RE: [sv-ac] mantis 1550

Hi,

I think that we should allow events to begenerated fro all sampled value
functions. Updated in postponed and if value changes, generates an event
which can trigger a process in active or reactive region of the
subsequent time step (in which the new vlaue can be read using $sampled)
depending where the process was suspended. (Excuse perhaps inexact
terminology.)

If you go for synthesis or formal, i.e., synchornous context, simply
ignore the sampled values, use the current value.

ed


> -----Original Message-----
> From: owner-sv-ac@eda.org [mailto:owner-sv-ac@eda.org] On 
> Behalf Of John Havlicek
> Sent: Friday, November 03, 2006 12:50 PM
> To: sv-ac@eda-stds.org
> Subject: Re: [sv-ac] mantis 1550
> 
> Hi Ed:
> 
> That does seem too drastic -- the assignment you show seems 
> like a clearly useful example of the sampled value function.
> I know our users would like to be able to write this kind of
> assignment.
> 
> My original point of view was that we should define when 
> changes in the sampled value function produce update events
> so that everything would (hopefully and, perhaps, wishfully)
> be defined.
> 
> It seems to me that we definitely do not want sampled value 
> functions to behave like $time with regard to update events.
> 
> Should we go back to this and try to identify the specific
> uses of sampled value functions that caused concern for
> Stu and others?
> 
> Best regards,
> 
> John H.
> 
> > X-MimeOLE: Produced By Microsoft Exchange V6.5.7226.0
> > Content-class: urn:content-classes:message
> > Date: Fri, 3 Nov 2006 08:09:09 -0800
> > Thread-Topic: [sv-ac] mantis 1550
> > Thread-Index: Acb/RzmiEzfY5rx5TP63YAxEjfldCQAGMJDAAACGRyA=
> > From: "Eduard Cerny" <Eduard.Cerny@synopsys.com>
> > X-OriginalArrivalTime: 03 Nov 2006 16:09:11.0345 (UTC) 
> FILETIME=[6637D210:01C6FF62]
> > 
> > Therefore, should I add a statement that no sampled value 
> function can
> > be used as the source of a value-change event? Does this 
> mean that the
> > following statement would only execute once at time 0?
> > assign v =3D $past(x);
> > Isn't that a bit drastic?
> > 
> > ed
> > 
> > 
> > > -----Original Message-----
> > > From: owner-sv-ac@eda.org [mailto:owner-sv-ac@eda.org] On=20
> > > Behalf Of Rich, Dave
> > > Sent: Friday, November 03, 2006 10:54 AM
> > > To: john.havlicek@freescale.com; sv-ac@eda-stds.org
> > > Subject: RE: [sv-ac] mantis 1550
> > >=20
> > > I agree with your conclusion.
> > >=20
> > >=20
> > > > -----Original Message-----
> > > > From: owner-sv-ac@server.eda.org 
> [mailto:owner-sv-ac@server.eda.org]
> > > On
> > > > Behalf Of John Havlicek
> > > > Sent: Friday, November 03, 2006 4:53 AM
> > > > To: sv-ac@server.eda-stds.org
> > > > Subject: Re: [sv-ac] mantis 1550
> > > >=20
> > > > Hi Dave:
> > > >=20
> > > > My intuition is that the generation of events from 
> changes in the
> > > > return values of sampled value functions will be useful 
> as a more
> > > > abstract view of the various changes in their arguments.
> > > >=20
> > > > Nevertheless, the reluctance to define such generated events is
> > > > understandable in the absence of a clear use model.
> > > >=20
> > > > In this case, though, I prefer that we make illegal the usage of
> > > > sampled value functions in contexts that require the 
> definition of
> > > > events generated from their changes.
> > > >=20
> > > > In that way, we can preserve the possibility of defining=20
> > > those events
> > > > at a later time.
> > > >=20
> > > > Best regards,
> > > >=20
> > > > John H.
> > > >=20
>