Subject: selling large (dull) development on free software
From: gnu@CYGNUS.COM
Date: Thu, 11 Feb 93 12:48:17 -0800

One thing about a free software business is that you have to run it
like a business.  I've been a consultant for close to twenty years,
off and on, and in a one-man consulting business you can get away with
looser operation because much of your overhead is covered by your clients,
and pricing is very negotiable.

But a one-"man" retail or support operation had better know how to
be a salesperson and marketeer as well as financier and hacker.

Brian Marick said:
> 					 People are clamoring for C++
> and ports to DOS.  Both of those would be a fair amount of work,
> pretty dull work.  The clamoring users will mostly not chip in to pay
> for it, though they all wish that others would.  (This is the "tragedy
> of the commons" as applied to freeware development.)

This is part of the "sales" function of a business.  Cygnus has put
together several deals in which no single party is willing to pay for
a large piece of development to happen, but all of them stand to 
benefit from it.  They realize that unless they chip in, it won't happen.
Or worse, it will eventually happen, but in a way that doesn't meet
their requirements (e.g. a DOS port that doesn't support Windows).

Seeing the need in the market is not too hard for anyone.  But pulling
together a multiway deal is hard work even for a good sales person.  Still,
it sometimes works, and this is one way that large developments can be

An example is the Solaris port of GCC and GDB.  We knew it would be a
lot of work, and we advertised for supporters who'd get a year of
service at a good price, but who would have to trust us that we'd
really deliver.  About 40 people and companies signed up at
$2000/each, and that was enough money for us to do the port.  (We'd
hoped to get more people, and had worked out a deal to donate a
percentage above a certain amount to FSF, but didn't get enough to
trigger that donation.)

We've also done other multi way deals to add C++ features, or to port
tools to a new embedded platform (e.g. splitting the cost between the
manufacturer of the chip, and one or more of their customers).  It's
easier to find two or three (rather than 40) people who'll benefit,
and they can often be sold on paying 1/2 or 1/3 of the cost because it's
less money than doing it themselves and worth it due to the benefit
to their own business.

I wish I could give you-all a magic formula for making it possible to sell
new development in free software -- because I want there to be a *lot* of
free software businesses!  But it's work like everything else...

	John Gilmore
	Cygnus Support

PS:  Take heart -- just surviving your first year in business means you
are doing better than 90% of the people who try.  I can't say that it gets
easier from there, but at least the challenges are different.