Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84 SMI; site sun.uucp
Path: utzoo!watmath!clyde!burl!ulysses!bellcore!decvax!decwrl!sun!kevin
From: kevin@sun.uucp (Kevin Sheehan)
Newsgroups: net.micro.68k
Subject: Timing loops
Message-ID: <3291@sun.uucp>
Date: Wed, 26-Feb-86 03:10:28 EST
Article-I.D.: sun.3291
Posted: Wed Feb 26 03:10:28 1986
Date-Received: Fri, 28-Feb-86 21:30:43 EST
Distribution: net
Organization: Sun Microsystems, Inc.
Lines: 23


Ok, anybody object to the last word on the subject? :-)

Actually, I am a little concerned about the reaction on two counts.

1) if you have a box where everthing but this chip cycles ok, and
you are looking for that sexy point in the curve between cost
and performance - KNOW that the choice will (and according to religion,
should :-) be to do it in software. bienvenidos al mundo..

2) ok, now if it's in software, and assuming you checked the point, or it
was made known to you as the driver writer, (seems to be assumed so far
that you didnt get lucky) then you write whatever it takes to work
(timing loop, simple assignment) and DOCUMENT THAT ASSUMPTION!!!
right there, right then, at that line, at the top - but make
it known to the next smuck.

I dont mind someone pulling stunts if it's NECESSARY, or sometimes just
the way it is.  (some are amusing, no?) I DO mind not knowing when it's
been done, and getting bit. 'nuff said?

				l & h,
				kev