SV-EC Meeting Minutes
30 September 2002

(--a-----) Adam Krolnik (LSI Logic)
(---a-a--) Alec Stanculescu (Fintronic)
(a-aaaa-a) Cliff Cummings (Sunburst)
(-aaa----) Dave Kelf (Co-Design)
(aaaaa-aa) David Smith (Synopsys)
(-aaa-a-a) Dennis Brophy (ModelTech)
(---aaa-a) Francoise Martinolle (Cadence)
(aa-aa---) Heath Chambers (HMC)
(aaaaaaaa) Karen Pieper (Synopsys)
(aaa-aaaa) Kevin Cameron (National)
(---a-aa-) Kurt Takara (0-in)
(--a--a-a) Michael McNamara (Verisity)
(aaaapaaa) Mehdi Mohtashemi (Synopsys)
(-aaaaaaa) Neil Korpusik (Sun)
(aaa-----) Paul Graham (Cadence)
(aapaa---) Peter Flake (Co-Design)
(a-------) Roy Armoni (Intel)
(aapa-a--) Simon Davidmann (Co-Design)
(-aaaaaaa) Stefen Boyd (Boyd Technology, Inc.)
(aa---a--) Steven Sharp (Cadence)
(-aaaaa--) Tom Fitzpatrick (Co-Design)
(-----a-a) Tim Corcoran (WHDL)
(-----a--) Stephen Meier (Synopsys)
(-----a--) Arturo Salz (Synopsys)
(-----a--) Zeev Kirshenbaum (Verisity)
(-----aa-) Brad Pierce (Synopsys)

a => Attended
p => Attended by proxy
- => Missed

Action Items:
    1) Kevin: Add detailed description of requirements for Data Channels
    2) David: Follow-up with Stu on meeting and understanding of LRM work
    3) Kevin and Mehdi: Discuss use of pointers in design work
    4) Kevin: Rework alias proposal based on feedback (keyword, what types will be supported,
	and compatibility rules)
    5) Mehdi: ask Arturo for more elaboration on $cast_assign

Thanks to Tim Corcoran and Stefan Boyd for taking the minutes. David Smith combined the results into these.

1. Review and approve minutes
    Proposal to accept minutes
    	Proposed: Tim C.
	Second: Karen P. seconded
	Abastain: Mac and Cliff
	Passed

2. Reviewed open action items from last meeting
    Old Items:
    1) Kevin still needs to add requirements for data channels.
    2) CC is not goind to do any enhancements for 3.1 - the FSM issues will be delayed
	past 3.1.
    New Items:
    1) David has gotten confirmation from Stu, Arturo, and Mehdi to be on the LRM sub-committee.
    2) Kevin sent the alias proposal for review
    4) David updated the web site.

3. Extensions list discussion
   A. ONLY Extensions 2,3,4 met deadlines for consideration and had proposals.
   B. Constraints & Randomization was proposed by another committee.
   C. Ref Pointer: 

	Kevin was asked to talk us through his proposal doc.

	Q:(?) What would these be used for in HW design?
	Q:(Mehdi) Should references apply to generic datatypes (e.g. integers)?

	Neil K: Calls to var (similar situation) can be confusing to the user
            is it pased by ref or by value.
		Solution 1: new syntax to indicate dereferencing
		Solution 2: Naming convention but cannot be enforced (bad idea).

	David: There are many different aspects and implications here that
	    we need to get a handle on. New semantics, pass by reference,
	    local declarations of references, operators for those references.
	    Some aspects of this proposal (for completeness?) conflict 
	    with TB donation, how do we resolve?

	David: Didn't we say we DON'T want to do pointers? Is this an end-run?

	Kevin: We can dissect the pieces people don't like and remove them.

	Francoise: I think pointers and the ability to traverse them (at least for
	    using opaque data) is essential to a system level language.

	David: What is different between the testbench use of var and the support for opaque
	    data pointers?
		            
	Suggestion: To save committee time, Kevin and Mehdi get together to 
	     discuss this proposal in person. (**AI for Kevin & Mehdi)    

   D. Alias proposal
	Cliff: Could we say "alias" instead of "assign {packed} union"

	Francoise: What are the datatype compatability rules?

	Cliff: Perhaps for simplicity it only applies to nets?.

	David: Can you alias to an interface?

	David: Can Kevin rewrite proposal with "alias" and other suggestions? (**AI for Kevin)

5. Issues from TB Donation

	David: These are misnumbered (two #19's) which I will correct

	REV-19a: Is $cast_assign bi-directional?
	    - Accept Arturos response

	REV-19b: $cast_assign on assignments only?
	    David: Answer may differ between TB and design code
	    Karen: Has runtime efficiency problems
	    David: enums may carry more information than their immediate value
		- Runtime checking issues TBD
		- Perhaps a choice of checking or not? TBD
	    (**AI for Mehdi: ask Arturo for more elaboration

	REV-20:  Is pass-by-reference or pass-by-value always controlled by the user?
	    ?: Currently pass-by-ref is unsupported in SV.

	    - Answer is helpful, but we need to resolve a larger issue
	      of which datatypes are supported by pass-by-ref control.

	REV-21:  Superclass instance to a derived class
	    - Arturos response is clear
	    - needed to detect run-time error
	    Mehdi: This is crucial to avoid "clobbering" data
		 tho' there may be runtime overhead
	    Mehdi: We need to addd more examples to LRM to clarify the issues

	REV-22: Strict FIFO? 
	    Mehdi: For many uses FIFO is sufficient, users can get more sophisticated
		using lists, arrays, objects etc. for systems with non-sequential access,
		LIFO's etc.
	    - Accept Arturos answer

6. Other Issues from TB donation

	1: $cast_assign
	    Francoise: Why is $cast_assign needed, can't it be implicit?
	    Mehdi: Dynamic checking

	    - No resolution. Still an issue
	    - same kind of question as as item REV-20/REV-22

	2: Should there be new '$' system tasks and functions?
	Discussion on other alternatives: operator, system functions, tasks
	   (**AI David: Add confusion with system task ($) issue to list)
	DAVID: We will resume with REV-6 at next meeting.

7. Meeting closed.