Subject: Re: FW: Why would I pay for Ximian software?
From: Ian Lance Taylor <ian@airs.com>
Date: 03 Jan 2002 13:42:32 -0800

"Perry E. Metzger" <perry@wasabisystems.com> writes:

[ I think the economics stuff has gotten too far off-topic. ]

> > An example of a successful pool in the free software community is gcc.
> > The gcc steering committee has 14 people at the point.  The members
> > are employees of at least 7 companies.  They help coordinate gcc
> > development efforts from a number of paid and unpaid contributors.
> 
> I don't think that's a pool in the money sense, though. It isn't much
> different from things like the NetBSD core team or the Apache group or
> what have you -- we understand pretty well how to coordinate large
> distributed groups of developers at this point. What we need are
> better ways to get a fraction of them paid.

When I look at the description of the NetBSD core team on netbsd.org,
I have no idea who those people work for.  When I look at the gcc
steering committee, the institutional affiliation is listed for all
but one person.  Those people are being paid to work on gcc, and their
employers are claiming due credit.

You are correct that the different companies are not chipping money
into a pool which is then being used to pay those people.  But it
could be set up that way.  Then you would have something like the Open
Software Foundation, only smaller.

I see two reasons why that is not done.  The first is that it wouldn't
make much difference--there is already a reasonable level of trust
among the members of the steering committee, so their institutional
affiliation is not terribly important.  The second is that the
plurality of steering committee members work for Red Hat (and used to
work for Cygnus); the steering committee is in part a political
fiction to cover the fact that Red Hat dominates the direction of gcc
work, and in part a political maneuver to keep Red Hat from being in
sole control.

Anyhow, my point is, gcc is effectively being run by an industry
consortium, and it shows.  That may be true for NetBSD and Apache, but
it is not obvious from reading the descriptions on the web page.

Ian