What's Fortran?!?!

[J. Clarke]

I feel old.

It occurred to me today to find out if my employer had an officially
blessed Fortran environment for the PC.

There wasn't anything in the site for software orders and there has
been a recent redesign so I couldn't find the general request.

So I called the help desk and asked. First response--"what is Fortran
and does it have another name?". Ultimately she found where the
general request had been moved to and I sent that off, so we'll see.

But this person, who provides IT support for a company that has
multiple mainframes, has never even heard of Fortran. Am I that much
of an antique?

[Quadibloc]

Post by J. Clarke
So I called the help desk and asked. First response--"what is Fortran
and does it have another name?". Ultimately she found where the
general request had been moved to and I sent that off, so we'll see.
But this person, who provides IT support for a company that has
multiple mainframes, has never even heard of Fortran. Am I that much
of an antique?

I will admit that I would be surprised, just like you were, initially.

After all, I remember that back in the days of 8-bit computers, enough people
remembered Fortran that there was Nevada Fortran made to run under CP/M, for
example.

GCC supports Fortran as well as C. Also, the community edition of a
professional Fortran compiler that runs under Windows is available for free
download from Nvidia.

But while Fortran is not dead, and it can be found _if you go looking for it_,
it has not been "where the action is" for quite some time. If you get into
computers today, you'll enocounter C and Pascal and Python and Ruby on Rails,
Java, JavaScript (aka ECMAScript), and maybe even Perl (although that's
starting to be an antique too)... but unless you have a personal interest in
the history of computing, you may never bump into Fortran, PL/I, or COBOL.
COBOL still gets used in its application area - and Fortran is still used to
program supercomputers - but if you're not working in those specific fields,
you indeed may never encounter them.

I recently learned that although IBM's current Fortran compiler for the system
Z is a Fortran 77 compiler, it actually has been updated to include a lot of
Fortran 90 features, even if it doesn't fully meet the Fortran 90 spec. But
the days when Fortran was used for all sorts of stuff because we didn't have
anything else handy are long gone. C is in that position now, along with C++
of course.

John Savard

[Quadibloc]

Post by Quadibloc
I recently learned that although IBM's current Fortran compiler for the system
Z is a Fortran 77 compiler, it actually has been updated to include a lot of
Fortran 90 features, even if it doesn't fully meet the Fortran 90 spec.

Come to think of that, this indicates that to some extent, Fortran has fallen
victim to the same thing that killed ALGOL.

Why aren't they making ALGOL compilers any more? ALGOL 60 was a nice simple
language, not terribly different from many other languages still in use today.

Ah, but if you put something in a box and call it an ALGOL compiler, people
would tend to expect that it would compile ALGOL 68 programs, since that's the
latest version of the language. Which, of course, is why we have Pascal
nowadays.

Similarly, while Fortran 77 is a ncie basic language that's easy enough to
implement, it's been succeeded by Fortran 90, Fortran 95, and Fortran 2003.
These languages are especially designed to be useful in writing parallel
programs to run on supercomputers. If you want a language for doing numerical
analysis on an ordinary computer, Pascal is a possible choice; today's Fortran
is overkill.

John Savard

[Peter Flass]

Post by Quadibloc

Post by Quadibloc
I recently learned that although IBM's current Fortran compiler for the system
Z is a Fortran 77 compiler, it actually has been updated to include a lot of
Fortran 90 features, even if it doesn't fully meet the Fortran 90 spec.

Come to think of that, this indicates that to some extent, Fortran has fallen
victim to the same thing that killed ALGOL.
Why aren't they making ALGOL compilers any more? ALGOL 60 was a nice simple
language, not terribly different from many other languages still in use today.
Ah, but if you put something in a box and call it an ALGOL compiler, people
would tend to expect that it would compile ALGOL 68 programs, since that's the
latest version of the language. Which, of course, is why we have Pascal
nowadays.

IMHO ALGOL 68 never really caught on, at least in Leftpondia.

Post by Quadibloc
Similarly, while Fortran 77 is a ncie basic language that's easy enough to
implement, it's been succeeded by Fortran 90, Fortran 95, and Fortran 2003.
These languages are especially designed to be useful in writing parallel
programs to run on supercomputers. If you want a language for doing numerical
analysis on an ordinary computer, Pascal is a possible choice; today's Fortran
is overkill.
John Savard

[Scott Lurndal]

Post by Peter Flass

Post by Quadibloc

Post by Quadibloc
I recently learned that although IBM's current Fortran compiler for the system
Z is a Fortran 77 compiler, it actually has been updated to include a lot of
Fortran 90 features, even if it doesn't fully meet the Fortran 90 spec.

Come to think of that, this indicates that to some extent, Fortran has fallen
victim to the same thing that killed ALGOL.
Why aren't they making ALGOL compilers any more? ALGOL 60 was a nice simple
language, not terribly different from many other languages still in use today.
Ah, but if you put something in a box and call it an ALGOL compiler, people
would tend to expect that it would compile ALGOL 68 programs, since that's the
latest version of the language. Which, of course, is why we have Pascal
nowadays.

IMHO ALGOL 68 never really caught on, at least in Leftpondia.

Except of course for the thousands of Burroughs systems.

[Peter Flass]

Post by Scott Lurndal

Post by Peter Flass

Post by Quadibloc

Post by Quadibloc
I recently learned that although IBM's current Fortran compiler for the system
Z is a Fortran 77 compiler, it actually has been updated to include a lot of
Fortran 90 features, even if it doesn't fully meet the Fortran 90 spec.

Come to think of that, this indicates that to some extent, Fortran has fallen
victim to the same thing that killed ALGOL.
Why aren't they making ALGOL compilers any more? ALGOL 60 was a nice simple
language, not terribly different from many other languages still in use today.
Ah, but if you put something in a box and call it an ALGOL compiler, people
would tend to expect that it would compile ALGOL 68 programs, since that's the
latest version of the language. Which, of course, is why we have Pascal
nowadays.

IMHO ALGOL 68 never really caught on, at least in Leftpondia.

Except of course for the thousands of Burroughs systems.

You used ALGOL 68? I guess the last time I used Burroughs it was 60. My
understanding is that 68 is quite a different language than 60.

[Ahem A Rivet's Shot]

On Wed, 19 Feb 2020 17:40:14 -0700

Post by Peter Flass
You used ALGOL 68? I guess the last time I used Burroughs it was 60. My
understanding is that 68 is quite a different language than 60.

Having heard a reading of the revised report on Algol 68C I have to
agree. Algol 60 is a nice simple, expressive language excellently suited to
its purpose. Algol 68 is none of the above. Even thinking about it makes me
stroppy.

[Bob Eager]

Post by Ahem A Rivet's Shot

Post by Peter Flass
You used ALGOL 68? I guess the last time I used Burroughs it was 60. My
understanding is that 68 is quite a different language than 60.

Having heard a reading of the revised report on Algol 68C I have to
agree. Algol 60 is a nice simple, expressive language excellently suited
to its purpose. Algol 68 is none of the above. Even thinking about it
makes me stroppy.

LOL!

The problem with ALGOL 60 was a lack of definition for I/O (transput in
ALGOL 68 terms). There were many different implementations.

Not to mention stropping conventions.

[Peter Flass]

Post by Bob Eager

Post by Ahem A Rivet's Shot

Post by Peter Flass
You used ALGOL 68? I guess the last time I used Burroughs it was 60. My
understanding is that 68 is quite a different language than 60.

Having heard a reading of the revised report on Algol 68C I have

to

Post by Ahem A Rivet's Shot
agree. Algol 60 is a nice simple, expressive language excellently suited
to its purpose. Algol 68 is none of the above. Even thinking about it
makes me stroppy.

LOL!
The problem with ALGOL 60 was a lack of definition for I/O (transput in
ALGOL 68 terms). There were many different implementations.
Not to mention stropping conventions.

That’s what I don’t understand. I would think that, in the absence of a
standard, someone writing a new compiler would try to model their I/O
facilities on an existing one. Why re-invent the wheel?

[Bob Eager]

Post by Peter Flass

Post by Bob Eager

Post by Ahem A Rivet's Shot

Post by Peter Flass
You used ALGOL 68? I guess the last time I used Burroughs it was 60.
My understanding is that 68 is quite a different language than 60.

Having heard a reading of the revised report on Algol 68C I have

to

Post by Ahem A Rivet's Shot
agree. Algol 60 is a nice simple, expressive language excellently
suited to its purpose. Algol 68 is none of the above. Even thinking
about it makes me stroppy.

LOL!
The problem with ALGOL 60 was a lack of definition for I/O (transput in
ALGOL 68 terms). There were many different implementations.
Not to mention stropping conventions.

That’s what I don’t understand. I would think that, in the absence of a
standard, someone writing a new compiler would try to model their I/O
facilities on an existing one. Why re-invent the wheel?

Becuase when ALGOL 58/60 were being defined, the only common model was
FORTRAN. "If I wanted to get there, I wouldn't start from here".

And in the case of ALGOL 68, theer was no defined mnodel from ALGOL 60.
What they defined was at least a definition!

[Neil Thompson]

ICL's S3 system programming language was largely based on ALGOL 68 - with the possible exception of 68000 assembler it's probably my favourite of all the languages I've seriously programmed in (can you call 68000 assembler a language?). It felt like it had a flow and a freedom of expression I've not had in any other language - yeah, I know that sounds sappy, but it is what it is, it agreed with my mindset.

Post by Bob Eager

Post by Peter Flass

Post by Bob Eager

Post by Ahem A Rivet's Shot

Post by Peter Flass
You used ALGOL 68? I guess the last time I used Burroughs it was 60.
My understanding is that 68 is quite a different language than 60.

Having heard a reading of the revised report on Algol 68C I have

to

Post by Ahem A Rivet's Shot
agree. Algol 60 is a nice simple, expressive language excellently
suited to its purpose. Algol 68 is none of the above. Even thinking
about it makes me stroppy.

LOL!
The problem with ALGOL 60 was a lack of definition for I/O (transput in
ALGOL 68 terms). There were many different implementations.
Not to mention stropping conventions.

