[cumbria_lug] New distro advice

[Michael Saunders]

On Thu, 19 Feb 2004, Trevor Pearson wrote:
>
> I would like to point out to Michael that neither I nor anyone in
> the Cumbria Linux USER group is any way responsible for Fedora Core
> or Gnome or even KDE so I doubt telling us that these projects are
> not doing what he thinks is best may generate much e-mail

You're right, but in six years of using Linux (and four of those
writing for mags on the subject), I've grown used to the fact that
open source development isn't a magic bullet. There's a gargantuan
amount of talent out there -- unfortunately, though, it's often being
used to impress others, or to scratch a specific itch. Fair enough.

But when you're not working to a spec, when you have no minimum target
platform, things are going to bloat uncontrollably. It's happening
right here, and right now. I should note that I'm not some foaming at
the mouth rant-machine here; this view is becoming increasingly common
among those who've used Linux for some time. Look around on Slashdot
and other boards, and you'll see other folks becoming concerned with
the direction the Linux desktop is taking.

> but little change.  This discussion belongs on the fedora-devl or
> equivalent gnome and KDE mailing list and if Michael is going to
> take it there I will join to watch the fireworks....

It'd achieve nothing. Fedora Core is a bleeding-edge testing distro,
GNOME is an attempt to recreate much of Windows' design philosophy on
Linux, and all the coders will naturally think it runs fine on their
2.2 GHz 512MB RAM boxes. These people aren't thinking long term.

> Michael can you get one line of your code into Longhorn ? I think not.

Trevor can you go into the GNOME and KDE project's codebases and
remove thousands of lines of unnecessary code? Nope, thought not. Your
point doesn't stand. Open source doesn't produce miracles; it needs
careful thought, design and structure -- not endless features,
buzzwords and bloat.

You can dismiss my views here, but in five years time when we're
sinking under the weight of all this unmaintained code, with more bugs
and security vulnerabilities than ever before, you might see what I'm
getting at. Featureitis and overengineering lead to intensely
difficult support problems later on. Let's think long term now!

Mike

-- 
Michael Saunders
www.aster.fsnet.co.uk