Shen computer algebra system

[Raoul Duke]

Claude says it is plausible. 

https://docs.google.com/document/d/10c1TD63O6l1pYB8xmRlRvzdGs7SUCI-KKbevyQJc5IU/edit?usp=drivesdk

[Reuben]

This would be amazing. Back in 2017 Mathematica was the only CAS i could find that would do symbolic integration (gnarly 3rd/4th order partial derivatives of Black-Scholes equation). To my knowledge it's still the only one...

[dr.mt...@gmail.com]

There is some mat'l in TBoS on algebra in Seh which might

have some relevance https://shenlanguage.org/TBoS/tbos_288.html.

Mark

[Joshua Scholar]

There has been some work on implementing a CAS in Julia, but I want to echo Richard Fateman who was heavily involved in developing Maxsyma and then in its open sourced version in Julia's forums, Maxima that there isn't a huge advantage to trying to duplicate hundreds of man years of work in a different language. 

Maxima is written in Common Lisp and the most common version of Shen is implemented in Common Lisp.  You could just load both in at the same time and have shen programs directly access Maxima data and programs and routines. 

https://discourse.julialang.org/t/cas-best-practices/58092/22

https://discourse.julialang.org/t/cas-best-practices/58092/27

I don't know the state of Julia's symbolics package now, whether they got an advantage out of it the last 5 years since that argument.

https://docs.sciml.ai/Symbolics/stable/

--
You received this message because you are subscribed to the Google Groups "Shen" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qilang+un...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/qilang/1e17ab57-fbf3-4f18-b346-46ba492f8254n%40googlegroups.com.

[Joshua Scholar]

Apologies.  I don't know how an editing error moved some words around in my last post. It should have read

On Thu, Jun 18, 2026 at 7:13 AM dr.mt...@gmail.com <dr.mt...@gmail.com> wrote:

[Joshua Scholar]

Arrgh!!! Here in Firefox in Linux I just found that I can't hit "return" without it deciding to send an unfinished email.  Idk.  I guess it decides to highlight the button instead of putting a carriage return in. Two messed up posts in a row! Once again it was supposed to say:

There has been some work on implementing a CAS in Julia, but I want to echo Richard Fateman who was heavily involved in developing Maxsyma and then in its open sourced version Maxima  in Julia's forums, that there isn't a huge advantage to trying to duplicate hundreds of man years of work in a different language.

Maxima is written in Common Lisp and the most common version of Shen is implemented in Common Lisp.  You could just load both in at the same time and have shen programs directly access Maxima data and programs and routines.

https://discourse.julialang.org/t/cas-best-practices/58092/22
https://discourse.julialang.org/t/cas-best-practices/58092/27

I don't know the state of Julia's symbolics package now, whether they got an advantage out of it the last 5 years since that argument.
https://docs.sciml.ai/Symbolics/stable/

[Raoul Duke]

0) sure, anybody is free to not reinvent wheels :-)
a) because it is there.
b) because shen can in theory find more errors, making everything more robust.
c) if one is willing/able to leverage AI then the open source man
hours can be paid forward into a stronger more robust system in less
time. (i can understand if some people are "pro" that idea, and if
some people are "con" that idea.)
d) i think if one is "pro" that idea, then one should incorporate
ai-agents into the new system, to help with things like explaining
errors, working on proofs, offering interactive help with the system,
etc.

[Joshua Scholar]

I just had an interesting idea. Say you decide that Julia's symbolic math system is better for your problem than Maxima is, then maybe you implement k-lambda in Julia and run your program in Shen on Julia using Julia's library.

If you want Maxima, then you use Marks version in Lisp and connect directly to Maxima.

And if you need Mathematica, maybe you can implement K-Lambda in Wolfram and connect to Mathematica natively. 

Maybe it's silly but for systems that are lisp like, that has to be a temptation.

I don't think you need Shen in C++ to interface with GiNaC. 

--
You received this message because you are subscribed to the Google Groups "Shen" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qilang+un...@googlegroups.com.

To view this discussion, visit https://groups.google.com/d/msgid/qilang/CAJ7XQb7qVaXJbXxtJ3bw1hXNCJj0tmkgjugD%3DiN-Ct2OO2C5oQ%40mail.gmail.com.