Subject: Re: Opportunity lost? Challenge declined!?
From: Ian Lance Taylor <>
Date: 10 May 2001 00:44:30 -0700 writes:

> The thread begun by Adam Theo represented both an opportunity and a challenge
> to this group.  One that so far has been squandered, yet may still be redeemed.

I was clearly in a slightly grumpy mode earlier today.  I'm going to
write a slightly different reply.

> Specifically, a developer has come to us with a product and a question on
> semi-proprietary licensing.  Rather than use this as an opportunity to brain-storm
> and develop a viable means for this person(and the community at large) to
> see a return on their investment(development time) we have squandered it
> shouting what amounts to opinions, rather than arguments, that closed source
> is bad.

Adam asked for a license which ``recognizes the ownership and control
of the developer (like traditional proprietary licenses).''  He
``recently developed a couple of programs.''  I think he got more or
less fair and reasonable responses, pointing to a couple of
possibilities.  I don't think people shouted at him, except maybe once
or twice.

You're quite right in that we could have asked him what the programs
were, and why he wanted a license which retained developer control.
We didn't.  We just answered his question.

Adam, if you are still reading, do you want to describe further what
you are doing and what your goals are?

> Here are the challenges:
> 1.  To brainstorm for solutions to allow developers to be compensated for
> their development work.

There are many ways, ranging from consulting and support to providing
a service to selling something else which is enabled by the software.

> 2.  To convince Mr. Theo that the among those solutions exists a business
> plan which provides an expected return on investment at least equal to his
> proprietary license model.

We would need more information for this.

> Along those lines, I shall begin with a discussion of the Red Hat model.
> Red Hat is essentially a support provider.  You pay Red Hat a fee and they
> help you set-up your Linux system.

Red Hat does a number of different things.  Their recent 10-K form
shows where they made their money last year.  44% of their revenue
came from subscription (sale of open source software solutions;
technical support contracts; and web advertising revenue) and 55% came
from services (customer training and education; software compiling,
debugging and optimization contracts; network infrastructure
consulting services; and engineering services).  Clearly they do not
make a majority of their money helping people set up their Linux

> Unfortunately, there are several problems with the Red Hat solution:
> 1.  The software it is supporting already had strong grass-roots momentum,
> and a non-trivial userbase when Red Hat began.

Yes, in particular, they did not develop Linux.  In fact, before Red
Hat purchased Cygnus, they had relatively few developers.  Even today,
they spend some $49 million on sales and marketing while they spend
some $16 million on research and development (G&A is some $30
million).  Red Hat does employ a number of developers.  But they are
not predominantly a developer organization.

> 2.  I am unsure of the return Linus Torvalds has seen from open-sourcing
> Linux, but it does not seem to be on the scale of Bill Gates, Larry Ellison,
> or Scott McNealy.  Granted, pure monetary success is not neccessarily the
> only metric, however, most developers(and people in general) want to see
> a return beyond the feel good of altruism for their efforts.

None of the three people you mention made their money because they are
developers.  As far as I know, Ellison and McNealy are not developers
at all.  So comparing them to Linus is not to the point.  If Linus
starts his own company, then we can reasonably compare them.  (Note
that Linus is not affiliated with Red Hat.)

A better question would be whether Linus makes a good return compared
to other successful developers.  I'm sure that Linus makes a
competitive salary at Transmeta, and I expect that he has a decent
number of shares of stock which he could sell, or has sold, for a
price significantly higher than he paid for them.  I would be
surprised if he is not a millionaire, at least on paper.  Being a
millionaire isn't what it used to be, but it's a darn site better than
most people in the U.S.  Transmeta, of course, is not really a free
software company, but so what?  They aren't a proprietary software
company either.  And Linus got his job there because of Linux.

> 3.  As a developer, what I am good at is developing software, not providing
> support.

As you can see from the above, Red Hat does not make a majority of
their money from support.  In any case, I think that what you are
really saying is that you would need to be part of a larger
organization.  The number of developers who work by themselves, just
do development, and make as much money as, say, Linus Torvalds, is
really rather small.

> Which brings me to a more focused question than my earlier challenges:
> How does a developer that wants to invest his time developing, get compensated
> in the open source arena?

He or she goes to work for somebody else.  Red Hat might be a good
choice.  Or how about IBM?  I believe they have more people working on
free software than any other company.

> I will add two constraints to this question:
> 1.  The income level must be comparable to other opportunities in the market
> place.

Red Hat and IBM pay competitive salaries.

> 2.  The individual does not need to be a world renowned expert on the topic/software.

Red Hat and IBM require skill, but certainly do not require world
renowned experts.