Re: "obsolete" hardware

Thomas Pornin (
Wed, 11 Jun 1997 20:33:18 +0200

In article <> you write:
> You're missing the point. I'm not talking about taking it out. I'm not
>even talking about the 386 specifically.
> Again my point is not to waste time developing for obsolete
> If you so love the 386 - its your prerogative. As far as I'm concerned
>it is deceased - it RIP - it is a dead parrot (oops chip).

That's right, obsolete hardware is, by definition, dead, and should not
be supported. But I disagree when you say that 386 is dead, for two

-- the first reason is that there are plenty of 386 out there, and, in
fact, they are all the more sufficient for what we do 95% of the time.
I am personnaly a researcher in cryptography, and I so run sometimes
some programs for two or three weeks; but my pentium is way too
powerful for everyday life, which is namely email reading and,
on times to times, word processing. I would still be happy with
a 386. Ten years ago, such a beast was considered as rather good.
It still is. It has not changed. Only there are better processors

-- the second reason is that a big pentium is not really more than a
big 386, from the software point of view. It has indeed some new
features, and runs way faster, but using it as a 386 is in no way
a performance eater. That's why the kernel should run on any 386.
If it does not run, then the answer is not "drop 386 support". The
answer is "find the bug".

So, in conclusion, I would say: a 386 is not obsolete, as, besides
being a powerfull router, en efficient VT and a good text editor, it is
a correct kernel-code-tester.

And, yes, I use a 386 as router.

--Thomas Pornin