Tipley; pci-Sig
Subject: Re: Value Message from PCI-SIG Chairman


Roger Tipley wrote:
> 
> With PCI evolving to include higher speeds, more
> efficient modes of operation, and new form factors,
> the PCI-SIG is at the forefront of leading the
> industry to keep PCI viable, and targeted at 21st
> century products.  With 100s of millions of PCI
> products sold each year, it behooves us all to keep
> the standard vibrant, alive and interoperable.

I completely agree. And that it why I think that PCI-SIG should
carefully consider the issues being raised here. In the past, the high
cost of entry kept all but big deep pocket players out of the process.
But the inevitable advance of technology has changed the equation so
that building truly compliant PCI boards is well within the reach of
even a college student on a limited budget, and PCI-SIG needs to
consider the implications of that.
 
> To fund SIG programs, your company pays an annual
> dues, and for that dues receives several rights,
> protections and benefits.  80% of your company's $3000
> annual membership fee in 2001 goes into programs that
> you as members see as direct, tangible benefits.  The
> remaining amount funds programs that enable the PCI
> standard to continue to be an industry success.

Oops, we've been using a number of $2500. Sigh...

Let me try to lay out the issues involved as I see them, from the point
of view of a very small company. All the issues raised in the rest of
the letter make sense if you are building 30,000 boards or chips or
whatever per year that are going to be sold widely. The membership fees
add a measly $0.10 or less to the cost of the board, and the value
obtained in relation may indeed be high.

But I am a little guy in a niche market, who is lucky to sell 30 boards
per year (and I am not interested in making that 30,000). So for $100 or
more per board:

Benefit 1: Fortunately, the spec is pretty well written. The Znyx forum
seems to work very well on clarifying minor points, and getting info on
real world implementations. I am definitely very grateful to them.

Benefit 2: I would have to carefully consider whether I could justify
the expense of going to a plugfest. Yes, it may save a large company
millions of dollars, but that certainly does not apply to me. I deal
with every single problem promptly on an individual basis.

Benefit 3: It is the nature of one person companies that the person
involved is unlikely to want this sort of training. We tend to be
aggressive self starters and self learners.

Benefit 4: I really don't need a unique vendor ID. I would be quite
happy to share one that was officially registered as "Brand X". And by
the way, I would be quite willing to pay a reasonable fee for this. But
nowhere near $3000.

Benefit 5: An official PCI logo does not add value in the kind of market
I am in.

Benefit 6: I am not interested in cutting edge early access. I am
satisfied with established standards, and am willing to wait for
official releases of new standards. And I will make every effort to
comply with those standards, as long as major roadblocks are not put in
the way.

I consider the $3000 to be a major roadblock, and a very large cost to a
small company. I think that PCI-SIG needs to recognize that times have
changed and they need to consider that small companies are going to put
out PCI boards and products, and it is in the interest of the PCI-SIG to
retain control of the process. The way to retain control is to put in
place a mechanism and cost that makes sense to our companies.

Duane
Leeward Engineering