Subject: Re: But why?
To: Larry McVoy <lm@neteng.engr.sgi.com>
From: Travis Hassloch x231 <travis@evtech.com>
List: tech-kern
Date: 10/24/1996 18:03:40
In message <199610242114.OAA13654@neteng.engr.sgi.com> you write: 
> : >Why do you think the federal government is specing lmbench numbers and
> : >other similar benchmarks for the systems that they will purchase?

Actually let's just drop this whole thread.
The use of popularity as any kind of metric for performance, elegance,
functionality, or anything but popularity is dubious at best, misleading
at worst, and yet it comes up in almost every debate.
Citing authorities' opinions falls into the same category.

I'd like to see a better sharing of ideas between these groups
-- especially in device drivers -- (email me if you have ideas)
but as long as people feel it necessary to bash one thing to
prove theirs is better, or worse yet attack people personally,
it's not going to speed its acceptance, except in the rare cases
of people big enough to overlook it.  In fact I think the more one
bashes the status quo, the more defensive and resistant to
your change people get, and you're not helping the world anymore.

Deciding what is worth doing seems to be a hotly debated
issue in technical circles, and probably is not worth arguing over.

I've stated my opinion on the system call design.  I haven't looked
too deeply into any of the restarting or faulting, but if it all
works cleanly let's dump this on the TODO list and get on with things.
I've had to edit this four times to sand the sharp edges off,
and I need to get real work done.