4 Feb
2019
4 Feb
'19
8:16 a.m.
On 2/3/19, Clem Cole <clemc(a)ccc.com> wrote:
On Sun, Feb 3, 2019 at 5:14 PM Henry Bent
<henry.r.bent(a)gmail.com> wrote:
The de facto standard for Fortran in the 1970s was IBM Fortran IV.
All of the important academic packages, both in the physical sciences
and the stats packages (SPSS, BMDP, etc.) used in Economics,
Psychology, Sociology, etc., were written in Fortran IV. The ANSI
(later ISO) standards committee was more or less an irrelevancy.
Vendors added their own proprietary extensions to Fortran IV. With
the VAX DEC was able to supplant IBM in the educational/research
marketplace by offering mainframe-level performance at a fraction of
the IBM price. VAX Fortran, and its extensions to Fortran IV, thus
became the new de facto standard in the education/research market in
the early 1980s. VAX Fortran did eventually implement all of the
features of Fortran 77, but for quite a long time we didn't bother
implementing the missing pieces because there was no market demand for
them. As Clem said, the important VAX Fortran extensions to Fortran
IV have over time made their way into the ISO standard.
-Paul W.
This code was apparently so pervasive and
long-lived that the GNU Fortran
compiler added compatibility for DEC extensions less than two years ago,
in
version 7. There must be enough demand for DEC's additions to have made
it
worthwhile.
Truth is most of the important ones went into Fortran-90 if I understand it
correctly (I'd trust Paul W.s comments if he knows). Again, I'm not a
compiler guy, but I've been known to eat lunch with a few of them :-)