*BSD News Article 64304


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!news.rmit.EDU.AU!news.unimelb.EDU.AU!munnari.OZ.AU!spool.mu.edu!pravda.aa.msen.com!nntp.coast.net!howland.reston.ans.net!swrinde!cssun.mathcs.emory.edu!cc.gatech.edu!cc.gatech.edu!byron
From: byron@cc.gatech.edu (Byron A Jeff)
Newsgroups: comp.os.linux.development.apps,comp.os.linux.development.system,comp.os.linux.x,comp.os.linux.hardware,comp.os.linux.setup,comp.unix.bsd.386bsd.misc,comp.unix.bsd.bsdi.misc,comp.unix.bsd.netbsd.misc,comp.unix.bsd.freebsd.misc
Subject: Re: Why to not buy Matrox Millennium
Date: 28 Mar 1996 06:08:00 GMT
Organization: Georgia Institute of Technology - College of Computing
Lines: 119
Message-ID: <4jdac1$edc@solaria.cc.gatech.edu>
References: <4j21ph$crr@slappy.cs.utexas.edu> <slrn4lbb5r.ai.mark@hunter.mas.org>
NNTP-Posting-Host: gemini.cc.gatech.edu
NNTP-Posting-User: byron
Xref: euryale.cc.adfa.oz.au comp.os.linux.development.apps:13674 comp.os.linux.development.system:20011 comp.os.linux.x:27622 comp.os.linux.hardware:34369 comp.os.linux.setup:47245 comp.unix.bsd.386bsd.misc:294 comp.unix.bsd.bsdi.misc:2787 comp.unix.bsd.netbsd.misc:2563 comp.unix.bsd.freebsd.misc:16024

In article <slrn4lbb5r.ai.mark@hunter.mas.org>,
Mark Swanson <mark@hunter.mas.org> wrote:
-On 23 Mar 1996 17:34:09 -0600, Peter F. McDermott <pmcdermo@cs.utexas.edu> wrote:
->Some companies just don't get it.  Valdimir Vukicevic on the XFree86-3D list 
->requested information on Matrox's card (first message) and received the
->second message as a reply.  Unless Matrox changes their mind on release of 
->information, I suggest no one interested in running a free Unix system 
->buy their products. 
->
-
-I think their policy is just fine. One must respect a companies right to
-protect intellectual property.  Their attitude is snotty and arrogant, but
-their policy shouldn't be too surprising. I should think that we (the Linux
-community) have the intelligence to route around this. Why can't the XFree86
-team just distribute the binaries for certain cards? Do what you can for free
-software but living or dying by it results in no support for this type of
-hardware. There will ALWAYS be this type of hardware around that Linux
-users would like to use.

I think the real problem is that the policy doesn't make much sense.

Consider Matrox, Diamond, Connectix, Xircom, Visioneer, et. al. All these
folks sell some type of hardware along with software so the hardware can
be used. 

Now comes along another market segment which wants to use the hardware but
have no use for the current software.  The company has three issues to
consider:

1) Does the company want to invest money and energy developing and supporting
this new segment?

2) Will the release of the information to develop other software compromise
any trade secrets the company uses to do business?

3) Is there a company objective to either make money off any software
developed for the hardware or to prevent others from making money from
software developed for the hardware?

Now interestingly enough most company policies focus on point #2. I quote
from the original Matrox message.

->We have a closed architecture chip set. We have the fastest
->board on the market. The reason for this is our chip set. We did not create
->our own chip set, in which we spent millions of dollars, so that our
->competitors can steal our designs. 

I fail to understand how releasing an interface specification for a piece
of hardware in any way enables a competitor to steal designs for the
underlaying hardware. It reeks of a half cocked argument. Explain how
Diamond or #9 will be able to steal and manufacture Matrox's chip from a
list of registers along with the values that make the chip do accelerated 3D?
I can't see it. I'm willing to listen to any arguments to the contrary.

So at every turn we should point out that fallacy of this line of attack. 

Now point #1 is much more grounded in reality and present somewhat
sticky problems for a company. It is difficult if not impossible to 
support software that isn't generated in-house. Again to quote the Matrox
response.

->We have nothing to do with unix/linux. We do not
->create the drivers, we do not support the drivers, most of us have never
->even seen unix/linux let alone the drivers. If you have any problems with
->the drivers, contact X-inside. They do the Drivers, Support.

This is quite prudent. The Connectix policy of making developers point out
that Connectix has nothing to do with non-Connectix drivers is important.
When things don't work folks tend to phone or E-mail customer support for the
hardware.  A lot of resources could be burned up trying to support something 
the company knows nothing about.

Lastly point #3. A back of the envelope analysis would point out if the
company really wanted to make money on the software that they would either
sell it separately or actively develop said software for all market segments.
They don't do either because the software isn't a real priority, the hardware 
is.

Now as for others selling software for the hardware it can be an interesting
call for the company because the hardware must still be bought thereby
making more profit for the hardware company. OTOH The company could possibly
acquire more profit by selling the software themselves, but then you're
right back at point #1: obligation to support. Like I said, a tough call...

Bottom line. Dismiss #2 because it's nearly impossible to reverse engineer
hardware from a software driver. More software drivers generate more hardware
sales but also generate more support calls. So encourage more third party
drivers (both commercial and freeware) but don't support them. So every
hardware company should have publicly available interface specifications for
their hardware.

So in my book Diamond and Connectix have the winning strategy while Matrox
and Xircom are losers. Diamond and Connectix are both selling more hardware
to a new market segment while having minimal support impact. Xircom is
losing out to folks like D-Link and At-Lan-tec in the parallel Ethernet
market while Matrox is in the middle of this minor fracus.

Has anyone else noticed how quickly Diamond turned from Bad guy to good guy
simply by releasing their interface specs?

These hardware companies should rethink their policies. I just don't think
it's good business to turn a customer away unless supporting that customer
costs quite a bit more than what you'll get from him/her. Satisfied
customers tend to tell others about good treatment while miffed customers
tend to disparage whoever treats them badly. I'm happy with any company
that says: "The interface specs are on our website, we will not write a
driver nor will we support any driver that we did not write in house. You're
on your own kid."

That's the way every video card, parallel ethernet, parallel CDROM/TAPE,
scanner, video capture, and other hardware manufacturer should be.

But alas it isn't a perfect world is it?

BAJ
-- 
Another random extraction from the mental bit stream of...
Byron A. Jeff - PhD student operating in parallel - And Using Linux!
Georgia Tech, Atlanta GA 30332   Internet: byron@cc.gatech.edu