(j3.2006) Comments on 14-117r1

Malcolm Cohen malcolm
Wed Feb 12 00:26:28 EST 2014

Hi folks,

>There is a similar problem concerning instances of subprograms.  The

This is not a problem.  We left this as subprograms because of the data object 
sharing between the multiple procedures defined by a single subprogram via 

It would be theoretically nice to rewrite it all to talk about procedure 
instances, but since we know there are traps lying in wait for us it would seem 
to be a high risk strategy.

>(i) The current wording of C533 is slightly defective, as it does not
>    clearly prohibit "REAL,DIMENSION(*) :: dummy,nondummy", seeing as
>    how that does indeed declare "the array bounds of a dummy data
>    object".  C533 should probably be reworded similarly to C534a.
>F08/0086 appears to address this problem.  C535 in 14-007 appears to
>work correctly.  C533 now refers to <assumed-size-spec>, not
><assumed-implied-spec>.  No edits appear to be needed.

I think the edit supplied is in fact needed for the very similar statement
   REAL,DIMENSION(10,*) :: dummy,nondummy

(You may deduce that I wrote this up before we changed the syntax which avoided 
the problem for the rank=1 case.)

>[179:28] Restrictions on DO CONCURRENT constructs, p1,

It should be straw-voted which of these to do, to give guidance to the editor. 
Discussion is in section 2 item (a) of the paper.

>[232:15-16] 9.7.1 Wait operation, p1,
>{EDITORIAL CHANGE (a), part 1.}

I note the typo that this is editorial item (f) not (a).  The same typo appears 
in the comment for the edit at [232:29+1-2].

It should be straw-voted which of these to do, unless no-one cares which one the 
editor does when he applies the paper.  Discussion is in section 2 item (f) of 
the paper.

In section 5,
>We are sympathetic to this view, but have not developed edits.

If there is consensus on the forms preferred that would be good info to have.

................................Malcolm Cohen, Nihon NAG, Tokyo. 

More information about the J3 mailing list