[Tfug] "Linux is a forking mess."

[Claude Rubinson]

On Thu, Jul 19, 2007 at 12:15:38PM -0700, Jeremy C. Reed wrote:
> > One of the points in the above essay is that the major BSD distros
> > continue to persist because there are compelling reasons for the
> > differentiation (i.e., stability vs portability vs security).
> 
> Ouch. That implies that some don't consider stability, or portability, or 
> security as important. That is not true. (I am a NetBSD and DragonFly 

Not what I meant, although I see that it could read that way.
Certainly none has focused on one area to the exclusion of the others
but there are traditional niches.  If you want the absolute best in
security, you go with Open; performance/stability on 386, Free; and an
esoteric arch, Net.  This isn't to say that these distros don't
perform well at other tasks (indeed, they all benefit from the sharing
of code) but simply that they have (traditionally) emphasized one over
another.  And while people might point to this changes such as the
ports of Free, it seems to me that these divisions have never been
hard and fast--they all bleed into each other's niches at some
point--but rather rough rules of thumb.

It might also be worth pointing out that niches maximize potential
which, of course, depends upon your admin.  In my hands
any distro is going to function equally well (or poorly, as the case
may be) as I don't really have the skills to take full advantage of,
e.g., Open's security focus.

The neat thing about the observation that distros/apps/etc (and here
I'm speaking of the entire f/oss ecosystem) occupy different niches is
that it's a way of incurring the advantages of competition but not the
disadvantages.  Features that benefit all distros get shared so that
each distro can focus on developing the features needed to satisfy its
niche.

Take GNOME and KDE, as an example.  GNOME has privileged ease-of-use
and user-friendliness above all.  And while a lot of people dislike
GNOME's lack of flexibility, they've produced software like
NetworkManager, HAL, Project Utopia, etc which a lot of people seem to
really like.  KDE is a bit harder to pigeon-hole because I wouldn't
want to say that the community doesn't care about ease-of-use but I
think that it would be fair to say that KDE privileges flexibility
and, as a result, has nurtured a much richer development environment.
The competition between the desktops, then, remains more-or-less
friendly because of the (roughly) distinct niches.

C.