That’s what I don’t understand. I would think that, in the absence of a
standard, someone writing a new compiler would try to model their I/O
facilities on an existing one. Why re-invent the wheel?

Becuase when ALGOL 58/60 were being defined, the only common model was
FORTRAN. "If I wanted to get there, I wouldn't start from here".
And in the case of ALGOL 68, theer was no defined mnodel from ALGOL 60.
What they defined was at least a definition!
--
Using UNIX since v6 (1975)...
http://www.mirrorservice.org

[Bob Eager]

Post by Neil Thompson
ICL's S3 system programming language was largely based on ALGOL 68 -
with the possible exception of 68000 assembler it's probably my
favourite of all the languages I've seriously programmed in (can you
call 68000 assembler a language?). It felt like it had a flow and a
freedom of expression I've not had in any other language - yeah, I know
that sounds sappy, but it is what it is, it agreed with my mindset.

One of my favourites, again used to write an operating system, was
Edinburgh IMP.

https://en.wikipedia.org/wiki/Edinburgh_IMP

[Peter Flass]

Post by Bob Eager

Post by Peter Flass

Post by Bob Eager

Post by Ahem A Rivet's Shot

Post by Peter Flass
You used ALGOL 68? I guess the last time I used Burroughs it was 60.
My understanding is that 68 is quite a different language than 60.

Having heard a reading of the revised report on Algol 68C I have

to

Post by Ahem A Rivet's Shot
agree. Algol 60 is a nice simple, expressive language excellently
suited to its purpose. Algol 68 is none of the above. Even thinking
about it makes me stroppy.

LOL!
The problem with ALGOL 60 was a lack of definition for I/O (transput in
ALGOL 68 terms). There were many different implementations.
Not to mention stropping conventions.

That’s what I don’t understand. I would think that, in the absence of a
standard, someone writing a new compiler would try to model their I/O
facilities on an existing one. Why re-invent the wheel?

Becuase when ALGOL 58/60 were being defined, the only common model was
FORTRAN. "If I wanted to get there, I wouldn't start from here".

Sure, but why wouldn’t the people who wrote the second, third, etc.
compilers think about compatibility with the first. Was that I/O model that
bad?

Post by Bob Eager
And in the case of ALGOL 68, theer was no defined mnodel from ALGOL 60.
What they defined was at least a definition!

[Bob Eager]

Post by Peter Flass

Post by Bob Eager

Post by Peter Flass

Post by Bob Eager

On Wed, 19 Feb 2020 17:40:14 -0700 Peter Flass

Post by Peter Flass
You used ALGOL 68? I guess the last time I used Burroughs it was 60.
My understanding is that 68 is quite a different language than 60.

Having heard a reading of the revised report on Algol 68C I have

to

agree. Algol 60 is a nice simple, expressive language excellently
suited to its purpose. Algol 68 is none of the above. Even thinking
about it makes me stroppy.

LOL!
The problem with ALGOL 60 was a lack of definition for I/O (transput
in ALGOL 68 terms). There were many different implementations.
Not to mention stropping conventions.

That’s what I don’t understand. I would think that, in the absence of
a standard, someone writing a new compiler would try to model their
I/O facilities on an existing one. Why re-invent the wheel?

Becuase when ALGOL 58/60 were being defined, the only common model was
FORTRAN. "If I wanted to get there, I wouldn't start from here".

Sure, but why wouldn’t the people who wrote the second, third, etc.
compilers think about compatibility with the first. Was that I/O model
that bad?

Different terams, different places, different companies, different
countries. Lock-in probably.

[Andy Walker]

Post by Peter Flass

Post by Bob Eager

Post by Peter Flass
That’s what I don’t understand. I would think that, in the absence of a
standard, someone writing a new compiler would try to model their I/O
facilities on an existing one. Why re-invent the wheel?

[Bob Eager:]

Post by Peter Flass

Post by Bob Eager
Becuase when ALGOL 58/60 were being defined, the only common model was
FORTRAN. "If I wanted to get there, I wouldn't start from here".

Sure, but why wouldn’t the people who wrote the second, third, etc.
compilers think about compatibility with the first. Was that I/O model that
bad?

In Algol 60 there virtually was no I/O model at all. It was
expected that there would be some way of reading/writing numbers and
of printing messages. That was it. At least in the UK, there was
almost no thought given to compatibility of any sort. You might be
able to take a deck of cards or a reel of paper tape from "your"
computer to the nearest other computer, perhaps 100 miles away, but
you had no expectation that they would have the same card reader,
or the same paper tape equipment, or a compatible lineprinter, and
certainly not compatible job control [no OS worthy of the name in
1958-60].

All that was the sort of thing we began to think about only
when we got our *second* computer; so much better than the old one,
but it meant we had to throw away all our existing programs and 5+
years of work and start again. You don't make that mistake twice,
which is why the big companies, producing tens, hundreds or even
thousands of computers, started making them in ranges so that you
could start with one and upgrade to a bigger or later model without
having to re-write everything.

[Andy Walker]

[...] Algol 60 is a nice simple, expressive language excellently suited to
its purpose.

It was nice; not *that* simple [there were all sorts of pitfalls
of implementation]; not *that* expressive [inadequate I/O, no character
handling, no data structures]; and of rather limited purpose [pretty
much the writing of algorithms for numerical processes]. It was nicer
than Fortran, as were many other languages, but otherwise it was not as
efficient as [for example] Atlas Autocode [designed as a compromise
between Fortran and Algol]. It was superseded by [eg] Pascal, which also
has its faults but is better than Algol 60, but doesn't yet have as much
of a nostalgia factor.

Algol 68 is none of the above.

Algol 68R was a breath of fresh air back in 1972. Nice [what
was there not to like?], simple [the User Guide got the syntax into
a page-and-a-bit], expressive, and excellently suited to writing all
manner of programs. I don't recognise your "none of the above", tho'
that may be explicable if you indeed tried to learn it by reading the
RR on A68C instead of by reading something, well, readable.

I lost touch with A68 when we moved from mainframes to Unix
on the PDP 11, and the choice was basically C or starve. I quite
like C, but you couldn't love it. It does the job. Somewhere around
2000, I realised that I had almost stopped writing C programs; I was
writing shell scripts, Sed/Awk scripts, Troff macros, almost anything
but Real Programs. Then A68G came along, and I remembered what I'd
been missing. I can, and do, again write programs for pleasure. I
suspect most modern programmers have never known that pleasure.

[Peter Flass]

Post by Andy Walker

[...] Algol 60 is a nice simple, expressive language excellently suited to
its purpose.

It was nice; not *that* simple [there were all sorts of pitfalls
of implementation]; not *that* expressive [inadequate I/O, no character
handling, no data structures]; and of rather limited purpose [pretty
much the writing of algorithms for numerical processes].

And, with a few extensions, operating systems.

Post by Andy Walker
It was nicer
than Fortran, as were many other languages, but otherwise it was not as
efficient as [for example] Atlas Autocode [designed as a compromise
between Fortran and Algol]. It was superseded by [eg] Pascal, which also
has its faults but is better than Algol 60, but doesn't yet have as much
of a nostalgia factor.

Algol 68 is none of the above.

Algol 68R was a breath of fresh air back in 1972. Nice [what
was there not to like?], simple [the User Guide got the syntax into
a page-and-a-bit], expressive, and excellently suited to writing all
manner of programs. I don't recognise your "none of the above", tho'
that may be explicable if you indeed tried to learn it by reading the
RR on A68C instead of by reading something, well, readable.
I lost touch with A68 when we moved from mainframes to Unix
on the PDP 11, and the choice was basically C or starve. I quite
like C, but you couldn't love it. It does the job. Somewhere around
2000, I realised that I had almost stopped writing C programs; I was
writing shell scripts, Sed/Awk scripts, Troff macros, almost anything
but Real Programs. Then A68G came along, and I remembered what I'd
been missing. I can, and do, again write programs for pleasure. I
suspect most modern programmers have never known that pleasure.

[Bob Eager]

Post by Scott Lurndal

Post by Peter Flass

Post by Quadibloc

Post by Quadibloc
I recently learned that although IBM's current Fortran compiler for
the system Z is a Fortran 77 compiler, it actually has been updated
to include a lot of Fortran 90 features, even if it doesn't fully
meet the Fortran 90 spec.

Come to think of that, this indicates that to some extent, Fortran has
fallen victim to the same thing that killed ALGOL.
Why aren't they making ALGOL compilers any more? ALGOL 60 was a nice
simple language, not terribly different from many other languages
still in use today.
Ah, but if you put something in a box and call it an ALGOL compiler,
people would tend to expect that it would compile ALGOL 68 programs,
since that's the latest version of the language. Which, of course, is
why we have Pascal nowadays.

IMHO ALGOL 68 never really caught on, at least in Leftpondia.

Except of course for the thousands of Burroughs systems.

But that was nothing like ALGOL 68.

[Louis Krupp]

Post by Scott Lurndal

Post by Peter Flass

Post by Quadibloc

Post by Quadibloc
I recently learned that although IBM's current Fortran compiler for the system
Z is a Fortran 77 compiler, it actually has been updated to include a lot of
Fortran 90 features, even if it doesn't fully meet the Fortran 90 spec.

Come to think of that, this indicates that to some extent, Fortran has fallen
victim to the same thing that killed ALGOL.
Why aren't they making ALGOL compilers any more? ALGOL 60 was a nice simple
language, not terribly different from many other languages still in use today.
Ah, but if you put something in a box and call it an ALGOL compiler, people
would tend to expect that it would compile ALGOL 68 programs, since that's the
latest version of the language. Which, of course, is why we have Pascal
nowadays.

IMHO ALGOL 68 never really caught on, at least in Leftpondia.

Except of course for the thousands of Burroughs systems.

Burroughs Large Systems offered Burroughs Extended ALGOL and used it
to write compilers. It had I/O and lots of features that tied it to
the system's architecture; there were intrinsic functions that
exploited the "link list lookup" operator and the "masked search,"
among other things, and there were features that implemented software
interrupts and coroutines and so on.

