Subject: RE: the GCC steering committee
From: "Larry Augustin" <lma@vasoftware.com>
Date: Sat, 16 Mar 2002 09:10:43 -0800

Under the conditions where:

	- A substantial majority of the core group of developers that
controls a project works for one company
	- Forking is a very difficult and painful process (perhaps
because the project is a large complex project needing a lot of in-depth
expertise)

There is an opportunity for a FSB model based on "integration
extortion."  This is where the controlling company refuses to integrate
patches into the main software distribution unless they are paid to do
so.

"I'm sorry, but we just haven't had time to integrate the patches to
support the new instructions in your chip.  Because it's such a pain to
build and maintain a fork, the distributions are unlikely to ship a
patched version, and so support won't be in the distributions.  Also,
we're about to issue a new release and your patches won't apply cleanly
once we've issued that release, making it even harder for other people
to use your patches.  However, if you paid us money, we could probably
find the time to integrate your patches."

Under this model the further "upstream" your FSB is in the process, the
more leverage you have.  For example, the company controlling the
distribution could still use this model even if they don't control the
core development of the project:

"I'm sorry, we just haven't had time to get that latest release of that
package into the distribution.  But if you paid us money, I'm sure we
could get the new version into the next release.  We might even be able
to get an early beta of that package into the distribution."

Larry