(j3.2006) [Fwd: C501 in TS 18508]

[Van Snyder]

Bill Long wrote:
> On 3/5/13 9:40 PM, Malcolm Cohen wrote:
>   
>> Since CRITICAL/DO CONCURRENT are already described as a mixture, resolving this
>> really should wait for integration.  The TS should not be attempting to modify
>> the standard for mere stylistic purposes.
>>     
>
> Perhaps, as part of integration, we could add CHANGE TEAM to the 
> constraints for EXIT, CYCLE, and RETURN [I agree that an incomplete list 
> in those constraints is unhelpful], and then add normative text in the 
> CHANGE TEAM section something like "The appearance of EXIT (\ref), CYCLE 
> (\ref), and RETURN (\ref) statements in a CHANGE TEAM construct are 
> constrained to ensure that the construct terminates by execution of its 
> END TEAM statement."  That gives the reader specific places to look for 
> the constraints, rather than requiring reading of the whole document.
>   

I don't object to this.  I don't like the possibility of having the 
relationship between EXIT/CYCLE and two constructs constrained by 
constraints described under EXIT and CYCLE, and one other described 
under that construct.  This inconsistency makes it look like the 
document wasn't planned (it wasn't) or integrated (it wasn't).  It would 
be more consistent to move the existing and new constraints on branching 
and RETURN to those discussions.  Either normative or non-normative text 
could remark that these are single-exit constructs, and refer the reader 
to those four places.  The alternative is twelve constraints.  I don't 
see an elegant way to combine the constraints on EXIT, CYCLE, RETURN and 
branching into a single constraint under CRITICAL etc., but it's easy to 
combine CHANGE TEAM, CRITICAL and DO CONCURRENT into one constraint on 
EXIT, etc.
> Cheers,
> Bill
>
>