I couldn't tell you if the language was based on ALGOL 60 or ALGOL 68.
I never thought about standards or interopability in those days; if it
ran on the big machine in the computer room, that was all that
mattered.

The Unisys (Burroughs + Sperry merger) MCP series still has a version
of ALGOL:
https://public.support.unisys.com/framework/publicterms.aspx?returnurl=%2faseries%2fdocs%2fClearPath-MCP-18.0%2f86000098-516%2findex.html

Louis

[John Levine]

Post by Louis Krupp
Burroughs Large Systems offered Burroughs Extended ALGOL and used it
to write compilers. It had I/O and lots of features that tied it to
the system's architecture; there were intrinsic functions that
exploited the "link list lookup" operator and the "masked search,"
among other things, and there were features that implemented software
interrupts and coroutines and so on.
I couldn't tell you if the language was based on ALGOL 60 or ALGOL 68.

Since those systems were developed starting around 1963 it'd be pretty
impressive if they used Algol 68.

[Louis Krupp]

On Sun, 23 Feb 2020 02:53:33 -0000 (UTC), John Levine

Post by John Levine

Post by Louis Krupp
Burroughs Large Systems offered Burroughs Extended ALGOL and used it
to write compilers. It had I/O and lots of features that tied it to
the system's architecture; there were intrinsic functions that
exploited the "link list lookup" operator and the "masked search,"
among other things, and there were features that implemented software
interrupts and coroutines and so on.
I couldn't tell you if the language was based on ALGOL 60 or ALGOL 68.

Since those systems were developed starting around 1963 it'd be pretty
impressive if they used Algol 68.

Burroughs always was ahead of its time...

Louis

[h***@bbs.cpcn.com]

Post by Quadibloc
I recently learned that although IBM's current Fortran compiler for the system
Z is a Fortran 77 compiler, it actually has been updated to include a lot of
Fortran 90 features, even if it doesn't fully meet the Fortran 90 spec. But
the days when Fortran was used for all sorts of stuff because we didn't have
anything else handy are long gone. C is in that position now, along with C++
of course.

I am not up on FORTRAN and don't know what IBM's official position
is on it for the Z series mainframe. But my impression (based
on conversations in this newsgroup) is that they no longer support
it for Z, but do so for other machines. Old stuff still runs.

I don't know what legacy heavy users, such as weather
forecasting on a 370-195, use today.

Note that modern IBM Z COBOL has a lot of FORTRAN features
in it, like trig functions and floating point for years.
I haven't studied it, but I'm guessing someone with a new
basic sci/eng task could do it in COBOL.

For what it's worth, engineers have long told me FORTRAN
is obsolete, replaced by other tools, such CAD/CAM.

I have no idea what the heavy-duty number crunchers use
on super computers.

[Scott Lurndal]

Post by h***@bbs.cpcn.com

Post by Quadibloc
I recently learned that although IBM's current Fortran compiler for the system
Z is a Fortran 77 compiler, it actually has been updated to include a lot of
Fortran 90 features, even if it doesn't fully meet the Fortran 90 spec. But
the days when Fortran was used for all sorts of stuff because we didn't have
anything else handy are long gone. C is in that position now, along with C++
of course.

I am not up on FORTRAN and don't know what IBM's official position
is on it for the Z series mainframe. But my impression (based
on conversations in this newsgroup) is that they no longer support
it for Z, but do so for other machines. Old stuff still runs.
I don't know what legacy heavy users, such as weather
forecasting on a 370-195, use today.

They use supercomuters and beowulf clusters (both running linux). And yes, many of the
older models are still in Fortran, but running on linux.

[Quadibloc]

Post by h***@bbs.cpcn.com
I have no idea what the heavy-duty number crunchers use
on super computers.

Fortran 2003.

John Savard

[d***@gmail.com]

Post by Quadibloc

Post by h***@bbs.cpcn.com
I have no idea what the heavy-duty number crunchers use
on super computers.

Fortran 2003.
John Savard

I worked with some supercomputer customers some years ago, and was surprised at how many massive FORTRAN programs they were using. and how little my FORTRAN IV experienced helped... FORTRAN today is very different. And the compilers these days are very sophisticated. But, unless you're in one of those fields, FORTRAN is really not something you'd ever hear about.

[Quadibloc]

Post by d***@gmail.com

Post by Quadibloc

Post by h***@bbs.cpcn.com
I have no idea what the heavy-duty number crunchers use
on super computers.

Fortran 2003.

I worked with some supercomputer customers some years ago, and was surprised
at how many massive FORTRAN programs they were using. and how little my
FORTRAN IV experienced helped... FORTRAN today is very different. And the
compilers these days are very sophisticated. But, unless you're in one of
those fields, FORTRAN is really not something you'd ever hear about.

Yes, you have described the current situation very accurately, in case anyone
might have doubts: and what you have pointed out are precisely its most salient
features.

John Savard

[d***@gmail.com]

On Thursday, February 20, 2020 at 5:31:54 PM UTC-6, Quadibloc wrote:
...

Post by Quadibloc
Yes, you have described the current situation very accurately, in case anyone
might have doubts: and what you have pointed out are precisely its most salient
features.
John Savard

It was interesting to look back at the history of FORTRAN recently. It originally was created to dove-tail with the earliest computers machine language, a lot of FORTRAN statements would equate to a single instruction on those old computers (like the old tri-nary IF statement). Considering it is still in use today (albeit much more advanced) is a testament to it's power. Is there any other language that has that long of a lifespan? What is it, almost 70 years now?

[J. Clarke]

Post by d***@gmail.com
...

Post by Quadibloc
Yes, you have described the current situation very accurately, in case anyone
might have doubts: and what you have pointed out are precisely its most salient
features.
John Savard

It was interesting to look back at the history of FORTRAN recently. It originally was created to dove-tail with the earliest computers machine language, a lot of FORTRAN statements would equate to a single instruction on those old computers (like the old tri-nary IF statement). Considering it is still in use today (albeit much more advanced) is a testament to it's power. Is there any other language that has that long of a lifespan? What is it, almost 70 years now?

You mean besides COBOL?

[Peter Flass]

Post by d***@gmail.com
...

Post by Quadibloc
Yes, you have described the current situation very accurately, in case anyone
might have doubts: and what you have pointed out are precisely its most salient
features.
John Savard

It was interesting to look back at the history of FORTRAN recently. It
originally was created to dove-tail with the earliest computers machine
language, a lot of FORTRAN statements would equate to a single
instruction on those old computers (like the old tri-nary IF statement).
Considering it is still in use today (albeit much more advanced) is a
testament to it's power. Is there any other language that has that long
of a lifespan? What is it, almost 70 years now?

Since FORTRAN was the first HLL (AFAIK), it’s probably the only one. COBOL
ALGOL, and PL/I are next.

[Quadibloc]

Post by Peter Flass
Since FORTRAN was the first HLL (AFAIK), it’s probably the only one. COBOL
ALGOL, and PL/I are next.

FORTRAN doesn't have _that_ distinction.

However, higher-level languages which preceded FORTRAN, while not interpreted
like BASIC, were much slower than machine code. For example, they might be
implemented in the form of a series of subroutine calls instead of machine
instructions.

FORTRAN was originally implemented on the IBM 704. This machine had hardware
floating-point. In order to make widespread adoption of FORTRAN a reasonably
likely possibility in an era when computer time cost hundreds of dollars per
hour, a great deal of effort was made to make code generated by the original IBM
704 FORTRAN as efficient as handwritten assembler.

So FORTRAN was the first widely popular higher-level language, and the first
higher-level language with an efficient implementation.

John Savard

[h***@bbs.cpcn.com]

Post by Quadibloc

Post by Peter Flass
Since FORTRAN was the first HLL (AFAIK), it’s probably the only one. COBOL
ALGOL, and PL/I are next.

FORTRAN doesn't have _that_ distinction.
However, higher-level languages which preceded FORTRAN, while not interpreted
like BASIC, were much slower than machine code. For example, they might be
implemented in the form of a series of subroutine calls instead of machine
instructions.
FORTRAN was originally implemented on the IBM 704. This machine had hardware
floating-point. In order to make widespread adoption of FORTRAN a reasonably
likely possibility in an era when computer time cost hundreds of dollars per
hour, a great deal of effort was made to make code generated by the original IBM
704 FORTRAN as efficient as handwritten assembler.
So FORTRAN was the first widely popular higher-level language, and the first
higher-level language with an efficient implementation.

While I believe a goal of FORTRAN was to create reasonably
efficient executable code, I believe handwritten assembler
was still more efficient.

In 709/7090 days, I wonder the proportion of sci/eng programs
written in Fortran vs. assembler. I can't help but suspect
(based on the books out there) that assembler certainly held
its own in the old days, maybe the majority.

Many former 1401 programmers told me COBOL and FORTRAN
were just too slow, including the compilation effort,
and so were rarely used on them.

FORTRAN was developed by IBM, but I think it was open to all.
Would anyone know when the other computer makers developed
their own FORTRAN compilers and got wide use?

My impression is that FORTRAN really didn't take off until
S/360 days. Also, my impression (and experience) was that
FORTRAN became popular among engineers and scientists who
could then write their own programs--it wasn't as complex
as assembler. One complex arithmetic statement with
functions would require a lot of hand coding.

[Quadibloc]

Post by h***@bbs.cpcn.com
While I believe a goal of FORTRAN was to create reasonably
efficient executable code, I believe handwritten assembler
was still more efficient.
In 709/7090 days, I wonder the proportion of sci/eng programs
written in Fortran vs. assembler. I can't help but suspect
(based on the books out there) that assembler certainly held
its own in the old days, maybe the majority.
Many former 1401 programmers told me COBOL and FORTRAN
were just too slow, including the compilation effort,
and so were rarely used on them.

IBM made many different FORTRAN compilers as it made many different machines.

And it's certainly true that the time required to compile a program is of
importance, since not all programs are intended to run for a long time.

IBM's original FORTRAN for the 704 was designed to come close to handwritten
assembler in performance. The compiler performed a number of optimizations, as a
result sometimes producing code faster than would have been written by hand.

Their FORTRAN II compiler was closely based on the original FORTRAN, despite
adding a number of vital features that compiler lacked, like subroutine calls.
So it, too, was a highly optimizing compiler.

