(j3.2006) Public Comment J32036 (Submodules)

[Bill Long]

Michael Ingrassia wrote:
>
>> As for the third of the "advantages" of submodules, nesting
>> of scopes to any depth, is something that I believe the
>> language shouldn't support.
>>     
>
> Setting a definite limit in nesting depth (looks like 3 would
> handle the examples in ISO/IEC TR 19767:2004(E) ?)
> would promote portability, if as I expect every implementation will have
> some nesting depth which is practically unattainable, just as DO loops
> may not be infinitely nestable.
>   

At yet the lack of a limit on DO loops has not crippled portability.

A restriction on the nesting depth for submodules to 3 would not really 
be aimed at portability as much as trying to discourage bad management 
of large programming projects.   However, we've already failed at that 
goal so miserably that it seems incongruent to change policy now.  
Besides, I expect we would pretty soon hear from someone who absolutely 
has to have 4 levels.  I think the reasonable choices for depth are 1 or 
unlimited.  Van and others argued against 1, and they won. 

Cheers,
Bill


-- 
Bill Long                                   longb at cray.com
Fortran Technical Support    &              voice: 651-605-9024
Bioinformatics Software Development         fax:   651-605-9142
Cray Inc., 1340 Mendota Heights Rd., Mendota Heights, MN, 55120