RE: ITC Meeting Minutes for May 26th

From: Bojsen, Per <bojsen_at_.....>
Date: Tue Jun 21 2005 - 14:00:10 PDT
> So, batching of output messages from different cycle stamp 
> times is not
> possible without changing the clock control we use, and that 
> would lose
> determinism, no?

Right, that's what I'm thinking Shabtay is worried about.  He wants
to allow the user to decide that it is OK to batch multiple messages
even if controlled time is running and then have the software side
process them at some later time.  This would work fine *and* be
deterministic
as long as the reception of those messages does not imply a feedback
to the input side of the software side.  One example would be a
switch test where checking of the packets received on the software
side can be done at any time.  This is where reactivity comes in.  Such
a test would have a uni-directional transaction flow SW->HW->SW without
feedback on the SW side.

> Can you conceive of a deterministic way of 
> operation that
> allows controlled clock time to progress after an output 
> message has been
> sent?

The above example is one such case.  However, I cannot conceive of
a way that would allow the infrastructure to automatically guarantee
determinism in this case.  The burden is on the user, which we are
trying to avoid . . .

Per

-- 
Per Bojsen                                Email: <bojsen@zaiqtech.com>
Zaiq Technologies, Inc.                   WWW:   http://www.zaiqtech.com
78 Dragon Ct.                             Tel:   781 721 8229
Woburn, MA 01801                          Fax:   781 932 7488
Received on Tue Jun 21 13:59:19 2005

This archive was generated by hypermail 2.1.8 : Tue Jun 21 2005 - 13:59:25 PDT