Subject: Re: Thoughts on GPL
From: John Gilmore <>
Date: Tue, 03 Mar 1998 00:01:58 -0800

>		A better way to look at this is perhaps that Mike got
> Stanford (grad stipend) and Penn (parents tuition, if you like) to
> support his work on GCC, which he later converted into a hidden
> investment in what would later become Cygnus.  Simultaneously, John
> Gilmore invested his own time in GDB.  The $20k number is therefore
> more than a bit fuzzy.
> Aside: people argue that "labors of love" are not investment.  That is
> both inaccurate and absurd.

A "hidden" investment in free software is there for anyone and
everyone to use.  Many people can benefit from it simultaneously
(generally benefiting from each others' efforts as well).  It's like
society's investment in science, ideas, and concepts; and like the
physical infrastructures each generation inherits.

The time I'd invested in GDB before forming Cygnus was pretty minor;
I'd worked on a port or two.  The time I spent on GDB at Cygnus was
salaried time, not a labor of love.

Mike put in a lot of work on GCC and G++ before Cygnus.  Certainly his
time at MCC (where he wrote G++) and at Sun (where he made it
reliable) was all fully funded, salaried work.  It would defame Mike
to say it wasn't a labor of love; he was a ball of capable energy
focused on making it happen.  But it's not like he did it as an unpaid
hobby; he used his substantial persuasive skills to convince his
employers to pay him to do it (for their own benefit).  The process of
starting Cygnus was in some ways the process of Mike realizing that he
could probably get companies to pay him *simultaneously*, rather than
serially, to write or improve a piece of free software.

>    ... If they'd had to develop gcc from scratch to sell it,
>    do you believe they could they have started the company on $20K?

I know of other free products that are being developed from scratch,
paid for up front by customers.  They'll be talked about more when
first released.  Keith, you're right in the small sense that the initial
marketing/sales/mgmt/engineering expense to just get those deals
closed was far more than $20K.  But wrong in the larger sense that 
"writing free software just can't pay for itself".

Maybe what the Cygnus experience says is that with "$20K and three
talented guys" you can start a free software company.  (Like all
senior managers, our first job was to bring in enough business to pay
ourselves.  We did, so didn't suffer financially.)  I'd guess that with
such resources you can probably start many kinds of company.

> What I realized today is that there is one kind of market condition
> under which a free software vendor can enter and dominate: any
> established market lacking a single dominant player (or a pair of
> them).

I think I agree.  The embedded cross-development market was one of
these.  I saw another in accounting software, which happened in the UK
in the mid '90s.  No company had more than about 5% of the market.  A
company that had heard of Cygnus got in touch, hoping to grow their
market share significantly by freeing a large central core of their
software (retaining some add-on modules as propriatary).  Then it
would slip into use in many sites, as mentioned by Jonathan, and some
of them would become new customers.  Cygnus got as far as getting
their source code under NDA for examination.  Through lack of
bandwidth to handle new products, particularly in a market we had
limited in-house expertise in, it never happened.  It would have been
an interesting test to be able to see the results of now.  (Is the UK
accounting software market still this way?...)