Hi,
Life is hectic, but I gotta add my two cents briefly. I do not use derived
data types very often and so can not judge the effectiveness and utility of
the proposal (which I skimmed through). But what I object to are statements
such as:
> becoming so
> more powerful than the traditional C++ or Java language class.
I believe Fortran 2K should not attempt to surpass or even become equal to
C++ and Java in terms of OOP and features such as templates, dynamic classesl
or whatever! Fortran should focus on arrays and other numerical features,
implementing only so much of these extra features as neccesary to give a nice
interface between Fortran and other languages. So, instead of taking the
developer's time away from numerical optimizations, which still have ways to
go, especially in terms of array-related matters, focus should be put on
developing powerful interfaces between Fortran 2K, C++ and Java. Now, that of
course poses troubles when each language is implemented by different vendors,
but shouldn't be that hard for big vendors like Sun or Compaq or the GNU
suite.
I know F2K has worked on this some, and I hope more will be put into it. Let
C++ and Java people work on better OO features, but enable Fortran programs
to use them as well via interfacing.
Just my thoughts on this issue ever more contested as languages become very
complex...
Aleksandar
--
_____________________________________________
Aleksandar Donev
http://www.pa.msu.edu/~donev/
[log in to unmask]
(517) 432-6770
Department of Physics and Astronomy
Michigan State University
East Lansing, MI 48824-1116
_____________________________________________
|