Subject: Economics of Free Software
From: Christopher Maeda <cmaeda@ERNST.MACH.CS.CMU.EDU>
Date: Wed, 17 Mar 93 01:33:45 EST

Let's not beat around the bush here.  L. Peter Deutsch seems to be
saying that it is economically irrational for someone to *develop*
free software.

It's not clear that self-funding, mass-market free software is
economically feasible.  It takes lots of time and money to do
development and the kind of bullet-proofing required for a mass market
product.  Let's look at 386BSD.  The Jolitz's and the volunteer
maintainers have done a prodigious amount of work with 386BSD.  But
the Q/A in 386BSD doesn't even begin to compare to a commercial
product.  It's an utter pain in the ass to install.  If you don't
think so, you either got lucky or you haven't done it.  In contrast,
BSDI is a joy.  I've also looked at the kernel sources for both
systems.  You can see the difference in quality between a system where
unpaid amateurs do a lot of the work and a system that generates
enough revenue to pay for experienced and talented people.

It is also not clear how you can fund large development projects in
the free software model.  GCC ports (and maybe programming tools in
general) are a special case since you have large hardware vendors that
want compilers to be available.  But how do you fund a general
application of which everybody will want at most one?  Do you write a
business plan that says you're going to spend hundreds of thousands of
dollars and several man years to develop something and then that
you're going to give it away and start a service business selling
support contracts, cd-roms, and documentation?  How many venture
capitalists will go for that?

As I see it, on the one hand we have this quixotic notion that
software should be free and that programmers will always be able to
earn a living selling their time and that everyone will be happy
because programming is inherently a fun activity.  On the other hand,
we have this harsh reality that it takes a huge amount of time, money,
and hard work to develop high quality software and that, apart from a
small group of true believers, people will be unlikely to commit this
time and money without substantial return on investment.