eliminating GPL apps on FreeBSD
beyert at cs.ucr.edu
Sat Dec 18 02:36:53 PST 2004
On Fri, Dec 17, 2004 at 07:58:53PM -0600, Jorge Mario G. wrote:
re-written parts, or at least the FreeBSD project
> so that way we would reduce the GPL code
> in the system
> shouldnt be there an initiative to do this?
I agree with this, however idealistic (or ideological if you will)
this might seem.
The project of eventually attaining license purity should be divided
(roughly) into 2 areas:
1) replace GPL'd userland utilities, (there are still quite a few, and it
isn't always obvious because of BSD modifications on the gnu
versions) This isn't nearly as easy as it's often made out to be,
but I think this might be possible in the next few years
2) significantly improve Tendra C compiler (or any equivalent,
non-proprietary, non-copylefted compiler) for base system. Should
produce optimized code and should be fairly platform independent.
I have heard the most objections to replacing gcc, and although the
reasons are perfectly rational, it is simply hypocritical for the BSD
community to forever rely on a GPL or proprietary licensed compiler.
In my view, it was somewhat of a mistake to rely on gcc in the
first place. In the present situation, there are no alternatives to
gcc, even in many proprietary situations. A competitive BSD licensed
compiler could change this situation.
It's going to take years of work to get a worthwhile BSD-licensed C
compiler, but I think the effort would be worth it. Desktop projects
will probably always use gcc, (but then again those are probably GPL'd
anyway) but the base system should have a native BSD compiler.
In the long term, there might be some compelling advantages to this
approach as well. Namely:
1) Fixes made by various BSD projects are not just "forks" of gcc
2) Compiler won't just be specialized for Linux and most importantly
3) The BSD community will grow larger to accommodate the development.
More projects means more developers, and more developers means more
And to everyone who says this project is "impossible," or a "waste of
time," how was it that Richard Stallman was able to get gcc usable
practically by himself (perhaps through the efforts of Cygnus, but I
think they came in later) in just a few years? Some might say that
gcc, like GNU/Linux, was a "fluke" project, but I would disagree. It
had a steady stream of development during its infant years, and this
made a tremendous difference. If only I could say the same for
Finally, to those who believe that licensing is irrelevant, (I
obviously disagree here) why exactly are you using BSD, much less
free software in the first place? Licenses are important, they just
don't seem relevant for the most part to developers who are trying to
get real work done.
More information about the freebsd-chat