[cumbria_lug] New distro advice

[Michael Saunders]

On Fri, 20 Feb 2004, Schwuk wrote:
>
> Performance really doesn't come into it, unless you're talking about
> workstations (which you've said we're not). If BeOS "attracted
> plenty of attention" why did it die? Lack of take-up, driven mainly
> by lack of compatibility.

Yep, that's why it died. It suffered from the start, but one of the
reasons people got excited about it was its SPEED. Had it been as slow
as Linux+GNOME, it might not have lasted beyond a week. I'm not saying
we should clone BeOS, or anything like that. I'm just pointing out
that a modern desktop Linux that performed as well as BeOS would
generate a whole lot more interest.

Why, where, how and what happened to BeOS is a separate issue. But it
developed a small and hugely loyal userbase for a few years because it
screamed along. I believe performance IS important -- people rush out
to upgrade their CPUs when the next AMD / Intel line arrives, hoping
that things will run faster. And instead of improving performance and
justifying the cost of upgrades with better productivity, we seem to
be doing a Microsoft, just filling up the extra CPU cycles with gunk.

No doubt most of us on this list started off with 8 bit machines, and
later moved onto Amigas/STs/Acorns/early PCs. We all have some idea of
the relative spec of those systems, and what they offered; looking at
the situation today leaves a very sour taste though.

My A1200 (7 MHz, 2 MB RAM, 170 MB HD) booted and ran so much faster
than this PC here (800 MHz, 320 MB RAM, 40 gig HD) with Slack and
GNOME. Sure, the PC does a _lot_ more. But the PC is several thousand
times more powerful (adding up the components), and yet it doesn't
actually DO several thousand times more. It doesn't make anyone
several thousand times more productive :-)

Why should it be so much slower than the Amiga? We can list all manner
of things the PC can do that the Amiga can't, but it should reflect on
the spec and at least perform a little better. But planning, design
and efficient coding practices have gone out the window. We're getting
bogged down, and at some point it'll hit us hard in the face.

To those who think I'm bitter/ranting/insane or whatever, I suggest
this: spend a full week using QNX, Zeta (BeOS) or even Syllable.
You'll gain a new perspective on performance, and when you go back to
Linux you'll be saddened (and possibly disgusted) by how sluggish it
is in comparison. Yeah, those OSes don't have all the hardware and
software support, but that's a different issue. The point is, they're
still fully-fledged 32-bit, multitasking, graphical networked
operating systems with all the bells and whistles. And they zip along.

If you spend a lot of time with one of those, you may start to see
what I'm getting at -- see that we have a bit of a problem here.

> I can see Sun and their Java Desktop and Java Enterprise System
> doing far more damage to MS than a better performing OS would.

I can see Sun and their Java Desktop and Java Enterprise System doing
EVEN MORE damage to MS if it ran like the clappers, allowing companies
to throw it on old NT4 boxes and letting the PHBs dream about improved
productivity etc. A faster, lighter Linux WOULD increase corporate
uptake -- maybe not amazingly, but it'd have an effect.

If a company could keep their old 64 MB NT4 boxes and not spend a
penny by putting Linux on them, they'd be far, far more interested.

Mike

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