Subject: Re: Back to business [was: the Be thread]
From: Crispin Cowan <crispin@cse.ogi.edu>
Date: Wed, 24 Nov 1999 18:25:00 +0000

"Stephen J. Turnbull" wrote:

> >>>>> "Crispin" == Crispin Cowan <crispin@cse.ogi.edu> writes:
>
>     Crispin> The kernel is much easier to contribute to than GCC.  The
> ...    Crispin> To contribute to GCC, you have to grok a great deal of
>     Crispin> abstraction and detail.  It takes a lot of spin-up time
>     Crispin> to develop a working compiler patch.
>
> Cool!  Um, quick question:  does "working" mean "implements your idea
> without breaking what you need" or "acceptable to the GCC maintainer"?

The former, but I haven't tried the latter.


> Obvious quick hack try: bazaar vs cathedral.  Is that still accurate
> with the change of maintainership to the EGCS steering committee?
> Does that account for past history?
>
> Opinions on whether GCC has to stay like that?

I have not been inside the GCC/EGCS development community.  However, I think
you could make a case for "compilers have to be that way, and kernels
don't."  Compilers are much more inter-connected than kernels.  As I said in
my original post, a kernel is a big pile of libraries with a common syscall
interface and shared resource pool.  You can add features without breaking
other things.  A compiler, on the other hand, is a pipeline, and often
optimization opportunities require different stages of the pipeline to know
about each other's internal parts, creating wide-spread inter-dependency.


>  Whether the kernel can
> stay like that?  (As I understand it, the kernel is continually being
> pushed in the direction of greater modularization---does that mean
> things get "better" in the sense of open to contribution from almost
> anybody?)

Yes, I think so.

Crispin
-----
Crispin Cowan, CTO, WireX Communications, Inc.    http://wirex.com
Free Hardened Linux Distribution:                 http://immunix.org