(j3.2006) Public Comment J32034

[Jim Giles]

Aleksandar Donev wrote:
> On Tuesday 15 July 2008 15:02, Jim Giles wrote:
>> Actually, for reasons given later, the idiom is more likely to be:
>> CRITICAL
>> BLOCK
>> declare local entities
>> ...
>> END BLOCK
>> END CRITICAL
> How about we add a [ specification-part ] to CRITICAL (or all the
> block constructs for that matter)? This avoid your making CRITICAL an
> attribute, which I still don't see nor agree with. It is a block
> construct ("8.1 Executable constructs containing blocks") and I don't
> see what is wrong with that.

I still disagree with proliferating the number of different kinds of
block constructs.  And I don't see how the current CRITICAL 
construct guarantees that the declarations of a procedure are
inside the critical section.  I don't even like the syntactic appearance
of END this; END that; END the_other; stacked one right after 
the other (as above).  The fewer constructs, the better.

And, I also don't think it's a good idea that all constructs be allowed
a specification part.

J. Giles