Minutes of IEEE P1800 SV-AC meeting #2006-19: 10/03/2006
Written by: Ed Cerny

Code for the conference call:

        Domestic:         888-635-9997
        International:    763-315-6815
        Participant code: 2638198

Attendance Record
	Legend:
		x = attended
		- = missed
		r = represented
		. = not yet a member
		v = valid voter (3 out of last 4 or 75% overall)
		n = not valid voter

New PAR, presence record re-initialized on 8/22/2006:

n[x-----] Faisal Haque (Cisco, Chairman)
v[xxxx-x] Eduard Cerny (Synopsys - Co-chair)	
v[xxxxxx] John Havlicek (Freescale)
v[xxx-xx] Doron Bustan (Freescale)
v[x-xxxx] Manisha Kulshrestha (Mentor Graphics)
n[------] Volkan Esen (Infineon)
v[xxxxxx] Bassam Tabbara (Synopsys)
v[xxx-x-] Hillel Miller (Freescale)
n[-x---x] Surrendra Dudani (Synopsys)
n[------] Joseph Lu (Altera)
n[------] Yaniv Fais (Freescale)
v[xx-xxx] Dmitry Korchemny (Intel)
v[xxxxxx] Lisa Piper (Cadence)
v[-xx-x-] Jiang Long (Mentor)
n[x-----] Don Mills (Microchip)
==|--------------------------------10/03/2006

Agenda:
-------

- Review IEEE working group rules
- Review errata reported in Mantis

Minutes:
--------

1) #1550: There are several related issues that have to be resolved at
the same time:

- define initial values for all sampled value functions as the initial
value of the type, rather than X.

- Define $sample(expression) as the value of expression in the
proponed region of the time step in which the function is called.

- Define the behavior of $sampled(expression, clocking_event) as the
behavior of the associated variable sampled_clocked:
var type(expression) sampled_clocked;
always clocking_event sampled_clocked = $sampled(expression).

Note: This does not seem to correspond to $past(expression, 1,
clocking_event) because its behavior although similar requires that
the assignemnt in $past be non-blocking (<=) to avoid race
conditions (?)

- Define the behavior of $sampled in disable iff: This requires to
change the text regarding clocking event specs in sampled value
functions used in disable iff on pp234, 17.4, as follows or similar:
"No clock inference applies to disable iff. The sampled value function
needs to have an explicit clock if required by the function. $sampled
will return the immediate sampled value if a clocking event is not
specified or the value synchronized with the clocking event if
specified."

AI: John - relationship with $past; Ed - proposal; Dmitry - verify.

2) #1551: To be discussed once 1550 is resolved.

3) #1549: There are several issues to be resolved regarding type
specification for formal arguments in properties and sequences.:

- Type compatibility for sequence and property types: property should
accept property, sequence and expression as actual arguments, sequence
should accept sequence or expression. 

- What kind of type checking should be imposed on variable types. How
strict? Compatible with functions / tasks, pass by ref or by value?

- when a property / sequence that has formal arguments is passed as an
actual argument to another property (or sequence), what actual
arguments can be supplied to the passed property / sequence? Can it
have a formal argument that is a local variable and if yes what can
that formal argument be bound to?

4) Result of vote on #805: 5 yes, 0 against. Accepted.

Next meeting: 

10/10/2006 at 9 am PT (12 pm ET, 5 pm GMT): Discussion on 1550, 1551,
1549, etc.

1/2h regular meeting, 
1/2h presentation by Dmitry - continuation of proposal for enhancements.

====