Wikipedia page

[Waldek Hebisch]

On FriCAS Wikipedia page there is claim that Scratchpad (initial
IBM project) was written in Fortran. I wander what is the
source of this claim. All texts about Scratchpad say that
it was written in Lisp (IIUC VMLISP was created as part of
computer algebra project at IBM, before the project was named
Scratchpad).

--
Waldek Hebisch

[Tim Daly]

I was actively involved in translating Scratchpad from VMLISP to about

a dozen different lisps, then eventually to Common Lisp. I am unaware

of any Fortran code. Some of the Scratchpad code was translated from

MIT's Lisp. There was a Meta language developed by Dick Jenks which

was compiled first in order to create the parser. I removed the Meta

language and its compiler.

The ultimate target became AKCL, Austin Kyoto Common Lisp, now GCL.

I worked on the garbage collector, tail recursion, and a bunch of Common

Lisp compatibility issues. I minimally, and usually second-hand behind

Fred Blair, was on the Common Lisp mailing list so I was carefully evaluating

the AKCL compatibility.

The only possible connection to Fortran was my effort to translate the

BLAS and LAPACK libraries to Common Lisp. The translation result is 

part of the Axiom repository and was done during the open source effort.

NAG also made a connection to their Fortran libraries but that was after

it left IBM.

If there is a Scratchpad / Fortran connection that is a surprise to me.

Tim Daly

[Kurt Pagani]

I missed this (and maybe a lot of others in 2025) and have seen it only
now. I don't know from where the Fortran myth originates, however, I
have found on Tim's site:

"James H. Griesmer, "A history of the SCRATCHPAD Project (1965-1977)",
IBM newsletter, vol. 1, no. 2, Jan. 15, 1986" (PDF). axiom-developer.org.

"R.D.Jenks, "A history of the SCRATCHPAD Project (1977-1986)", IBM
newsletter, vol. 1, no. 3, May. 15, 1986" (PDF). axiom-developer.org.

They are also referenced at https://fricas.org/history.html

So it seems that Fortran never played a role in the development. I think
LISP 1.5 or S/360 LISP respectively was the lang they started with (?)
Anyway I just updated https://en.wikipedia.org/wiki/FriCAS by removing
the "Fortran" passage and adding the two links above (ref 6 and 8).

> --
> You received this message because you are subscribed to the Google
> Groups "FriCAS - computer algebra system" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to fricas-devel...@googlegroups.com <mailto:fricas-
> devel+un...@googlegroups.com>.
> To view this discussion visit https://groups.google.com/d/msgid/fricas-
> devel/d42eaef6-d45f-49ec-9795-24772631c102n%40googlegroups.com <https://
> groups.google.com/d/msgid/fricas-devel/d42eaef6-
> d45f-49ec-9795-24772631c102n%40googlegroups.com?
> utm_medium=email&utm_source=footer>.

[Kurt Pagani]

Addendum:


When looking for the "Fortran" myth, I found a lot of interesting
documents. The two below may be especially interesting.

https://cs.uwaterloo.ca/research/tr/1980/CS-80-49.pdf

https://ntrs.nasa.gov/api/citations/19940009325/downloads/19940009325.pdf

The frequent mentioning of "FORMACS"
(https://en.wikipedia.org/wiki/FORMAC) may be the source of the
confusion ...

[Tim Daly]

Scratchpad I was James Griesmer and Dick Jenks which was in Fortran
I believe.

Scratchpad II, later Axiom, was a complete redesign and as far as I know
was written in LISPVM. I was brought on board to convert it to Common
Lisp since LISPVM only ran on the IBM Mainframe. I was a very early
user of Common LIsp before it was widely available and before it became
a standard. I was a minor contributor to the standards debate, mostly
through Fred Blair.

I was also a user of LISPVM. Indeed I started with Lisp in college around
1971. Fred Blair at IBM Research made his version available to me.

