RE: [sv-cc] 'Scalar' term for SV function results, etc.

I have no problem with Jim's suggestion.

 

Shalom

 

________________________________

From: Jim Vellenga [mailto:vellenga@cadence.com] 
Sent: Tuesday, June 05, 2007 4:33 PM
To: Bresticker, Shalom; Ralph Duncan; Stickley, John; SV-CC
Subject: RE: [sv-cc] 'Scalar' term for SV function results, etc.

 

Personally, I think we're trying to solve the wrong problem.  We're

trying to come up with a brief but precise way of saying what

34.5.5 says in many more words.  That appears to be an impossible

task -- like trying to put 10 pounds of flour in a five-pound sack.

 

All we really need to say, somehow, is that fewer types can be

used as function return values than as arguments.  There are all

sorts of ways of being appropriately less precise, such as

 

"Function return values are restricted to a smaller subset (See

34.5.5)."

 

"Function return values are restricted to a subset of singular

values, however (See 34.5.5)."

 

This also has the desirable side effect of specifying legal return

values in only one place.  If we try to be any more precise, as

in 

 

(Not desirable) "Function return values are restricted to singular

values (See 34.5.5)."

 

then there's a danger that someone could argue that class variables

are allowed as return values, since they are in fact singular values.

 

Regards,

Jim

 

---------------------------------------------------------
James H. Vellenga                            978-262-6381
Engineering Director                   (FAX) 978-262-6636
Cadence Design Systems, Inc.         vellenga@cadence.com
270 Billerica Rd
Chelmsford, MA 01824-4179
"We all work with partial information."
---------------------------------------------------------- 

 

	 

	
________________________________


	From: owner-sv-cc@eda.org [mailto:owner-sv-cc@eda.org] On Behalf
Of Bresticker, Shalom
	Sent: Tuesday, June 05, 2007 4:03 AM
	To: Ralph Duncan; Stickley, John; SV-CC
	Subject: RE: [sv-cc] 'Scalar' term for SV function results, etc.

	Although Verilog did define scalar differently than in other
use, it retained the meaning of a contrast to vector.

	 

	And it actually has a logic to it. If you visualize a vector as
a line, as in elementary physics or in elementary computer graphics,
then a scalar is a point. Apply that to computer storage, and a vector
is a string (line) of bits, whereas a scalar is a single bit.

	 

	Shalom

	 

	
________________________________


	From: Ralph Duncan [mailto:RDuncan@CloudShield.com] 
	Sent: Monday, June 04, 2007 6:51 PM
	To: Bresticker, Shalom; Stickley, John; SV-CC
	Subject: RE: [sv-cc] 'Scalar' term for SV function results, etc.

	 

	Shalom,

	 

	1. The intent certainly is to distinguish a single value from
aggregate/composite values.

	2. Despite the original, non-vector meaning, common use of
'scalar' does seem to mean single value:

	        http://en.wikipedia.org/wiki/Scalar_%28computing%29

	3. The SV definition of 'scalar' as 1-bit wide is in 6.8.

	4. String appears to be included because we are indicating the
pointer to the string, rather than the string, itself.

	5. 'Singular' seems reasonable; the only obvious alternative is
to spell out that it's not 

	    an aggregate or composite (e.g., array, vector, structure).

	 

	Does anyone some memorable and snappy alternative?

	 

	Ralph 

		-----Original Message-----
		From: owner-sv-cc@server.eda.org
[mailto:owner-sv-cc@server.eda.org]On Behalf Of Bresticker, Shalom
		Sent: Monday, June 04, 2007 7:09 AM
		To: Stickley, John; SV-CC
		Subject: RE: [sv-cc] import/export function result types

		In SV, a scalar is a one-bit value. "Singular" would be
closer.

		 

		Shalom

		 

		
________________________________


		From: Stickley, John [mailto:john_stickley@mentor.com] 
		Sent: Monday, June 04, 2007 5:03 PM
		To: Bresticker, Shalom; SV-CC
		Subject: RE: [sv-cc] import/export function result types

		 

		A possible replacement for this term is "scalar values".
		
		That is the term used in VHDL and is probably less
vague.
		
		-- johnS
		
		
		-----Original Message-----
		From: owner-sv-cc@server.eda.org on behalf of
Bresticker, Shalom
		Sent: Mon 6/4/2007 5:30 AM
		To: SV-CC
		Subject: [sv-cc] import/export function result types
		
		 
		
		Hi,
		
		In Draft 3, 34.2.2 says,
		
		A rich subset of SystemVerilog data types is allowed
		
		for formal arguments of import and export functions,
although with some
		restrictions and with some notational
		
		extensions. Function result types are restricted to
small values,
		however (see 34.5.5).
		
		and 34.5.5 has a similar statement and details the exact
restrictions.
		
		However the phrase "small values" sounds strange.
		
		In what way are real, longint, and string 'small', for
example?
		
		Thanks,
		
		Shalom
		
		
		
		Shalom Bresticker
		
		Intel Jerusalem LAD DA
		
		+972 2 589-6852
		
		+972 54 721-1033
		
		
		
		
		--
		This message has been scanned for viruses and
		dangerous content by MailScanner, and is
		believed to be clean.

		
		-- 
		This message has been scanned for viruses and 
		dangerous content by MailScanner
<http://www.mailscanner.info/> , and is 
		believed to be clean. 

	
	-- 
	This message has been scanned for viruses and 
	dangerous content by MailScanner <http://www.mailscanner.info/>
, and is 
	believed to be clean. 


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