19 Mar
2018
19 Mar
'18
4:46 p.m.
arrgh -- dyslexia -- VT-100's are NOT full ansi [they use the ANSI
sequences, but do not implement all of the features/behaviors in the
spec]. VMS Fortran started the same way, although it did conform in time
because it had to pass the Fortran validation tests.
ᐧ
On Mon, Mar 19, 2018 at 11:43 AM, Clem Cole <clemc(a)ccc.com> wrote:
On Mon, Mar 19, 2018 at 10:50 AM, Dan Cross <crossd(a)gmail.com> wrote:
I'm curious at the FORTRAN effort: what was
that about, where did it come
from, and why was it abandoned?
I'll let Ken, Steve or Doug answer definitively that but I would suspect
- it is a lot of work and at time t0, it was less valuable than some of the
other efforts going at the time.
Second, 7th Edition came with the "f77" command implementing
(unsurprisingly) Fortran 77. A paper by Stu Feldman and Peter Weinberger in
Volume 2 describes the compiler and includes this line: "This is believed
to be the first complete Fortran 77 system to be implemented." (
https://s3.amazonaws.com/plan9-bell-labs/7thEdMan/vol2/f77.txt)
Mumble, although probably true in absolute fact. The DEC VAX/VMS Fortran
compiler was contemporary. I have always said that the best piece of work
DEC Marketing ever did was convince the world that VMS Fortran was F77 (it
was not). It ended up being a super-set, although it did not start out
that way (similar to people believing VT-100's are ANSI - they are and in
that case never did fully conform).
Was that true? Notable in this paper is mention that the Fortran compiler
can drive the backend of either Ritchie's PDP-11 C compiler *or* Johnson's
portable C compiler. What was the local story? Did this see local use?
I used the PCC/VAX version extensively (as well as ratfor) for the 'users
space' part of my thesis work. My housemate, Tom Quarles, had developed
SPICE3 (in C) and Ellis Cohen had written SPICE2 in FORTRAN. FPS had done
a great deal of development on the array processor that was the basis for
my work, all in Ratfor but assuming VMS under the coveres (they wrote an
optimizing parallel Fortran compiler in same -- those guys are now the
Portland Compiler Group).
I worked, although moving stuff from VMS to BSD was huge because F77 !=
VMS Fortran. Much of the 'grunt' work I had was making all that work. In
fact, it was this work that I found a bug in the C compiler runtimes, that
I have written about elsewhere. The ratfor code called F77, which shared
C's runtime.
ᐧ