The other Fortran based system was a Lisp from Bell Labs.
I have a source listing of it in my library somewhere.

So there is a connection in history of Fortran software with Lisp
but none that I know of with Scratchpad II.

NAG is famous for its Fortran-based math library which was added
to the commercial version of Axiom. However I believe I removed
all of those connections in the open source version as NAG was
not going to release its library as open source.

As mentioned, I converted the Fortran versions of BLAS and LAPACK
(both in Fortran) to Common Lisp. The converted code is in the Axiom
repository. Still to do is to provide test code and Spad covers for
these numeric libraries. However I'm wasting my time connecting to
LEAN proof code so that is test code is far in the future.

Tim

> --
> You received this message because you are subscribed to a topic in the Google Groups "FriCAS - computer algebra system" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/fricas-devel/MZLEWUsXhQM/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to fricas-devel...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/fricas-devel/037c3d1b-cd0b-4df6-8461-339c8fe75675%40gmail.com.

[Waldek Hebisch]

On Thu, Jul 09, 2026 at 03:16:39PM -0400, Tim Daly wrote:
> Scratchpad I was James Griesmer and Dick Jenks which was in Fortran
> I believe.

I was not there but all evidence that I saw says different thing.
Griesmer biography says that LISPVM was created in order to do
symbolic computation. Other texts say that they got pieces of Lisp
code from other teams and build upon this.

> You received this message because you are subscribed to the Google Groups "FriCAS - computer algebra system" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to fricas-devel...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/fricas-devel/CAJn5L%3DJ8hfMfcRqQBoZJq-jWMA_rAPJ_jcciFXhKPmtst%3D6FNg%40mail.gmail.com.

--
Waldek Hebisch

[Kurt Pagani]

On 09/07/2026 21:16, Tim Daly wrote:
> Scratchpad I was James Griesmer and Dick Jenks which was in Fortran
> I believe.

That's the crux of the matter. Was FORTRAN really involved or not? When
reading

"James H. Griesmer, "A history of the SCRATCHPAD Project (1965-1977)",
IBM newsletter, vol. 1, no. 2, Jan. 15, 1986" (PDF). axiom-developer.org.

then my understanding is that they used S/360 LISP from beginning, but
the mentioning of FORMAC might be indicating that Fortran played a role.
You know the connections of all those systems, so perhaps you can shed
light on this? For the moment (at least) the mention of 'Fortran' has
been avoided in the WP page.

>
> Scratchpad II, later Axiom, was a complete redesign and as far as I know
> was written in LISPVM. I was brought on board to convert it to Common
> Lisp since LISPVM only ran on the IBM Mainframe. I was a very early
> user of Common LIsp before it was widely available and before it became
> a standard. I was a minor contributor to the standards debate, mostly
> through Fred Blair.
>
> I was also a user of LISPVM. Indeed I started with Lisp in college around
> 1971. Fred Blair at IBM Research made his version available to me.
>
> The other Fortran based system was a Lisp from Bell Labs.
> I have a source listing of it in my library somewhere.

Well, you mean that the "Lisp" was written in Fortran?

>
> So there is a connection in history of Fortran software with Lisp
> but none that I know of with Scratchpad II.
>
> NAG is famous for its Fortran-based math library which was added
> to the commercial version of Axiom. However I believe I removed
> all of those connections in the open source version as NAG was
> not going to release its library as open source.

The "nagd" sources still exist ...

/*--------------------------------------------------------------------------
source file for the nag daemon
Michael Mc Gettrick (m...@maths.bath.ac.uk)
--------------------------------------------------------------------------*/

Their Fortran library is still closed of course.

>
> As mentioned, I converted the Fortran versions of BLAS and LAPACK
> (both in Fortran) to Common Lisp. The converted code is in the Axiom
> repository. Still to do is to provide test code and Spad covers for
> these numeric libraries.