IBM's FORTRAN IV compiler for the descendants of the 704, such as the 7090, on
the other hand, was not an optimizing compiler. It took less time to compile
programs, but the resulting code was not as highly efficient.

FORTRAN for the 1401 - and, indeed, most other IBM mainframes of those days -
usually wasn't implemented with a highly optimizing compiler either.

It wasn't until the IBM System/360 that one had a choice of plain or optimizing
compilers - FORTRAN IV Level G and FORTRAN IV Level H - for the same language on
the same computer.

John Savard

[Charlie Gibbs]

Post by Quadibloc
IBM made many different FORTRAN compilers as it made many different machines.
And it's certainly true that the time required to compile a program is of
importance, since not all programs are intended to run for a long time.

On the other hand, for programs that are used in daily production
the extra time spent by a good optimizing compiler is paid for
many times over.

<snip>

Post by Quadibloc
It wasn't until the IBM System/360 that one had a choice of plain or
optimizing compilers - FORTRAN IV Level G and FORTRAN IV Level H -
for the same language on the same computer.

In a school environment, where most programs were small and compiled
many times (during debugging) but executed only a few times, the
other choice was WATFOR. Perhaps execution was not too efficient,
but compiles were lightning-fast.

[Peter Flass]

Post by Charlie Gibbs

Post by Quadibloc
IBM made many different FORTRAN compilers as it made many different machines.
And it's certainly true that the time required to compile a program is of
importance, since not all programs are intended to run for a long time.

On the other hand, for programs that are used in daily production
the extra time spent by a good optimizing compiler is paid for
many times over.
<snip>

Post by Quadibloc
It wasn't until the IBM System/360 that one had a choice of plain or
optimizing compilers - FORTRAN IV Level G and FORTRAN IV Level H -
for the same language on the same computer.

In a school environment, where most programs were small and compiled
many times (during debugging) but executed only a few times, the
other choice was WATFOR. Perhaps execution was not too efficient,
but compiles were lightning-fast.

FORTRAN H was also pretty buggy initially.

[Quadibloc]

Post by Charlie Gibbs

Post by Quadibloc
It wasn't until the IBM System/360 that one had a choice of plain or
optimizing compilers - FORTRAN IV Level G and FORTRAN IV Level H -
for the same language on the same computer.

In a school environment, where most programs were small and compiled
many times (during debugging) but executed only a few times, the
other choice was WATFOR. Perhaps execution was not too efficient,
but compiles were lightning-fast.

I learned how to program in FORTRAN with WATFIV, so I should have mentioned it
as a third step in the chain - but I was looking at the history of IBM products
in this area, and surveying more than one hardware system.

John Savard

[h***@bbs.cpcn.com]

Post by Charlie Gibbs
In a school environment, where most programs were small and compiled
many times (during debugging) but executed only a few times, the
other choice was WATFOR. Perhaps execution was not too efficient,
but compiles were lightning-fast.

We used WATFOR. My compsci prof said that while WATFOR was
intended as a quickie student compiler, it actually generated
reasonably efficient code.

[h***@bbs.cpcn.com]

Post by Quadibloc

Post by h***@bbs.cpcn.com
In 709/7090 days, I wonder the proportion of sci/eng programs
written in Fortran vs. assembler. I can't help but suspect
(based on the books out there) that assembler certainly held
its own in the old days, maybe the majority.
Many former 1401 programmers told me COBOL and FORTRAN
were just too slow, including the compilation effort,
and so were rarely used on them.

IBM made many different FORTRAN compilers as it made many different machines.

Which was one of its motivations for creating S/360--a single
product line so it'd need less software. Indeed, IBM was drowning
in the need for compilers, utilities, and operating systems for
its pre-S/360 product line.

Of course, S/360 still required numerous variations of compilers.
Just in 360-DOS we had the D and F version of COBOL (D was
free, F was rented for a slight fee. F was a lot better).

Post by Quadibloc
And it's certainly true that the time required to compile a program is of
importance, since not all programs are intended to run for a long time.

One problem colleges had (discussed here before) was the overhead
to start a batch program was high. Sometimes schools would
clump together a bunch of student jobs which would save a lot
of time.

[snip]

[h***@bbs.cpcn.com]

Post by Quadibloc
IBM's FORTRAN IV compiler for the descendants of the 704, such as the 7090, on
the other hand, was not an optimizing compiler. It took less time to compile
programs, but the resulting code was not as highly efficient.

Given the importance of efficient code, why wasn't the 7090
compiler not optimizing?


In COBOL, one can set the optimization switch on/off. On
complex programs handling large volumes of data, it does
save a lot of time. But I've been told there are times
you do not want optimization on, not sure why.

[Peter Flass]

Post by h***@bbs.cpcn.com

Post by Quadibloc
IBM's FORTRAN IV compiler for the descendants of the 704, such as the 7090, on
the other hand, was not an optimizing compiler. It took less time to compile
programs, but the resulting code was not as highly efficient.

Given the importance of efficient code, why wasn't the 7090
compiler not optimizing?
In COBOL, one can set the optimization switch on/off. On
complex programs handling large volumes of data, it does
save a lot of time. But I've been told there are times
you do not want optimization on, not sure why.

One-shot where it takes longer to compile than to run
Cases where the optimizer generates bad code

[Charlie Gibbs]

Post by Peter Flass

Post by h***@bbs.cpcn.com

Post by Quadibloc
IBM's FORTRAN IV compiler for the descendants of the 704, such
as the 7090, on the other hand, was not an optimizing compiler.
It took less time to compile programs, but the resulting code
was not as highly efficient.

Given the importance of efficient code, why wasn't the 7090
compiler not optimizing?
In COBOL, one can set the optimization switch on/off. On
complex programs handling large volumes of data, it does
save a lot of time. But I've been told there are times
you do not want optimization on, not sure why.

One-shot where it takes longer to compile than to run
Cases where the optimizer generates bad code

Code is easier to debug (by single-stepping the processor
through it in those days) if an optimizer hasn't meddled
with it. Some compilers had an option to generate equivalent
assembly language code, which wouldn't correspond to the
source code nearly as well if it had been optimized.

[Scott Lurndal]

Post by h***@bbs.cpcn.com

Post by Quadibloc
IBM's FORTRAN IV compiler for the descendants of the 704, such as the 7090, on
the other hand, was not an optimizing compiler. It took less time to compile
programs, but the resulting code was not as highly efficient.

Given the importance of efficient code, why wasn't the 7090
compiler not optimizing?

Probably because it would take 24 hours to compile :-)

[Quadibloc]

Post by h***@bbs.cpcn.com
Given the importance of efficient code, why wasn't the 7090
compiler not optimizing?

Having a choice between an optimizing compiler and a non-optimizing compiler is
not a bad thing. If the intended program is not going to run for a long time,
the cost of producing a highly optimized version may not be paid for by the
savings in shorter run time. This was already noted in another reply to your
post.

I think there were two major reasons, though, that specifically applied in the
actual historical situation:

One was that the Fortran II compiler still worked on the 7090. The 709 was an
upgrade of the 704, and the 7090 was a transistor version of the 709. There were
minor incompatibilities between the 709 and 704, but IBM did not have much
difficulty adjusting the compiler to deal with them. The 7094 introduced some
other added features which again required small changes to the compilers.

The other was that Fortran IV was a significantly extended language compared to
Fortran II. So this time modifying the original Fortran compiler didn't seem
like a reasonable option.

Back to the more general question of "optimizing" versus "non-optimizing", some
other context may be helpful.

Also, with transistorized computers, now computer time was a bit cheaper - and
the initial skepticism about higher-level languages that IBM felt it needed to
overcome by showing that efficient code was possible had been overcome.

A non-optimizing compiler still generated code of reasonable quality. Before the
first IBM Fortran, higher-level languages were very slow because they didn't
compile to machine code at all, but to some type of interpretive layer - instead
of P-code, more common then was code made up of subroutine calls.

Compared to the original 704 FORTRAN, FORTRAN II added vital features, like the
subroutine call. On the other hand, FORTRAN IV added features like the logical
IF statement and CHARACTER variables. So the added features of FORTRAN IV were,
to some extent, luxuries, and if someone had serious number-crunching to do, it
was not too painful to do it in FORTRAN II.

My suspicion is that it was the Logical IF statement that made it just a little
bit too hard, at the time, to modify the original 704 FORTRAN to handle FORTRAN
IV, because it optimized programs by tracing the possible paths of execution,
and the Logical IF was just a little bit too flexible. (Admittedly, that doesn't
really make sense; either a Logical IF has a single statement it controls, or a
GO TO statement, and either way that should not be any more complex than an
Arithmetic IF.)

John Savard

[Quadibloc]

Post by Quadibloc
(Admittedly, that doesn't
really make sense; either a Logical IF has a single statement it controls, or a
GO TO statement, and either way that should not be any more complex than an
Arithmetic IF.)

I just checked the 7090 FORTRAN IV manual.

The statement in a Logical IF may not be another _logical_ IF, and it may not be
a DO statement. But it can be anything else.

Including an Arithmetic IF, a computed GO TO, or an assigned GO TO.

That still should not have made it impossible to modify the optimizing FORTRAN
II compiler to compile FORTRAN IV instead, though. It's not like they weren't
smart back then, or that it would have required some algorithm that hadn't been
invented yet.

John Savard

[Dan Espen]

Post by h***@bbs.cpcn.com

Post by Quadibloc
IBM's FORTRAN IV compiler for the descendants of the 704, such as the 7090, on
the other hand, was not an optimizing compiler. It took less time to compile
programs, but the resulting code was not as highly efficient.

Given the importance of efficient code, why wasn't the 7090
compiler not optimizing?
In COBOL, one can set the optimization switch on/off. On
complex programs handling large volumes of data, it does
save a lot of time. But I've been told there are times
you do not want optimization on, not sure why.

Usually debug tools and the optimizer don't get along.

A shop I worked in started with C/370 and went through
years of IBM trying to get their C compiler up to snuff.
During those years I found more than one compiler malfunction that went
away with the optimizer off.

