(j3.2006) Public Comment J32034
Wed Jul 16 22:52:13 EDT 2008
j3-bounces at j3-fortran.org wrote on 07/16/2008 09:41:58 PM:
> On Wed, 2008-07-16 at 19:28 -0600, Jim Giles wrote:
> > ASSOCIATE isn't a new consruct, so I didn't comment on it at all.
> > Actually I find it appalling.
> It's just an inline internal procedure. What's appalling about that?
Actually not just users find it appalling, implementers too found it
appalling. Since it looks more like a pointer assignment than an inline
function call (or argument association), I'll bet you most of people will
try to understand the feature from that angle.
> It's a degenerate case of SELECT TYPE, with prettier and less confusing
> spelling than
> SELECT TYPE ( A => <messy designator> )
> CLASS DEFAULT
> stuff involving A instead of <messy designator>
> END SELECT
Except that the <messy designator> here in the SELECT TYPE has to be
polymorphic, and there is no such restrictions on ASSOCIATE construct. I
found this ASSOCIATE construct is much more appalling than SELECT TYPE. So
far I haven't found a convincing evidence for its necessity in the
RL Fortran Compiler Test
IBM Toronto Lab at 8200 Warden Ave, Markham, On, L6G 1C7
Phone (905) 413-3444 Tie-line 313-3444
email: jimxia at ca.ibm.com
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the J3