[MATRIX-SIG] A proposal (LONG) - "reverse of take"

Phil Austin phil@geog.ubc.ca
Thu, 3 Jul 1997 18:34:48 -0700


>>>>> "Andrew" == Andrew P Mullhaupt <amullhau@ix.netcom.com> writes:

    Andrew> At 07:29 PM 7/2/97 -0700, Phil Austin wrote:
    >> And C++ scientific programmers have adopted the Fortran90 Index
    >> notation.

    Andrew> I find that while the F90 indexing is better than C,
    Andrew> Pascal, etc., it's pretty thin soup compared to most array
    Andrew> languages. Are they going to improve it for F95?

There's an "enhanced where" (slightly more powerful boolean
constructs, see the appended features list).
The cutting edge is now off wrestling with OO-Fortran.

The glacial calm among the language conservatives in comp.lang.fortran
make it a very restful place to visit when the tumult in some of the
other newsgroups is raising your blood pressure.

* Enhanced WHERE construct [with nested WHERE, masked ELSEWHERE, and construct 
names]
* The FORALL statement and construct
* PURE user-defined procedures
* ELEMENTAL user-defined procedures
* Pointer initialization [i.e., the NULL( ) intrinsic]
* Derived-type structure default initialization
* Automatic deallocation of allocatable arrays [at exit from a procedure]
* CPU_TIME intrinsic
* Enhanced CEILING and FLOOR intrinsic functions [with KIND argument]
* Enhanced MAXLOC and MINLOC intrinsic functions [with DIM argument]
* Enhanced SIGN intrinsic function [can detect negative real zero]
* Printing of -0.0
* Generic identifier allowed in END INTERFACE statement
* Zero-length formats [I, B, O, Z, F expands field width to fit data]
* Comments allowed in namelist input
* New obsolescent features [as specified in F95]




_______________
MATRIX-SIG  - SIG on Matrix Math for Python

send messages to: matrix-sig@python.org
administrivia to: matrix-sig-request@python.org
_______________