That's especially dangerous if you test with the optimizer off.

Not to put everything on IBM other vendors C compilers would sometimes
incorrectly optimize code.

[John Levine]

Post by h***@bbs.cpcn.com

Post by Quadibloc
IBM's FORTRAN IV compiler for the descendants of the 704, such as the 7090, on
the other hand, was not an optimizing compiler. It took less time to compile
programs, but the resulting code was not as highly efficient.

Given the importance of efficient code, why wasn't the 7090
compiler not optimizing?

I heard a story that when they added a development mode option to the
original compiler to compile faster and produce worse code, everyone,
and I mean everyone, used it. So the second compiler only had
development mode.

When Fortran was new, the conventional wisdom was that compiler code
would always be worse than hand written code, so compilers were a
waste of time. The original compiler had to produce superb code to
get people even to try it.

Once they got used to compilers and how much easier it was to write in
Fortran than in assembler, the efficiency issue kind of got forgotten.
It probably also occurred to people that if your assembler program ran
five minutes faster than the compiled one, but the programmer spent an
extra week and a dozen test runs getting it to work, we can live with
slower code.

[Peter Flass]

Post by h***@bbs.cpcn.com

Post by Quadibloc

Post by Peter Flass
Since FORTRAN was the first HLL (AFAIK), it’s probably the only one. COBOL
ALGOL, and PL/I are next.

FORTRAN doesn't have _that_ distinction.
However, higher-level languages which preceded FORTRAN, while not interpreted
like BASIC, were much slower than machine code. For example, they might be
implemented in the form of a series of subroutine calls instead of machine
instructions.
FORTRAN was originally implemented on the IBM 704. This machine had hardware
floating-point. In order to make widespread adoption of FORTRAN a reasonably
likely possibility in an era when computer time cost hundreds of dollars per
hour, a great deal of effort was made to make code generated by the original IBM
704 FORTRAN as efficient as handwritten assembler.
So FORTRAN was the first widely popular higher-level language, and the first
higher-level language with an efficient implementation.

While I believe a goal of FORTRAN was to create reasonably
efficient executable code, I believe handwritten assembler
was still more efficient.
In 709/7090 days, I wonder the proportion of sci/eng programs
written in Fortran vs. assembler. I can't help but suspect
(based on the books out there) that assembler certainly held
its own in the old days, maybe the majority.
Many former 1401 programmers told me COBOL and FORTRAN
were just too slow, including the compilation effort,
and so were rarely used on them.

I heard that too, a least about COBOL. That’s why Autocoder was so heavily
used.
.

Post by h***@bbs.cpcn.com
FORTRAN was developed by IBM, but I think it was open to all.
Would anyone know when the other computer makers developed
their own FORTRAN compilers and got wide use?
My impression is that FORTRAN really didn't take off until
S/360 days. Also, my impression (and experience) was that
FORTRAN became popular among engineers and scientists who
could then write their own programs--it wasn't as complex
as assembler. One complex arithmetic statement with
functions would require a lot of hand coding.

I think FORTRAN was heavily used on the 709x machines. NASA had a few.

[Robert Swindells]

Post by d***@gmail.com
Considering it is still in use today (albeit much more advanced) is a
testament to it's power. Is there any other language that has that long
of a lifespan? What is it, almost 70 years now?

Lisp is a year younger than FORTRAN.

[Rich Alderson]

Post by Robert Swindells

Post by d***@gmail.com
Considering it is still in use today (albeit much more advanced) is a
testament to it's power. Is there any other language that has that long
of a lifespan? What is it, almost 70 years now?

Lisp is a year younger than FORTRAN.

If you hadn't said it, I would have.

The first implementation of LISP was actually written in FORTRAN on a 704.

[h***@bbs.cpcn.com]

Post by d***@gmail.com
It was interesting to look back at the history of FORTRAN recently. It originally was created to dove-tail with the earliest computers machine language, a lot of FORTRAN statements would equate to a single instruction on those old computers (like the old tri-nary IF statement). Considering it is still in use today (albeit much more advanced) is a testament to it's power. Is there any other language that has that long of a lifespan? What is it, almost 70 years now?

Univac had developed FLOWMATIC* in 1958 which looks a lot like COBOL.
But I don't know how widely it was utilized. One employer had a
Univac III and wrote in its assembler language ("SALT"?)

* http://bitsavers.org/pdf/univac/flow-matic/U1518_FLOW-MATIC_Programming_System_1958.pdf


Given how slow and small computers and peripherals were in
the old days, I know many old sites used assembler to
squeeze out every bit they could. For instance, it
was very common to store flags as bits, eight to a
byte, rather than as a byte containing Y or N. Assembler
programmers knew all sorts of tricks in binary to maximize
efficiency and minimize space. With a 16k or 32k machine and
complex applications, they had to squeeze it down.

Per our conversation on tapes, they had to keep tapes
down, too. An 800 bpi tape didn't hold too much. One
issue was tape mounts--that was labor intensive and
ran out the clock, so keeping a tape file small was
important. Disk space was expensive. A 1311 held only
two meg, though the 2311 improved on it, and the 2314
improved on it further. But there were lots of 1311's
still in service in the 1970s.

[Peter Flass]

Post by h***@bbs.cpcn.com

Post by d***@gmail.com
It was interesting to look back at the history of FORTRAN recently. It
originally was created to dove-tail with the earliest computers machine
language, a lot of FORTRAN statements would equate to a single
instruction on those old computers (like the old tri-nary IF statement).
Considering it is still in use today (albeit much more advanced) is a
testament to it's power. Is there any other language that has that long
of a lifespan? What is it, almost 70 years now?

Univac had developed FLOWMATIC* in 1958 which looks a lot like COBOL.
But I don't know how widely it was utilized. One employer had a
Univac III and wrote in its assembler language ("SALT"?)
* http://bitsavers.org/pdf/univac/flow-matic/U1518_FLOW-MATIC_Programming_System_1958.pdf
Given how slow and small computers and peripherals were in
the old days, I know many old sites used assembler to
squeeze out every bit they could. For instance, it
was very common to store flags as bits, eight to a
byte, rather than as a byte containing Y or N.

No reason not to do that now. Setting and testing a bit is probably as
efficient as using a byte.

Post by h***@bbs.cpcn.com
Assembler
programmers knew all sorts of tricks in binary to maximize
efficiency and minimize space. With a 16k or 32k machine and
complex applications, they had to squeeze it down.
Per our conversation on tapes, they had to keep tapes
down, too. An 800 bpi tape didn't hold too much. One
issue was tape mounts--that was labor intensive and
ran out the clock, so keeping a tape file small was
important. Disk space was expensive. A 1311 held only
two meg, though the 2311 improved on it, and the 2314
improved on it further. But there were lots of 1311's
still in service in the 1970s.

[Charlie Gibbs]

Post by h***@bbs.cpcn.com
Given how slow and small computers and peripherals were in
the old days, I know many old sites used assembler to
squeeze out every bit they could. For instance, it
was very common to store flags as bits, eight to a
byte, rather than as a byte containing Y or N.

At the other extreme, a meddlesome boss once wanted us
to set aside three bytes for flags so that they could
hold the value YES or NO. Sanity eventually prevailed...

Post by h***@bbs.cpcn.com
Assembler programmers knew all sorts of tricks in binary
to maximize efficiency and minimize space. With a 16k
or 32k machine and complex applications, they had to
squeeze it down.

When we expanded our machine's memory from 16K to 32K,
we wondered what we were going to do with all that space.
We eventually figured it out. :-)

Post by h***@bbs.cpcn.com
Per our conversation on tapes, they had to keep tapes
down, too. An 800 bpi tape didn't hold too much.

Again, the trade-off between block size and memory usage
became apparent. Given an inter-block gap of 0.6 inch,
small block sizes were _very_ inefficient. On the other
hand, large block sizes meant you ran out of memory just
allocating buffers. A lot of people talked about the
wonders of double-buffering, but usually there just
wasn't enough room for it.

Post by h***@bbs.cpcn.com
One issue was tape mounts--that was labor intensive and
ran out the clock, so keeping a tape file small was
important.

Form changes also ate a lot of time - we tried to
standardize on as few forms as possible. If you
wanted four copies of one report and three of another,
it was sometimes better to run both reports on four-part
paper.

Post by h***@bbs.cpcn.com
Disk space was expensive. A 1311 held only
two meg, though the 2311 improved on it, and the 2314
improved on it further. But there were lots of 1311's
still in service in the 1970s.

It still amazes me how the cost of disk storage has
fallen by seven orders of magnitude since then.

[Ahem A Rivet's Shot]

On 22 Feb 2020 20:24:51 GMT

Post by Charlie Gibbs

Post by h***@bbs.cpcn.com
Disk space was expensive. A 1311 held only
two meg, though the 2311 improved on it, and the 2314
improved on it further. But there were lots of 1311's
still in service in the 1970s.

It still amazes me how the cost of disk storage has
fallen by seven orders of magnitude since then.

Capacity has done similar things ten terabyte drives are commodity
devices while two meg isn't even enough for a decent on-chip CPU cache and
this ageing desktop has a hundred times more RAM than my first home unix
workstation had hard disk.

[d***@gmail.com]

One cold, dark, boring, winter day I decided to see how much a 16G thumbdrive would be in punch cards. I did my best at calculating it. I figured it would be 572 tons of punch cards, and would take about 15 semi-tractor-trailers to haul, and would required over a year just to read through a typical card reader into a computer. I guess we've made *some* progress since the 1960's.

[J. Clarke]

Post by d***@gmail.com
One cold, dark, boring, winter day I decided to see how much a 16G thumbdrive would be in punch cards. I did my best at calculating it. I figured it would be 572 tons of punch cards, and would take about 15 semi-tractor-trailers to haul, and would required over a year just to read through a typical card reader into a computer. I guess we've made *some* progress since the 1960's.

Now take a look at a terabyte microSD. Friggin thing's so small that
if you drop it on a carpet you'll likely never find it.

