Subject: Requirements for debug (Questions from Erich)
From: Miller Hillel-R53776 (r53776@motorola.com)
Date: Tue Oct 01 2002 - 03:53:25 PDT
Erich,
Thanks for answering this. I was deeply concerned on not getting any
feedback from
my question. I think this should stay on the reflector because it is an
issue that should
concern all.
Core CBV was designed to be simple and to avoid things like regular
expressions. This
has made it very easy to develope tools around the language that provide
many capabilities,
which one of them is debug.
Sugar and OVA use regular expressions and no one has proposed a real debug
solution for this
and I have started wondering if such a solution can exist. Accessing
assertion states
is certainly not a solution and such a proposal has made me even more
concerned.
We cannot have a language which is inherently not debugable.
A possible requirement on the same basis that one needs to prove that the
language is
formally verifiable, one could add a usage requirement to prove that the
language
facilitates debugging
Many Thanks
Hillel
-------- Original Message --------
Subject: Shalom; could you forward this to Hillel please? Thanks.
Date: Mon, 30 Sep 2002 04:38:39 -0700
From: "Erich Marschner"<erichm@cadence.com>
To: "Shalom Bresticker"<Shalom.Bresticker@motorola.com>
Hillel,
(I'd like to move this discussion off the reflector for just a bit ...)
I appreciate your concern about debugging. Formal verification in
general has needed good debugging solutions for many years - and still
does. Simulation tools have evolved more or less standard debugging
features, but it is not at all clear that they are the most appropriate
ones.
I have to admit I'm at a loss to understand how we might define a
language requirement that would help ensure that tool implementations
support good debugging facilities. As some have noted already, there
has been a discussion about providing access to assertion states, but
this doesn't immediately translate into effective debug facilities. Do
you have any ideas about this? In particular, does CBV have
characteristics that ensure "debug-ability", which might be considered
in this context? Is there anything specific that should perhaps be
proposed to the SV-AC? Or for that matter, should be considered in the
FVTC?
Regards,
Erich
-------------------------------------------
Erich Marschner, Cadence Design Systems
Senior Architect, Advanced Verification
Phone: +1 410 750 6995 Email: erichm@cadence.com
Vmail: +1 410 872 4369 Email: erichm@comcast.net
| -----Original Message-----
| From: Shalom.Bresticker@motorola.com
| [mailto:Shalom.Bresticker@motorola.com]
| Sent: Friday, September 27, 2002 5:38 AM
| To: sv-ac@eda.org
| Subject: [sv-ac] FW: Requirements for debugging capabilities
|
|
| -----Original Message-----
| From: Miller Hillel-R53776
| Sent: Thursday, September 26, 2002 10:56 AM
| To: 'sv-ac@eda.org'
| Subject: Requirements for debugging capabilities
|
| Hi,
|
| I am very suprised that there is no requirements for
| sufficient debugging capabilites.
|
| Lately I am finding it very difficult to find/define tools
| that facilitate for debugging capabilities
| that involve regular expressions. Regular expression for
| temporal languages has been around
| for some time now. I would expect that these capabilities
| would exist by now.
|
| I would hate to find out that a standard language would be
| developed which inherently could not have
| good debugging tools. Afterall the users always spend most
| of their time debugging.
|
| Does anyone else share this concern ?
| Can we add a usage requirement that will prove debugging facilities ?
|
| Hillel
|
|
This archive was generated by hypermail 2b28 : Tue Oct 01 2002 - 05:59:49 PDT