[sv-ac] RE: cyclic dependency definition

Hi John,
It may not hurt to have explicit description of constructing a directed
graph in each case, but, to me, it seems sufficient to explain the term
"cyclic dependency" once and use it where necessary. In each case, the
context is quite explicit and clear.   
That is just my reading, others may think differently.
Surrendra  

-----Original Message-----
From: John Havlicek [mailto:john.havlicek@freescale.com] 
Sent: Thursday, November 08, 2007 2:54 PM
To: surrendra.dudani@synopsys.COM
Cc: sv-champions@eda.org; sv-ac@eda.org
Subject: cyclic dependency definition

Hi Surrendra:

In the Champions meeting today, you suggested having one definition of
cyclic dependency, and I agreed that I thought that sounded like a good
idea.

On second thought, I am not sure it will work out so well.

It's not a question of whether we can write one definition of what it
means to have a cycle in a directed graph.  That is well understood.

The problem is that the text in the footnotes is aimed primarily at
saying how the directed graph is formed in the various cases.  We will
still have to do this even if we have a single definition of cycle.

To put this another way, we cannot give an abstract definition of cyclic
dependency among a set of entities because we also need an abstract
representation of the relation among the entities that is used to define
the arcs in the directed graph.

After reflecting on this, I don't think making the single abstract
definition is really going to help, and I think we should just move the
footnotes into the text as I had originally suggested.

What do you think?

J.H.

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.