[Peter Flass]

Post by d***@gmail.com
One cold, dark, boring, winter day I decided to see how much a 16G
thumbdrive would be in punch cards. I did my best at calculating it. I
figured it would be 572 tons of punch cards, and would take about 15
semi-tractor-trailers to haul, and would required over a year just to
read through a typical card reader into a computer. I guess we've made
*some* progress since the 1960's.

Why does it seem that things were better in the old days. Maybe because it
was new and we were a lot younger?

[Quadibloc]

Post by Peter Flass
Why does it seem that things were better in the old days. Maybe because it
was new and we were a lot younger?

When a computer cost a million dollars, they could afford to provide helpful
manuals with it.

When a terminal cost over a thousand dollars, they could afford to put a nice
keyboard on it.

When there was no Internet, computers weren't attacked by hackers every other
day.

Some things _were_ better, as a direct consequence of most things being worse...
is the quick way to explain it.

John Savard

[J. Clarke]

On Sat, 22 Feb 2020 22:12:14 -0800 (PST), Quadibloc

Post by Quadibloc

Post by Peter Flass
Why does it seem that things were better in the old days. Maybe because it
was new and we were a lot younger?

When a computer cost a million dollars, they could afford to provide helpful
manuals with it.

When they sell a hundred million computers they can afford to provide
helpful manuals even if the computers are cheap.

Post by Quadibloc
When a terminal cost over a thousand dollars, they could afford to put a nice
keyboard on it.

You can buy keyboards today that are as good as or better than
anything that ever came on a thousand dollar terminal. However since
the keyboard interface is standardized you have a wide range of
choices and aren't stuck with what one vendor decided to provide.

Post by Quadibloc
When there was no Internet, computers weren't attacked by hackers every other
day.
Some things _were_ better, as a direct consequence of most things being worse...
is the quick way to explain it.
John Savard

[Michael LeVine]

Post by J. Clarke
On Sat, 22 Feb 2020 22:12:14 -0800 (PST), Quadibloc

Post by Quadibloc

Post by Peter Flass
Why does it seem that things were better in the old days. Maybe because it
was new and we were a lot younger?

When a computer cost a million dollars, they could afford to provide helpful
manuals with it.

When they sell a hundred million computers they can afford to provide
helpful manuals even if the computers are cheap.

The manuals are there, you just have to find them, either on line or
hidden in the firmware of the device.

Post by J. Clarke

Post by Quadibloc
When a terminal cost over a thousand dollars, they could afford to put a nice
keyboard on it.

You can buy keyboards today that are as good as or better than
anything that ever came on a thousand dollar terminal. However since
the keyboard interface is standardized you have a wide range of
choices and aren't stuck with what one vendor decided to provide.

Post by Quadibloc
When there was no Internet, computers weren't attacked by hackers every other
day.
Some things _were_ better, as a direct consequence of most things being worse...
is the quick way to explain it.
John Savard

[Ahem A Rivet's Shot]

On Sun, 23 Feb 2020 01:31:16 -0800

Post by Michael LeVine

Post by J. Clarke
When they sell a hundred million computers they can afford to provide
helpful manuals even if the computers are cheap.

The manuals are there, you just have to find them, either on line or
hidden in the firmware of the device.

The word 'sometimes' needs to be in that sentence, sometimes the
manuals are only available after negotiation and signing of NDAs. There are
some quite powerful ARM boxes going surprisingly cheaply under the guise of
TV boxes - they'd probably make decent workstations if only the
documentation were available to port an OS to them. You can run a standard
Linux distribution under the custom kernel but you can't upgrade the kernel
so when that runs out of steam you're stuck. OTOH there's no reason for the
manufacturers of these things to want to release that documentation, it
won't help them sell TV box SOCs.

[Quadibloc]

Also: can today's computers inspire such feelings of awe:

https://www.rediff.com/business/slide-show/slide-show-1-tech-a-look-at-the-early-computers/20130529.htm#17

John Savard

[Melzzzzz]

Post by Quadibloc
https://www.rediff.com/business/slide-show/slide-show-1-tech-a-look-at-the-early-computers/20130529.htm#17

Feelings of awe were because they were rare. Nowadays everyone has
computers and they are part of everyday life. Nothing unusual...

Post by Quadibloc
John Savard

[John Levine]

Post by Quadibloc
https://www.rediff.com/business/slide-show/slide-show-1-tech-a-look-at-the-early-computers/20130529.htm#17

People sure dressed better, suits and dresses everywhere. Even the
three year old girl in #17 is wearing a dress, tights, and proper
buckle shoes.

[Peter Flass]

Post by John Levine

Post by Quadibloc
https://www.rediff.com/business/slide-show/slide-show-1-tech-a-look-at-the-early-computers/20130529.htm#17

People sure dressed better, suits and dresses everywhere. Even the
three year old girl in #17 is wearing a dress, tights, and proper
buckle shoes.

This was the UK, though.

[Quadibloc]

Post by Peter Flass

Post by John Levine
People sure dressed better, suits and dresses everywhere. Even the
three year old girl in #17 is wearing a dress, tights, and proper
buckle shoes.

This was the UK, though.

I see the URL didn't take people directly to #17, which had been my intent.

Several of the images in the sequence note something in the captions indicating
they were taken in the UK, but that particular image didn't have anything to
indicate that, so I hadn't realized this.

John Savard

[h***@bbs.cpcn.com]

Post by John Levine
People sure dressed better, suits and dresses everywhere. Even the
three year old girl in #17 is wearing a dress, tights, and proper
buckle shoes.

It is weird. I don't think my father ever left the house--even
just to go to the local store, without a tie and hat. In
family vacation photos, he's wearing trousers and a tie
even in the summer.

Nixon was from that generation. They tried to make him
casual, but he ended up wearing his wingtips on the beach.
That was his world.

You see this in old TV shows. The men wore suits,
even when lounging at home. Even in the early 1970s
they were dressy. Mannix, in his early seasons, always
wore a coat and tie.

What kills me is that air conditioning was sparse years
ago but men still wore suits.

[Peter Flass]

Post by h***@bbs.cpcn.com

Post by John Levine
People sure dressed better, suits and dresses everywhere. Even the
three year old girl in #17 is wearing a dress, tights, and proper
buckle shoes.

It is weird. I don't think my father ever left the house--even
just to go to the local store, without a tie and hat. In
family vacation photos, he's wearing trousers and a tie
even in the summer.
Nixon was from that generation. They tried to make him
casual, but he ended up wearing his wingtips on the beach.
That was his world.
You see this in old TV shows. The men wore suits,
even when lounging at home. Even in the early 1970s
they were dressy. Mannix, in his early seasons, always
wore a coat and tie.
What kills me is that air conditioning was sparse years
ago but men still wore suits.

My father had to wear a suit and tie to work, and they had no A/C.

[J. Clarke]

On Mon, 24 Feb 2020 14:14:00 -0700, Peter Flass

Post by Peter Flass

Post by h***@bbs.cpcn.com

Post by John Levine
People sure dressed better, suits and dresses everywhere. Even the
three year old girl in #17 is wearing a dress, tights, and proper
buckle shoes.

It is weird. I don't think my father ever left the house--even
just to go to the local store, without a tie and hat. In
family vacation photos, he's wearing trousers and a tie
even in the summer.
Nixon was from that generation. They tried to make him
casual, but he ended up wearing his wingtips on the beach.
That was his world.
You see this in old TV shows. The men wore suits,
even when lounging at home. Even in the early 1970s
they were dressy. Mannix, in his early seasons, always
wore a coat and tie.
What kills me is that air conditioning was sparse years
ago but men still wore suits.

My father had to wear a suit and tie to work, and they had no A/C.

Mine, in Florida, had the option of a short sleeve open collar shirt
or a jacket and tie and he picked the jacket and tie.

[John Levine]

Post by h***@bbs.cpcn.com
What kills me is that air conditioning was sparse years
ago but men still wore suits.

Yes, but in those days nobody expected to get much work done on hot
summer days.

[h***@bbs.cpcn.com]

Post by Peter Flass
Why does it seem that things were better in the old days. Maybe because it
was new and we were a lot younger?

If I may offer some observations on that...

. Even though computers of the past were pretty crude--by today's
standards--they were incredibly powerful tools in their day.
That is, they were a huge advance over the tools we had.

For instance, I remember using BASIC timesharing--pretty
simple stuff--to do my trig, chemistry, and physics homework.
Without the computer, we'd have to do various table lookups
for log and trig functions, and use the slide rule. Sometimes
we'd resort to plain long division. Jeez, remember extrapolation
with the trig tables? P.I.T.A. A ten char/sec teletype was
pretty slick compared to that.

. They had slick mechanical aspects. Watching the card
reader, printer, and tape drives fly along was neat.
(This did not apply if you were using an IBM 1130,
glacially slow.)

. They gave us power. We controlled what the computer did--
we told it when to read a card, print a line, select
the tape drive.

. They were fun and satisfying. Writing a program that
run correctly was a source of satisfaction. Mastering
the innards of software and hardware was satisfying.

. They were new and novel. Exciting.

. We could experiment. For many of us, tinkering with
the innards of a CP/M machine (both software and hardware)
was fun. Likewise with the earliest PCs. For whatever reason,
tinkering with Windows isn't as much fun.


For myself, an Excel spreadsheet allows me to generate
a sophisticated report easily and quickly. Click click
click. Control breaks, sub-totals, counts, sort order.
But somehow 'rolling my own' in code was more fun.

[h***@bbs.cpcn.com]

Post by d***@gmail.com
One cold, dark, boring, winter day I decided to see how much a 16G thumbdrive would be in punch cards. I did my best at calculating it. I figured it would be 572 tons of punch cards, and would take about 15 semi-tractor-trailers to haul, and would required over a year just to read through a typical card reader into a computer. I guess we've made *some* progress since the 1960's.

Yes, but punched cards were a lot more fun! (At least until
you dropped your deck.)

[Peter Flass]

Post by h***@bbs.cpcn.com

Post by d***@gmail.com
One cold, dark, boring, winter day I decided to see how much a 16G
thumbdrive would be in punch cards. I did my best at calculating it. I
figured it would be 572 tons of punch cards, and would take about 15
semi-tractor-trailers to haul, and would required over a year just to
read through a typical card reader into a computer. I guess we've made
*some* progress since the 1960's.

