Re: List of software programs written in fortran (for engineers and scientists)

<nmm1@xxxxxxxxx> wrote in message news:h5ug3r$urn$1@xxxxxxxxxxxxxxxxxxxxxxxxxxxx
In article <NCygm.11240$ze1.6109@xxxxxxxxxxxxxxxxxxxxxxxxxx>,
robin <robin_v@xxxxxxxxxxx> wrote:

These sorts of programs were originally written in languages
other than FORTRAN.

Eh? Some were, some weren't. Fortran was THE implementation language
of choice for portable scientific applications (including statistical
and engineering ones) between 1960 and 1980. It was pretty heavily
used for non-portable ones, too.

That wasn't what the OP asked.

No, it wasn't.

That's right. It wasn't.

I was responding to your categorical statement, which
you may have made rather more absolute than you intended.

Let me say it again. Those sorts of programs
were written originally in languages other than Fortran.

Do you think, for instance, that the early algorithms that appeared
in CACM were the first instances of those algorithms?
Of course not; they were written in machine code or some other
computer language five to ten years prior. Some were implementations of
algorithms that had been known for decades and solved by hand
computation.

In any case, ALGOL was the language of choice for UK and Europe.

Sigh. Do look at my Email address. No, it wasn't. Many people
used ALGOL 60, many used Fortran, and many used other languages.

And some wrote their own Algol compiler.
Not language of choice?

You should note that early British and European
computers in the public eye used mainly Algol and languages
other than FORTRAN. Even in the US that was the case.

In the 1960s (but no later), many people used Autocode, but
abandoned it because it wasn't portable.

It was portable to other machines of the same kind.
But autocodes were generally primitive, and were
abandoned not because they weren't portable, but because
some better language came along.

Most of Rothamstead's

Rothamsted

and RAL's codes were in Fortran between those dates, to name
only two important sites.

PL/I also was the language of choice.

Not for portable applications.

Yes, for portable applications. Also for general algorithm design,
like Algol.
You're forgetting that IBM's PL/I SSP was able to run on
any PL/I compiler, including their DOS subset.

You are also forgetting that FORTRAN codes were no more
portable than a log. Apart from the need to change the code
to run on another model of the same machine, porting to
another manufacturer involved removing/changing extensions,
in particular changing the way subroutines handled arrays,
changing any character handling, and the like.

In the statistical field, there were
many dozens of Fortran ones, and the very few PL/I ones (SAS etc.)
were nor very portable or even specific to a single system. That
also seemed to be true in engineering (ICES etc.) and in the other
'scientific' areas I had some contact with.

and the applications would be regarded as minor utilities
nowadays.

Nonsense. They were used for structural design, aircraft design,
nuclear design, and the like.

Those computers had 1-4 KB of memory.

The amount of memory a system had (or has)
has very little to do with the complexity of algorithn that can be written.

In comparison with current machines, memories then were small.
But codes were compact.

Yes, they were used for those
purposes (I did so on a Mercury Meteor), but the applications were
what would be regarded as minor utilities nowadays.

I repeat that you are wrong.
Programs for solving engineering problems used OO programming.
In any case I know of at least one problem required the solving
of more than 1,000 simultaneous equations.
One compiler occupied 8KB, such was the compact code
of some early systems.-
Important sites in the UK included UKAEA,
Wharton (Atomic Power Division of English Electric),
Central Electricity Generating Board (CEGB),
RAE at Farnborough; British Aero Engine Co.;
BP, Ministry of Agriculture, Fisheries and Food,
and of course National Physical Laboratory --
to mention a few of the sites where DEUCE was installed.

Why were those important? They were are the forefront
of engineering and scientific development. Pilot ACE
was one of the fastest computers (if not the fastest) when it
was introduced, and DEUCE (re-engineered Pilot ACE)
had the same clock speed as Pilot ACE.*)

Some even predate any implementation of FORTRAN.

The number of such things written before 1956 (which is when I believe
the first implementation of Fortran was released) was pretty damn
small,

Programs for solving engineering and numerical problems
were written and solved on Pilot ACE prior to 1955,
which date precedes any implementation of FORTRAN.
These programs were used extensively on DEUCE
and ACE.
Much of the numerical work was pioneered by J. H. Wilkinson.
For DEUCE, the published programs numbered in excess
of 1,000 in all fields.

Considering that the first DEUCE computer was delivered in May 1955,
I doubt that many of those predated Fortran.

That's irrelevant, as you well know (and I didn't say that anyway).
The programs were originally written for Pilot ACE in the early 1950s
(i.e., pre 1955) and continued on DEUCE.
The software for Pilot ACE was developed from the
late 1940s.
All of that predates FORTRAN, does it not!

As for DEUCE, the first machines (plural) were
delivered in 1955, and came with a large library
of published code inherited from Pilot ACE.

____________
* but faster because it has some better instructions
and doubled I/O capacity.


.

Relevant Pages