Re: [sv-ac] JH comments on non-Annex F 1932

Hi John,

See my comments below


On Sun, Jan 06, 2008 at 08:25:03PM -0600, John Havlicek wrote:
> 
> * p. 6, Table 16-25.  Consider lowering the precedence of "next" family.
>   Perhaps it could go between "|->" and "always"?
> 
>   REMARK:  I have gone back through some of the mail on operator 
>   precedence.  I'm not sure that it is good to rely on deeper operand
>   analysis as a part of parsing to resolve precedence conflicts. 
>   I don't think we have done this in past discussions of similar
>   operator relationships.
> 
>   For example, I saw Johan's comments that interpreting
> 
>       next r |-> p
> 
>   as
> 
>       (next r) |-> p
> 
>   is not possible because of syntax error.  But I would read this
>   instead as saying that if "next" is higher precedence than "|->", 
>   then parentheses are required to avoid syntax error:
> 
>       next(r |-> p)  // legal
>       next r |-> p   // syntax error, since this groups as (next r) |-> p
> 

I thought this way at first too, but then I realized that there is a
difference between the semantics of the precedence rules and their
implementation. For example for 

next r |-> p

the relative precedences of next and |-> don't matter for the semantics
of the formula because there is only one possible legal placing of
parentheses which is

next (r |-> p)

how parsing should be implemented to get this result seems to be another
matter.

>   This is similar to
> 
>       a ##1 (b throughout c[->1])  // legal
>       a ##1 b throughout c[->1]    // syntax error

In fact at least in our implementation (Verific's)

      a ##1 b throughout c[->1]  

is legal and parsed as 

      a ##1 (b throughout c[->1])

notwithstanding the fact that ##1 has higher precedence than throughout.

Best Regards,

Johan

> 
>   
> * p. 6, Table 16-25.  Consider lowering the precedence of "iff".  I'm
>   not sure where the logical implication operators are going, but maybe
>   "iff" should be with them.  My guess is that these should be lower
>   precedence than "|->".  My rationale has to do with the way we 
>   interpret operator precedence with respect to syntax as described above.

-- 
------------------------------------------------------------
Johan M�rtensson                 Office: +46 31 7451913
Jasper Design Automation         Mobile: +46 703749681 
Kvarnbergsgatan 2                Fax: +46 31 7451914
411 05 Gothenburg, Sweden        Skype ID: johanmartensson
------------------------------------------------------------

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