Yes, but punched cards were a lot more fun! (At least until
you dropped your deck.)

Agreed. There was something very satisfying about watching (and listening)
when the card reader sucked your deck in at 1000cpm

[Charlie Gibbs]

Post by Peter Flass

On Saturday, February 22, 2020 at 5:22:33 PM UTC-5,

Post by d***@gmail.com
One cold, dark, boring, winter day I decided to see how much a 16G
thumbdrive would be in punch cards. I did my best at calculating it. I
figured it would be 572 tons of punch cards, and would take about 15
semi-tractor-trailers to haul, and would required over a year just to
read through a typical card reader into a computer. I guess we've made
*some* progress since the 1960's.

Yes, but punched cards were a lot more fun! (At least until
you dropped your deck.)

Agreed. There was something very satisfying about watching (and listening)
when the card reader sucked your deck in at 1000cpm

Plus the satisfying physicality of joggling a deck of cards
and dropping it into the hopper.

Some card jams were quite exciting (and funny in hindsight).
I once had a stacker jam where subsequent cards bounced off
and shot around the room.

[Mike Spencer]

Post by Charlie Gibbs
Some card jams were quite exciting (and funny in hindsight).
I once had a stacker jam where subsequent cards bounced off
and shot around the room.

There was a movie -- late 50s or early 60s -- in which a card sorter
that was classifying Army recruits repeatedly fired the lead
character's card through the air into a nearby waste basket. Until
the operator cleverly place one of the receiver bins on top of the
waste basket. Zochet...zinnngggg...plip. And the recruit was
classified.

No recollection of title, actor or anything else except that scene.

[Scott Lurndal]

One cold, dark, boring, winter day I decided to see how much a 16G thumbd=

rive would be in punch cards. I did my best at calculating it. I figured it=
would be 572 tons of punch cards, and would take about 15 semi-tractor-tra=
ilers to haul, and would required over a year just to read through a typica=
l card reader into a computer. I guess we've made *some* progress since the=
1960's.
Yes, but punched cards were a lot more fun! (At least until
you dropped your deck.)

What, you didn't draw a triangle over the top edges?

[Dan Espen]

Post by Scott Lurndal

One cold, dark, boring, winter day I decided to see how much a 16G thumbd=

rive would be in punch cards. I did my best at calculating it. I figured it=
would be 572 tons of punch cards, and would take about 15 semi-tractor-tra=
ilers to haul, and would required over a year just to read through a typica=
l card reader into a computer. I guess we've made *some* progress since the=
1960's.
Yes, but punched cards were a lot more fun! (At least until
you dropped your deck.)

What, you didn't draw a triangle over the top edges?

Usually a diagonal line.

I once visited another sites computer room.
This was a shipping company.
The object decks had these amazing pictures of ships drawn with markers.
The culprit said he didn't mind getting new decks.
I hope he took them home and sold them on Ebay.

[Peter Flass]

Post by Charlie Gibbs

Post by h***@bbs.cpcn.com
Given how slow and small computers and peripherals were in
the old days, I know many old sites used assembler to
squeeze out every bit they could. For instance, it
was very common to store flags as bits, eight to a
byte, rather than as a byte containing Y or N.

At the other extreme, a meddlesome boss once wanted us
to set aside three bytes for flags so that they could
hold the value YES or NO. Sanity eventually prevailed...

Post by h***@bbs.cpcn.com
Assembler programmers knew all sorts of tricks in binary
to maximize efficiency and minimize space. With a 16k
or 32k machine and complex applications, they had to
squeeze it down.

When we expanded our machine's memory from 16K to 32K,
we wondered what we were going to do with all that space.
We eventually figured it out. :-)

Post by h***@bbs.cpcn.com
Per our conversation on tapes, they had to keep tapes
down, too. An 800 bpi tape didn't hold too much.

Again, the trade-off between block size and memory usage
became apparent. Given an inter-block gap of 0.6 inch,
small block sizes were _very_ inefficient. On the other
hand, large block sizes meant you ran out of memory just
allocating buffers. A lot of people talked about the
wonders of double-buffering, but usually there just
wasn't enough room for it.

We had some programs that had a lot of files open at the same time, and we
had to plan pretty carefully. Sometimes block sizes for a system were
determined based on the worst-case program.

[John Varela]

Post by d***@gmail.com
...

Post by Quadibloc
Yes, you have described the current situation very accurately, in case anyone
might have doubts: and what you have pointed out are precisely its most salient
features.
John Savard

It was interesting to look back at the history of FORTRAN recently. It originally was created to dove-tail with the earliest computers machine language, a lot of FORTRAN statements would equate to a single instruction on those old computers (like the old tri-nary IF statement). Considering it is still in use today (albeit much more advanced) is a testament to it's power. Is there any other language that has that long of a lifespan? What is it, almost 70 years now?

I was coding in FORTRAN at Chance Vought Aircraft in the summer of
1958, almost 72 years ago.

[Quadibloc]

Post by John Varela
I was coding in FORTRAN at Chance Vought Aircraft in the summer of
1958, almost 72 years ago.

You certainly have me beat. Back then, I was still a very little child.

John Savard

[Ahem A Rivet's Shot]

On Sat, 22 Feb 2020 22:13:15 -0800 (PST)

Post by Quadibloc

Post by John Varela
I was coding in FORTRAN at Chance Vought Aircraft in the summer of
1958, almost 72 years ago.

You certainly have me beat. Back then, I was still a very little child.

I was still in two parts in summer 1958, but that's only 62 years
ago.

[John Varela]

On Sun, 23 Feb 2020 07:14:47 UTC, Ahem A Rivet's Shot

Post by Ahem A Rivet's Shot
On Sat, 22 Feb 2020 22:13:15 -0800 (PST)

Post by Quadibloc

Post by John Varela
I was coding in FORTRAN at Chance Vought Aircraft in the summer of
1958, almost 72 years ago.

You certainly have me beat. Back then, I was still a very little child.

I was still in two parts in summer 1958, but that's only 62 years
ago.

I know. I woke up in the middle of last night with the realization
that 72 years ago I was only 12 years old. In the summer of '58 I
was 22 and had a shiny new Master's degree.

[Dan Espen]

Post by Quadibloc

Post by John Varela
I was coding in FORTRAN at Chance Vought Aircraft in the summer of
1958, almost 72 years ago.

You certainly have me beat. Back then, I was still a very little child.

Beat me by 6 years.
I'm impressed.

[J. Clarke]

Post by h***@bbs.cpcn.com

Post by Quadibloc
I recently learned that although IBM's current Fortran compiler for the system
Z is a Fortran 77 compiler, it actually has been updated to include a lot of
Fortran 90 features, even if it doesn't fully meet the Fortran 90 spec. But
the days when Fortran was used for all sorts of stuff because we didn't have
anything else handy are long gone. C is in that position now, along with C++
of course.

I am not up on FORTRAN and don't know what IBM's official position
is on it for the Z series mainframe. But my impression (based
on conversations in this newsgroup) is that they no longer support
it for Z, but do so for other machines. Old stuff still runs.

The most assuredly do support it. What they don't do is make changes
that would break existing code--that means that it's mostly frozen at
a Fortran 77-and-a-half level.

Post by h***@bbs.cpcn.com
I don't know what legacy heavy users, such as weather
forecasting on a 370-195, use today.

Supercomputers or "the cloud".

Post by h***@bbs.cpcn.com
Note that modern IBM Z COBOL has a lot of FORTRAN features
in it, like trig functions and floating point for years.
I haven't studied it, but I'm guessing someone with a new
basic sci/eng task could do it in COBOL.
For what it's worth, engineers have long told me FORTRAN
is obsolete, replaced by other tools, such CAD/CAM.

If CAD/CAM does what you want to do, sure. I don't think I can buy a
CAD/CAM system that can calculate Jupiter entry.

Post by h***@bbs.cpcn.com
I have no idea what the heavy-duty number crunchers use
on super computers.

[Louis Krupp]

<snip>

Post by h***@bbs.cpcn.com
For what it's worth, engineers have long told me FORTRAN
is obsolete, replaced by other tools, such CAD/CAM.

Some CAD/CAM systems were once -- and might still be -- written in
FORTRAN. I worked at a mechanical CAD software company about thirty
years ago; most of the code was in FORTRAN, and some was in C.

Louis

[Gareth Evans]

Post by J. Clarke
I feel old.
It occurred to me today to find out if my employer had an officially
blessed Fortran environment for the PC.
There wasn't anything in the site for software orders and there has
been a recent redesign so I couldn't find the general request.
So I called the help desk and asked. First response--"what is Fortran
and does it have another name?". Ultimately she found where the
general request had been moved to and I sent that off, so we'll see.
But this person, who provides IT support for a company that has
multiple mainframes, has never even heard of Fortran. Am I that much
of an antique?

Such is the pace of change that we all become nobbut dinosaurs.

I read Computer and Communications Engineering, a 3rd year
specialisation of electronics, at Essex from
1969 to 1972, and much of the final year, both for hardware
and software, used the PDP8 as a model to study. (There's no
call these days to understand half-select noise in core stores ! )

This was 50 years ago, and yet 50 years before that
in 1920 there was no electronics and very few had access
to thermionic valves (Tubes for the Yanks).

As I said, such is the pace of change.

I last visited the Science Museum in London in 1999, 21 years
ago and was disgusted to see various PDP8 models displayed
as historical entities.

It having been the formative years of my career, 10 years
in assembler on the PDP11, my interest in computers
remains as hands-on at the very low level and,
as such, it is frustrating not to be able to get
details of the GPU in the Raspberry Pi model 4,
where for only a few pounds (dollars to the Yanks)
one can get one's hands on a 64 bit multi-core
processor running at several GHz with umpteen
GBytes of RAM to hand.

[Quadibloc]

Post by Gareth Evans
I last visited the Science Museum in London in 1999, 21 years
ago and was disgusted to see various PDP8 models displayed
as historical entities.

Even in 1999, that would have been an appropriate way to deal with them.

