Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!iuvax!uceng!dmocsny
From: dmocsny@uceng.UC.EDU (daniel mocsny)
Newsgroups: comp.unix.i386
Subject: Re: Purpose of .i386 newsgroup
Summary: Bugs find me
Keywords: purpose basis
Message-ID: <2404@uceng.UC.EDU>
Date: 11 Oct 89 00:30:15 GMT
References: <385@csense.UUCP> <16165@vail.ICO.ISC.COM>
Distribution: na
Organization: Univ. of Cincinnati, College of Engg.
Lines: 64

In article <16165@vail.ICO.ISC.COM>, rcd@ico.ISC.COM (Dick Dunn) writes:
> Also, Dan Mocsny wrote:
> 
> > Maybe the other ones simply work as advertised, so nobody has anything
> > to say about them? :-)
> 
> We'll get you for that, Dan!  Quick, everybody from *.isc.com over to
> rec.bicycles!  Start posting pro-automobile articles...make them real
> flames, and put "Followup-To: rec.auto"!

OK, I'm ready this time. I've been refining my arguments, and I'm much
more used to being flamed now. C'mon, Gaswasters!  Let's see what you
can do! :-)

> Seriously, I think that one of the things that makes for more ISC-related
> discussion is that, unlike AT&T or IBM, we don't sell hardware.  That
> probably results in more questions about "how do I make this magic card
> work?" or "will this card work?"

Also, I seem to have a magnetic attraction for computer trouble. Things
happen to me that just don't happen to other people. Like my local
hardware vendor who has been helping get 386/ix running says: "Believe
me, Dan, it's not really as hard as this!" Maybe I should work for
ISC quality control?

But I do have to believe that the entire computer industry needs to
beef up its inter-vendor communication. Vendors must never forget that
success is giving value to the customer, not just dumping crates on a
loading dock. If a board manufacturer wants to sell into a market,
they should insure that their boards are going to work in that market.
If an OS vendor is selling into a market, they should insure that they
support the hardware that exists in that market. If they don't support
everything, then they should make perfectly clear what they do support
*before* the customer buys.

In the computer industry no single vendor supplies all of the value of
its product. Hardware is only valuable because it runs software,
add-in boards both get value from an existing bus and add value to it,
and so on. To a very large extent, the value a vendor adds is not just
from the technical excellence of its product, but also from the effort
the vendor has put into working out the details with other vendors so
the customer does not have to.

Now historically, in a market economy vendors view themselves as
competing with each other on the basis of how their products "perform."
They usually express "performance" with some raw measure that may
have little or no relation to how the customer perceives value. For
many customers, the only valid benchmark is what I call "Loading
Dock Time," that is the time that elapses between the box landing
on the loading dock and useful work finally coming out of it. The
traditional measure of "performance" has something to do with this,
but much, much more is involved.

A box, an OS, a board, or an application that gets most of its
value from an existing market must work in that market. If the
vendor cannot cope with testing the astronomical number of
product combinations users are likely to get into, then all the
vendors must commit part of their resources toward getting
together and reducing the chaos they have created. If they do
this, they will give more value to the customers, who will then
in turn have more wealth to spend with the vendors.

Dan Mocsny
dmocsny@uceng.uc.edu