Re: A question for my personal knowledge.

["William B. Clodius" <wclodius@lanl.gov>]

bglbv@my-dejanews.com wrote:
> <snip>
> 
> That would have been an interesting scoop. The current Fortran is in
> fact Fortran 95. (It was approved in 1995 November, so Ada 95 is a few
> months older.)
> <snip>

The draft that became Fortran 95 was submitted for public review in
November 95. It passed the nominal review process with minor comments
six months later. However, both correcting one minor comment
(translitterating the title into French did not work which led to a
change in the English title) and getting lost in the ISO beaurocracy,
put off official approval until late spring 1997, and publication until
late 1997.

> Work is in progress to make Fortran 2000 suitable for serious object-oriented
> programming. (How successfully? Wait and see... Some of us are already
> unhappy that neither generics nor exceptions will make it into Fortran this
> time around.)

A committee of eleven unpaid (except for a few vendors's
representatives) part time workers (and one unpaid part time editor that
is not a committee member) has to put their priorities somewhere. If you
want to change their priorities get involved. Another four or five
(competant) members, or ten observers, would do wonders.

The 8x committees apparently put significant time into examining
exception handling, and the Fortran 95 committees put in about two years
(late 94 through early 96) with no significant progress. While a
majority of the F95 committees wanted exception handling, they were
unable to arrive at a concensus on the details of the syntax and
semantics. In the absence of a strong defacto implementation standard,
or a leader to decide things arbitrarily, there will be no true
exception handling in Fortran. In the end they did agree on a floating
point exception model, so at least they got in the most important needs
for their user community, albeit too late for F95 itself.

In 1995/96, when most of the official planning for Fortran 2000 occurred
there was *NO* significant public demand for generics (unless you count
one person as significant.) That, of course, changed by late 1997.

As it is the current committee has trouble handling all it committed to
in early 97, e.g., C interoperability, object orientation, derived type
I/O, polishing floating point exceptions.  The work needs more people to
review it for minor problems, but in their absence that task is getting
dumped on the editor who is very competant but overwhelmed.

The primary worker on object orientation, Malcolm Cohen, is also very
competent. He wrote the first Fortran 90 compiler, apparently unaided,
in one year and understands every aspect of the language. With a little
more help he could do a great job, but the help is almost absent.

http://www.ionet.net/~jwagener/j3/