27 Jun
2011
27 Jun
'11
9:26 a.m.
On 27 Jun 2011 at 8:35, Dan Gahlinger wrote:
Vax fortran in the 70s was kind of an interesting
time,some of it was
Fortran-IV, some of it was Fortran-77.
This was a big bone of contention with X3J3 when Fortran 8x (became
Fortran 90 because of some of the donnybrooks--it was supposed to
have been Fortran-88). Is the role of the ANSI working group to
invent language or to certify vendors' extensions to the language?
Should X3J3 prohibit the addition of extensions if a vendor wanted to
claim ANSI conformance? What should the compiler default be?
(CODASYL had gone through the same with COBOL 75 and got to be
pretty draconian about extensions to the "standard". But then COBOL
was a real mess before 75).
Both IBM and DEC tried to throw their weight around by threatening
to withdraw from X3J3 on some of these issues.
So those dialects based on FORTRAN 66 and FORTRAN II could be wildly
different. A vendor was supposed to be able to support at a minim
something called "USA Basic FORTRAN", which was a very small subset
(no logical IF statements, etc.) to claim conformance. But
extensions abounded, at least one of which lent itself to ambiguous
interpretation--and they were all different--even among compilers
from the same vendor.
N.B. It's "FORTRAN-77", but "Fortran-90" or
"FORTRAN-90".
--Chuck