Did you use f2cl?
https://github.com/rtoy/f2cl
A great tool. I converted SLATEC to CL and I think that MAXIMA also uses
a f2cl conversion of SLATEC for special functions.

> However I'm wasting my time connecting to
> LEAN proof code so that is test code is far in the future.

Mastering LEAN is certainly no waste of time :)

[Qian Yun]

On 7/10/26 12:48 AM, Kurt Pagani wrote:
>
> I missed this (and maybe a lot of others in 2025) and have seen it only
> now. I don't know from where the Fortran myth originates, however, I
> have found on Tim's site:
>
> "James H. Griesmer, "A history of the SCRATCHPAD Project (1965-1977)",
> IBM newsletter, vol. 1, no. 2, Jan. 15, 1986" (PDF). axiom-developer.org.
>

Wikipedia also mentions "Ralph Gomory" and I was confused about it.
Search engine returns no link between Ralph and Scratchpad.

But in the pdf above, first sentence gives:

"
In the spring of 1965, Ralph Gomory approached me and asked me to
begin a project whose goal would be to develop a computer system
suitable for a mathematician to use in trying out ideas and conjectures.
"

- Qian

[Tim Daly]

I might be mistaken about Scratchpad I. I will have to research it further.

My interactions with James Griesmer were related to an IBM effort
to create an artificial intelligent advisor for IBM mainframe configuration.
I wrote the AI Expert System they used which was a combination
of Forgy's RETE and May's K-Rep[0] in LISP. I don't recall discussing
Scratchpad I with him.

Ralph Gomery, when I knew him, was the head of IBM Research.
My only connection with him was when he would visit our robot lab
to chat with me about the latest robot task I was programming
or when I spent time making jokes with him. I believe Ralph's PhD
work was on integer bin packing. I remember him explaining it to me.

Dick Jenks was my manager on the Scratchpad II effort.

Barry Trager (who had an office next to mine) worked on his PhD
thesis while working on Scratchpad II. HIs advisors were Joel Moses
(author of SIN) and Richard Zippel.

A possible path of publications might be:

Slagle, James R. "A Heuristic Program that Solves Symbolic Integration
Problems in
Freshman Calculus, Symbolic Automatic Integrator." MIT Ph.D., 1961.

Moses, Joel "Symbolic Integration", MIT 1967

Wang, Paul S-H. "Evaluation of Definite Integrals by Symbolic
Manipulation." MIT Ph.D.,
1971.

Fateman, Richard. "Essays in Algebraic Simplification." MIT Ph.D., 1972.

Trager, Barry. "Algorithms for Manipulating Algebraic Functions." MIT
S.B., 1976.

Yun, David. "The Hensel Lemma in Algebraic Manipulation." MIT Ph.D., 1974.

Zippel, Richard. "Probabilistic Algorithms for Sparse
Polynomials."QA76.M41.P96 no. 138)
MIT Ph.D., 1979.

Trager, Barry. "Integration of Algebraic Functions." MIT Ph.D., 1984.

Joel Moses was a thesis advisor of Barry Trager (per the thesis)
so the lineage goes way back.

MIT has a few boxes of papers on the subject. If anyone has access to
those materials I'd love to know. It would be fun (and painful) to try to
write the history up to the present.

Of course MIT was not the only place doing symbolic math.
For example, see the contribution list on each Axiom document.
James Davenport did work on symplification of polynomials and
cylindrical algebraic decomposition in England.

Back in the last century when I was working, "symbolic computation" was
considered "artificial intelligence" (as was robotics). Anyone who did
anything in AI at that time almost certainly did it in LISP. As for LISP
itself, I have probably a dozen printouts of various implementations;
I contributed a box of my original materials to the Lisp History Collection[1]
so you can see them there.

The best of the pile, in my opinion is Queinnec 's "Lisp In Small Pieces" [2],
which is a "must read" for anyone who programs.

