[lug] Fedora *MEETS* KRUD comments wanted

Tom Tromey tromey at redhat.com
Sun Sep 28 04:14:16 MDT 2003

>>>>> "Nate" == Nate Duehr <nate at natetech.com> writes:

Nate> Just how *much* do developers from each of the distros contribute to
Nate> the raw packages of software that make up Linux as we know it today.
Nate> And how would one figure that out...?   Hmmm....

I can't speak for other distributions, but for Red Hat, I know that at
least for some packages, Red Hat employees contribute quite a bit.
Take a look through the gcc or gdb mailing list archives, or
ChangeLogs.  There are many, many contributions from redhat.com

I'm sure the same holds for the kernel, and for gtk, though I'm less
familiar with these projects.

Nate> Then of course, is the "how do you quantify it?" question... lines of
Nate> code?  Uggh.

You can get a rough sense just by reading mailing list archives.

Nate> It would be interesting to try to quantify where the "real work" in
Nate> Free Software comes from.  Might lead to some interesting
Nate> discoveries...

Someone tried this a few years ago, but as I recall, their methodology
was a bit flawed, and they ended up with obviously skewed results.

Nate> There are probably a
Nate> lower percentage of "distro maker" type developers who are regularly
Nate> contributing upstream to the root packages these days -- as they spend
Nate> their time working on the distribution of choice.  They work hard on
Nate> packaging and distribution.  ??

I'd suppose this is pretty accurate.  You'd expect to find people
working on packaging making changes to fix build problems.  Also you'd
expect distros to have an interest in certain key components: kernel,
compiler, libc, maybe a couple others.

But when looking at a company it is probably a mistake to assume that
such folks are a majority of those doing engineering work.  And you
could expect to see actual upstream patches skewed against this group
if they work in a company -- companies tend to have people specialize,
so if you find a compiler bug, you hand it to the compiler group and
forget about it.


More information about the LUG mailing list