In 1980, one might ask why they weren't in active service.

John Savard

[David Wade]

Post by J. Clarke
I feel old.
It occurred to me today to find out if my employer had an officially
blessed Fortran environment for the PC.
There wasn't anything in the site for software orders and there has
been a recent redesign so I couldn't find the general request.
So I called the help desk and asked. First response--"what is Fortran
and does it have another name?". Ultimately she found where the
general request had been moved to and I sent that off, so we'll see.
But this person, who provides IT support for a company that has
multiple mainframes, has never even heard of Fortran. Am I that much
of an antique?

These days "provides first level IT support" roughly translated means
"can ask a list of questions and work out who can actually fix it".

I would say in my previous job ringing the help desk and asking about
Fortran would have had the same answer. Probably even asking about Cobol
as well even though we were once a Cobol shop..

Dave

[Peter Flass]

Post by David Wade

Post by J. Clarke
I feel old.
It occurred to me today to find out if my employer had an officially
blessed Fortran environment for the PC.
There wasn't anything in the site for software orders and there has
been a recent redesign so I couldn't find the general request.
So I called the help desk and asked. First response--"what is Fortran
and does it have another name?". Ultimately she found where the
general request had been moved to and I sent that off, so we'll see.
But this person, who provides IT support for a company that has
multiple mainframes, has never even heard of Fortran. Am I that much
of an antique?

These days "provides first level IT support" roughly translated means
"can ask a list of questions and work out who can actually fix it".

Probably with a simple script of questions and answers such as “does it
display a cursor?”

Post by David Wade
I would say in my previous job ringing the help desk and asking about
Fortran would have had the same answer. Probably even asking about Cobol
as well even though we were once a Cobol shop..

[Andreas Kohlbach]

Post by Peter Flass
Probably with a simple script of questions and answers such as “does it
display a cursor?”

Did anybody watch the UK crime scene drama The Sweeney?

Regan reading the manual of a new office computer:

| It says here "The cursor will guide the operator..." What's a cursor?
|
| Det. Sgt. George Carter:
| Someone we nick for obscene language.

*g*

[JimP]

On Tue, 18 Feb 2020 19:27:41 -0500, J. Clarke

Post by J. Clarke
I feel old.
It occurred to me today to find out if my employer had an officially
blessed Fortran environment for the PC.
There wasn't anything in the site for software orders and there has
been a recent redesign so I couldn't find the general request.
So I called the help desk and asked. First response--"what is Fortran
and does it have another name?". Ultimately she found where the
general request had been moved to and I sent that off, so we'll see.
But this person, who provides IT support for a company that has
multiple mainframes, has never even heard of Fortran. Am I that much
of an antique?

I was asked at my last two jobs what programming languages I had used
or heard of. I was the only person who had heard of FORTRAN and one
other person knew about COBOL. That was this century.

[Charlie Gibbs]

Post by JimP
On Tue, 18 Feb 2020 19:27:41 -0500, J. Clarke

Post by J. Clarke
I feel old.
It occurred to me today to find out if my employer had an officially
blessed Fortran environment for the PC.
There wasn't anything in the site for software orders and there has
been a recent redesign so I couldn't find the general request.
So I called the help desk and asked. First response--"what is Fortran
and does it have another name?". Ultimately she found where the
general request had been moved to and I sent that off, so we'll see.
But this person, who provides IT support for a company that has
multiple mainframes, has never even heard of Fortran. Am I that much
of an antique?

I was asked at my last two jobs what programming languages I had used
or heard of. I was the only person who had heard of FORTRAN and one
other person knew about COBOL. That was this century.

You'd really bend their minds if you mentioned RPG.
"What do rocket-propelled grenades have to do with programming?"
However, the really savvy ones might think you write role-playing games.

[JimP]

Post by Charlie Gibbs

Post by JimP
On Tue, 18 Feb 2020 19:27:41 -0500, J. Clarke

Post by J. Clarke
I feel old.
It occurred to me today to find out if my employer had an officially
blessed Fortran environment for the PC.
There wasn't anything in the site for software orders and there has
been a recent redesign so I couldn't find the general request.
So I called the help desk and asked. First response--"what is Fortran
and does it have another name?". Ultimately she found where the
general request had been moved to and I sent that off, so we'll see.
But this person, who provides IT support for a company that has
multiple mainframes, has never even heard of Fortran. Am I that much
of an antique?

I was asked at my last two jobs what programming languages I had used
or heard of. I was the only person who had heard of FORTRAN and one
other person knew about COBOL. That was this century.

You'd really bend their minds if you mentioned RPG.
"What do rocket-propelled grenades have to do with programming?"
However, the really savvy ones might think you write role-playing games.

Actually, one of those employers, several of the developers liked RPG
and had written software using it.

[Quadibloc]

Post by Charlie Gibbs
You'd really bend their minds if you mentioned RPG.
"What do rocket-propelled grenades have to do with programming?"
However, the really savvy ones might think you write role-playing games.

Indeed, Report Program Generator has had its acronym recycled...

John Savard

[h***@bbs.cpcn.com]

Post by Charlie Gibbs
You'd really bend their minds if you mentioned RPG.
"What do rocket-propelled grenades have to do with programming?"
However, the really savvy ones might think you write role-playing games.

There used to be huge installed base of AS/400 machines.
I think their successors, the "i" series, are still
around. Are they still popular?

Likewise, Visual/RPG/400 used to be very popular on that
platform. Is it still?

IBM still publishes a Z series magazine, and I think an
i series magazine, but can't find them at the moment.

Then they got an "N" series. No clue on that one.
There is a subway train "N",from Astoria (Queens)
to Coney Island (Brooklyn).

[Ahem A Rivet's Shot]

On Wed, 19 Feb 2020 09:14:54 -0600

Post by JimP
I was asked at my last two jobs what programming languages I had used
or heard of. I was the only person who had heard of FORTRAN and one
other person knew about COBOL. That was this century.

So SNOBOL, ML, Simula, Prolog, Forth, Occam, APL, LISP, Modula
et al would have been familiar to all ?

[JimP]

On Wed, 19 Feb 2020 17:43:12 +0000, Ahem A Rivet's Shot

Post by Ahem A Rivet's Shot
On Wed, 19 Feb 2020 09:14:54 -0600

Post by JimP
I was asked at my last two jobs what programming languages I had used
or heard of. I was the only person who had heard of FORTRAN and one
other person knew about COBOL. That was this century.

So SNOBOL, ML, Simula, Prolog, Forth, Occam, APL, LISP, Modula
et al would have been familiar to all ?

I had heard of LISP and Forth. And saw mentions of the rest in here,
about 1992. Forth on some home computer, not elsewhere.

Maybe Prolog, but I'm not sure.

[Ahem A Rivet's Shot]

On Wed, 19 Feb 2020 14:43:57 -0600

Post by JimP
On Wed, 19 Feb 2020 17:43:12 +0000, Ahem A Rivet's Shot

Post by Ahem A Rivet's Shot
On Wed, 19 Feb 2020 09:14:54 -0600

Post by JimP
I was asked at my last two jobs what programming languages I had used
or heard of. I was the only person who had heard of FORTRAN and one
other person knew about COBOL. That was this century.

So SNOBOL, ML, Simula, Prolog, Forth, Occam, APL, LISP, Modula
et al would have been familiar to all ?

I had heard of LISP and Forth. And saw mentions of the rest in her

Most of the above were covered in my one year computer science
course in 80/81 along with a couple of Algols, BCPL and of course FORTRAN
and COBOL. The others I bumped into later along with better known languages
such as Smalltalk and XSLT. There are good reasons languages like C, C++,
Python and Perl are popular.

[Peter Flass]

Post by Ahem A Rivet's Shot
On Wed, 19 Feb 2020 09:14:54 -0600

Post by JimP
I was asked at my last two jobs what programming languages I had used
or heard of. I was the only person who had heard of FORTRAN and one
other person knew about COBOL. That was this century.

So SNOBOL, ML, Simula, Prolog, Forth, Occam, APL, LISP, Modula
et al would have been familiar to all ?

Have programmed SNOBOL and APL. May have touched LIST. Read some stuff
about Forth. Know the others mostly as names.

[Quadibloc]

Post by Ahem A Rivet's Shot
On Wed, 19 Feb 2020 09:14:54 -0600

Post by JimP
I was asked at my last two jobs what programming languages I had used
or heard of. I was the only person who had heard of FORTRAN and one
other person knew about COBOL. That was this century.

So SNOBOL, ML, Simula, Prolog, Forth, Occam, APL, LISP, Modula
et al would have been familiar to all ?

Not hardly, but as those languages were, even in their heyday, less widely used
than Fortran and COBOL, a lack of familiarity with them would seem less jarring.

John Savard

[Peter Flass]

Post by JimP
On Tue, 18 Feb 2020 19:27:41 -0500, J. Clarke

Post by J. Clarke
I feel old.
It occurred to me today to find out if my employer had an officially
blessed Fortran environment for the PC.
There wasn't anything in the site for software orders and there has
been a recent redesign so I couldn't find the general request.
So I called the help desk and asked. First response--"what is Fortran
and does it have another name?". Ultimately she found where the
general request had been moved to and I sent that off, so we'll see.
But this person, who provides IT support for a company that has
multiple mainframes, has never even heard of Fortran. Am I that much
of an antique?

I was asked at my last two jobs what programming languages I had used
or heard of. I was the only person who had heard of FORTRAN and one
other person knew about COBOL. That was this century.

Fair enough, I’ve never heard of a lot of stuff that’s being used today.

[h***@bbs.cpcn.com]

Post by JimP
I was asked at my last two jobs what programming languages I had used
or heard of. I was the only person who had heard of FORTRAN and one
other person knew about COBOL. That was this century.

For what it's worth, I know a lot of former FORTRAN programmers.
They don't miss it nor care about it.

Actually, again FWIW, my impression is that for most programmers,
when they hang up their flowchart template to retire, they
have zero interest in computers, that is, they are done with it
(beyond personal use as a tool). For those who keep on
working, they go into other fields.

I do know one sprogrammer who retired and returned to work
as a data entry operator. That's what she's happy doing.