And there is McCarthy's paper[3] which reduced LISP to a single page[4,pp 70-71]
you could fit on a t-shirt. I even tried to implement it in Verilog for my
FPGA (Field Programmable Gate Array) chip work. Everyone who is anyone
has to drop everything to write a lisp :-)

FORTRAN was involved in early efforts. Even COBOL (which I
taught at UCONN) was involved. But all of those efforts eventually just
became a LISP implementation. LISP was AI.

Tim

[0] Tim Daly, John Kastner, Eric Mays HICSS 1988
"Integrating Rules And Inheritance Networks In A Knowledge-Based
Financial Marketing Consultation System"

[1] Paul McJones "Lisp History Collection"
https://softwarepreservation.computerhistory.org/LISP/

[2] Christian Queinnec "Lisp in Small PIeces"
https://www.amazon.com/Lisp-Small-Pieces-Christian-Queinnec/dp/0521545668

[3] John McCarthy 1960
"Recursive Functions Of Symbolic Expressions And Their Computation By
Machine, Part I"

[4] John McCarthy
https://softwarepreservation.computerhistory.org/LISP/book/LISP%201.5%20Programmers%20Manual.pdf

> --

> You received this message because you are subscribed to a topic in the Google Groups "FriCAS - computer algebra system" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/fricas-devel/MZLEWUsXhQM/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to fricas-devel...@googlegroups.com.

> To view this discussion visit https://groups.google.com/d/msgid/fricas-devel/9cce6dd6-7686-4c3e-b2fa-da36f55e9538%40gmail.com.

[Kurt Pagani]

Quite revealing, thanks.

On 10/07/2026 08:22, Tim Daly wrote:
> I might be mistaken about Scratchpad I. I will have to research it further.
>

What's your opinion on:

SCRATCHPAD: Page 59
https://cs.uwaterloo.ca/research/tr/1980/CS-80-49.pdf

SCRATCHPAD: Page 8
https://ntrs.nasa.gov/api/citations/19940009325/downloads/19940009325.pdf

Can you confirm the stories? The authors (seemingly unrelated to the
development) must have had some inside knowledge?

[Tim Daly]

About the "Symbolic Computation And Constructive Mathematics" paper.
It seems the paper reviews Scratchpad I not Scratchpad II.
I never used Scratchpad I so I can't say much about it.

About the "Applications of Symbolic and Algebraic Manipulation..." paper

It was perfectly normal to do numeric integration, such as using
the method of "false position". There were other algorithms.

I wrote a whole series of numeric integration routines in Fortran as
part of a mathematics class. At the time there was no "computer
science" to speak of so my undergraduate degree is in mathematics.
I also wrote Fortran routines to solve Runga-Kutta ODEs. But it was
all just classwork to keep the math professors happy. Certainly none
of it was symbolic mathematics.

I ran the "computer lab" all 4 years of college. We started with 5 TTY
machines connected to Rutgers over a phone line, ending up with
16 Selectric typewriters. This was in the 1971 timeframe, predating
Scratchpad II by several years.

I spend a day a week going to Rutgers to listen to graduate students
defend their PhD thesis work. My co-worker Dave and I worked our
way through Ann Yasuhara's book "Recursive Function Theory And
Logic" which was as close as it comes to "computer science" at that
time. The math professors were just a chapter ahead of the students
from the book on BASIC. Dave and I wrote some of the exams for the
professors. It was VERY early in the game. (Amusing note: One of
my math professors was named George Google :-) )

The various pre-1971 systems mentioned in the paper were not available.
We were lucky to have a working punched card machine and overnight
printed results. :-)

If you look at the table on page 24 none of those systems were know
to us in 1971.

Tim



As to the availability of Scratchpad II, I regularly made full source
tapes of the Common Lisp version of Scratchpad that were sent
to anyone who asked.

