[lug] The changing Linux Community was Re: cp and rm

[Tom Tromey]

>>>>> "Wayde" == J Wayde Allen <wallen at lug.boulder.co.us> writes:

Wayde> I too have been troubled by the appearance of this attitude in
Wayde> the Linux ranks.  When I first got involved with Linux there
Wayde> was a real sense of community.

I think there is still an actual community.  In fact I think it is
bigger than ever.  Back in the olden days (my first free software
experience was hacking bash and Emacs (18!) back in 1990) you could go
a long time without meeting a free software person.  They just didn't
exist in the flesh as far as I could tell.  Nowadays if you wear a
Debian t-shirt someone might stop you on the street.

Things do feel different though.  "When I was a kid" things were more
intimate.  I knew the names of most of the people doing real work in
the area (or so it seemed).  You could track all the projects
(unbelievable).  The energy barrier to hacking seemed lower; I used to
write miscellaneous patches more than I do now.

Size has its disadvantages.  Maybe the "real sense of community" you
speak of was somehow a feature of the smaller size.

Wayde> Once people started making this a business things changed.
Wayde> Instead of a distribution being a system constructed by an
Wayde> individual or a cooperative group of people, it has become a
Wayde> system created by a corporate entity.  In a sense, this makes
Wayde> the people behind the distribution kind of faceless

I think it is important to differentiate here.

The existing commercial distributions are faceless.  I even find the
Red Hat distribution faceless, and I work here (I'm quite far from the
OS Devel group in what I do, geographically, and on the org chart).

I think this is disappointing.  It doesn't have to be that way.  And,
as proof that Red Hat is now a big company, I have a contradictory
example from the same source: if you look at how the gcc hackers at
Red Hat work, you'll see it is very transparent.  Sure, there is stuff
you don't know (why is Mr. Hacker suddenly checking in tons of patches
for Obscure Back End X?).  But in general the feeling is very open.
Work is done on the public tree; if you watch you can see some of it
is really developed there (you can see this by watching the bug
reports :-).  Particular RH employees are responsible for particular
parts of the public compiler.

In fact, this has even gotten better over the years.  Before I worked
at Cygnus, it was a black box.  It was very hard to find out what
anybody really did there.  In those days, gcc development was very
closed (the choice of the FSF), and gdb was basically run as a
Cygnus-only project -- gdb releases were done straight out of the
internal Cygnus cvs repository.  Nowadays there is a whole web site
and a bunch of overactive public mailing lists for these projects.

So this kind of transparency and community involvement doesn't
necessarily accompany corporate involvement.  Instead, my experience
is that corporations would be smart to approach things this way.  It
has paid off big time for gcc and gdb.

I don't know why Red Hat OS Devel is more closed.  It's at least
partly a cultural issue (perhaps "we've always done it that way" -- I
can't say for sure).

My impression is that SuSE has this problem even worse.  As I
understand it their distribution isn't free software.  I really think
that is too bad.  Actually, I think it is worse -- I think that is a
bad decision, which could cause them problems in the future.  The big
lesson of the free software businesses, I think, is that more openness
is better.

Tom