Subject: Re: My customers
From: Mike Stump <mrs@cygnus.com>
Date: Thu, 27 Oct 1994 16:00:53 -0700

In article <m0qevhq-000I8SC.cygnus.fsb@crynwr.com> you write:
>People being people, they try to get someone else to pay for its
>development.

Yes, this is true of free-software.  Someone that doesn't understand
this, and the ramifications of it, won't last in the free software
business.

>The best leverage point is the Ethernet board manufacturer.  When an
>end user goes out to purchase such a board, he can choose from a very
>large number of manufacturers.  And some of these end users require a
>packet driver of the manufacturer (e.g. Rutgers University).
>
>A single large order of adapters can pay for development of a packet
>driver.  And from the manufacturer's point of view, they might not
>make any money on that deal, but they've got the repeat business from
>that customer, and they're all set for the next packet driver
>customer.

If this is how your doing business, than your market is not yet
mature.  The right way to do it, is to hit up the manufacture of the
chipset, by using the leverage of a board maker that is trying to
decide which chipset to use, for the development costs and
maintenance.  Your working software is a feature of the chipset, not
really the board (it sounds like).

This is what Cygnus does to some extent...

Hum, this is interesting...  I think the best places to expand
free-software market may in fact be places in which the software
enables hardware.  The way to get money, is to get a good following
with some hardware, to create a situation, in which the competitors of
the hardware in question, see the value of the enabling that your
software does for the hardware, and then gives you a small percent of
their income to enable their hardware.  They can afford the small
initial investment, and if they don't pony up the money, they will
financially be hurt.

Cygnus gets a bit of money this way, sounds like Crynwr gets money
this way.  Hum, X kinda works this way, and the person that was
talking about boards for Linux match also.

A software only application, has no hardware it specifically enables,
other than the basic machine.  One cannot get money (or can they? see
NeXT->{Sun,HP}, Apple->IBM, Linus->DEC) for enabling a basic machine,
in general.

One might think that A good avenue for basic applications software may
be to convince CD-ROM publishers to divert 1-5% of their income, to a
product or two, to create something that distinguishes their CD-ROM
from all the other CD-ROMs out there, but this seems doomed to
failure, if it is really hot software, everyone would include it on
their CD-ROM next pressing season.  One way around the general
problem, may be through customer loyalty to the LABEL on the CD-ROM.
If a business can gain loyalty to the label, even if it is higher in
price, then they could fund development on new software, inorder to
keep customer loyalty and retain the business.  This could be
efficient, as the CD-ROM manufacture can guess what the masses want
(or it can ask them fairly directly), and then pay a team to do the
software that is most wanted.  It is self-throttling, as if the
company guesses wrong, they loose money.  If they guess right, demand
will go up, profits will rise, and they can spend even more money on
funding software.

But, until there is a massive CD-ROM industry and it matures (large
companies going out of business), we cannot hope to see much money
from them.  Adam, what is your take on this?  Do you see any holes in
my thinking?