> To view this discussion visit https://groups.google.com/d/msgid/fricas-devel/2cdd830a-9875-48cd-ae4f-13e8b3e13714%40gmail.com.

[Kurt Pagani]

On 10/07/2026 13:39, Tim Daly wrote:
> About the "Symbolic Computation And Constructive Mathematics" paper.
> It seems the paper reviews Scratchpad I not Scratchpad II.
> I never used Scratchpad I so I can't say much about it.

Well, then I'm afraid some dark spots will remain.
Anyway, whether SP I was written in Fortran or not will hardly change
the orbit :)

>
> About the "Applications of Symbolic and Algebraic Manipulation..." paper
>
> It was perfectly normal to do numeric integration, such as using
> the method of "false position". There were other algorithms.
>
> I wrote a whole series of numeric integration routines in Fortran as
> part of a mathematics class. At the time there was no "computer
> science" to speak of so my undergraduate degree is in mathematics.
> I also wrote Fortran routines to solve Runga-Kutta ODEs. But it was
> all just classwork to keep the math professors happy. Certainly none
> of it was symbolic mathematics.
>
> I ran the "computer lab" all 4 years of college. We started with 5 TTY
> machines connected to Rutgers over a phone line, ending up with
> 16 Selectric typewriters. This was in the 1971 timeframe, predating
> Scratchpad II by several years.
>
> I spend a day a week going to Rutgers to listen to graduate students
> defend their PhD thesis work. My co-worker Dave and I worked our
> way through Ann Yasuhara's book "Recursive Function Theory And
> Logic" which was as close as it comes to "computer science" at that
> time. The math professors were just a chapter ahead of the students
> from the book on BASIC. Dave and I wrote some of the exams for the
> professors. It was VERY early in the game. (Amusing note: One of
> my math professors was named George Google :-) )
>
> The various pre-1971 systems mentioned in the paper were not available.
> We were lucky to have a working punched card machine and overnight
> printed results. :-)

Conceivable, according to WP "Open source as a term emerged in the late
1990s. " only.
https://en.wikipedia.org/wiki/Open_source

Moreover, the 'transport' of punch cards is/was always a risk, wasn't it?

[Tim Daly]

The concept of Open Source wasn't really a problem before companies
like Microsoft got in the game. Most people were in the game for the
education. Microsoft was in the game for money. That's why I had the
source code for the Bell Labs Lisp and their chess program (which
regularly beat Dave and I no matter how hard we tried).

I have given away the source code of everything I've been involved
with long before it had a name. Scratchpad II tapes contained the
complete source. This was important because anyone who wanted to
use the system could see and improve the algorithm. This is mathematics
not magic. I do wish the algorithms had bibliographic references.

The fact that the Scratchpad code is open source is quite useful
since the LLMs can read the code to find boundary cases and bugs.

As for punched cards "transport" being a problem... Every deck used
columns 73-80 with a linear sequence number. The first step of every
punched card job was to sort the records on those columns. This was
especially important as the card readers sometimes misordered the cards.
The sequence numbers incremented by 10 to allow you to insert a card
without repunching the whole deck. As a side-effect Fortran program line
length was 72 characters.

Another war story. Except for alphanumeric data within a FORMAT
statement, DATA statement, or literal constant, blanks (spaces) and
TABS are ignored and may be used freely for appearance purposes.
Thus "GO TO 5" and "GOTO5" are the same thing.

One of my jobs in the "computer room" was to answer student questions
for things like homework (which, unbeknownst to the student I was the
likely author of the homework question given out in class). One student
discovered this fact about Fortran and came to me with a solid block of
text 72 characters wide which was a valid though incorrect Fortran
program. He was trying to minimize the number of cards in his deck.
Fortunately for both of us I had already written a "pretty print" program
which reformats code to text I liked.

> To view this discussion visit https://groups.google.com/d/msgid/fricas-devel/e164aff8-1b4a-4668-a737-7eaf4a1426b7%40gmail.com.