14 Jan
2025
14 Jan
'25
3:31 p.m.
On Tue, 2025-01-14 at 22:43 +0000, Chuck Guzis via cctalk wrote:
On 1/14/25 14:03, Van Snyder via cctalk wrote:
I corresponded with X3J3 starting in about 1980, but had no funding to
participate officially. From 1997 until I retired in 2020, I was my
employer's representative to J3, as INCITS renamed X3J3 when they took
over from CBEMA, and then PL22.3, as they renamed it to correspond to
ISO nomenclature. I was also my employer's representative to ISO/IEC
JTC1/SC22/WG5. I was liaison from J3/PL22.3 to WG9 (Ada), IEEE P1722
(interval arithmetic), and IFIP WG 2.5 (numerical software), of which I
was vice-chair for about five years until I retired.
Fortran 2025, the sixth edition, is rather
different from 1956.
F90 was a huge break from earlier versions; the target release date,
IIRC was supposed to be 1988, but there was a great amount of
negotiation. I was an alternate on our firm's representative to X3J3
for vector extensions. I recall that, at one point, both IBM and DEC
threatened to walk out of X3J3 over the committee's not choosing to
use
VECTRAN syntax.
I haven't kept up with releases after F90,
unfortunately. I can say
that F90 bears little resemblance to F77.
F90 was an extension to F77 and was entirely upwardly compatible with
it, not an entirely new language. The most important syntax extension
was free-form input to get away from the rigid card-oriented format,
but vector operations, dynamic memory, modules, etc. were also very
important. Almost all of F66 and F77 are still present in F2025. The
column-1 vertical format control was removed in 2003 because there was
no way to detect or specify that it was enabled for a particular file.
Assigned GO TO and assigned FORMAT and the ASSIGN statement were
removed in 2008. A few things are marked as "obsolescent" with a caveat
that they might be removed, but the vendors say "go ahead and remove
it, but I still have to compile it forever."
--Chuck