Re: SEUL: Is it too soon for me to comment?

[William T Wilson <fluffy@dunadan.com>]

On Mon, 5 Jan 1998, Paul Anderson wrote:

> > 2) Starting from scratch - but I suspect that all "new" OSs start like that
> > 
> You have got to be NUTS!  It would take us a month just to get something
> that printed 'A' then hung the computer!  It took 6 years of development

I think he means start a new distribution from scratch, not start a new OS
from scratch.  :)

> > Usually, yes.  :)  What I'm saying is that Linux out of the box has
> > IRQ's tuned wrong, has bad modem init strings, often has the UART
> > speed set wrong, and that this confuses a multitude of new users and
>
> HUH?!  Which distribution are you talking about that's THAT buggy?  I
> have NEVER used a distribution that bad...  I've used Slackware 1.2,
> Rh3.0.3 and 4.2 and have never encountered a problem, nor heard of

You obviously haven't talked to the same users I have, then.  :)  It's not
a distribution problem.  Have you heard of 'irqtune'?  It's frequently if
not always necessary to set that up to get any kind of throughput in a
loaded system.  The setting of the DTE speed is another common trap.  Lots
of modems default to 2400, 9600, or 19200, which will kill your
throughput; setserial is needed to solve this.  Modem init string is
another trap... Sometimes they work with ATZ, some of them need 40 and 50
character init strings.  And there's nothing the distribution can (does?) 
really do about any of these.

Do modern Redhats include irqtune?  Do they use setserial for anything
productive?  I haven't used a modem on any sort of Redhat system, ever.
(sheepish-looking) But I've had to help a lot of other people sort them
out... setserial, irqtune, and the init string.  Works every time.

Modem init strings can only be solved the way the monitor problem can be.
Guess, or collect a gigantic database of init strings (which wouldn't be
a bad thing to have around anyway).