bglenden@colobus.cv.nrao.edu (Brian Glendenning) (04/11/91)
A lot of people (including, obviously, big name vendors) think that
FORTRAN extended is a bad thing.
I would be interested in learning what these reasons are. For our
purposes the "it's not FORTRAN any more, it's a new language" argument
isn't very important. Otherwise what are the (purported) flaws?
If possible, please respond by email and I will post a summary of
responses when they have finished trickling in.
Brian
--
Brian Glendenning - National Radio Astronomy Observatory
bglenden@nrao.edu bglenden@nrao.bitnet (804) 296-0286bglenden@colobus.cv.nrao.edu (Brian Glendenning) (04/16/91)
A few days I posted a summary asking people to email me the problems
in F90 and promised to post a summary. Thanks to:
khb@chiba.Eng.Sun.COM, vsnyder@jato.Jpl.Nasa.Gov, jlg@woodsy.lanl.gov,
kauff@neit.cgd.ucar.EDU, mcdonald@aries.scs.uiuc.edu,
corbett@Eng.Sun.COM, tom.kovar@itc.univie.ac.at,
cflatter@zia.aoc.nrao.edu, rbe@yrloc.ipsa.reuter.COM,
pfkeb@EBNEXTK.SLAC.Stanford.EDU
Here is the quick summary. I will forward the unedited messages to
anyone who want to see them.
1. F90 will invalidate some valid F77 codes (in fairly easy to fix
ways)
2. Vendors will have to work very hard to get good compilers - it is a
big, complicated language.
3. Pointers are not as straightforward to use as one would like, e.g.
the pointers to slices are hard to implement but not all that useful.
Moreover the pointers aren't done very elegantly or intuitively.
4. Array sections in general may not allow efficient operations.
Brian
--
Brian Glendenning - National Radio Astronomy Observatory
bglenden@nrao.edu bglenden@nrao.bitnet (804) 296-0286