[Maxima-discuss] operator precedence & display, was: Stop simplifying a subexpression

[Robert D. <rob...@gm...>]

On Fri, Aug 8, 2025 at 1:50 AM Oleg Nesterov <ol...@re...> wrote:

>         (%i1) '"+"(2,2) = 4;
>         (%o1)                              2 + 2 = 4

Huh, I'm a little surprised (although not very much) that this has the
expected effect ... I guess that's a %MPLUS expression instead of
MPLUS, and the simplification rules for MPLUS aren't also implemented
for %MPLUS. OK by me.

>         (%i2) '"."(matrix([1, 2], [3, 4]),  matrix([1], [2])) = matrix([5],[11]);
>                                   ┌      ┐     ┌   ┐    ┌    ┐
>                                   │ 1  2 │     │ 1 │    │ 5  │
>         (%o2)                    (│      │) . (│   │) = │    │
>                                   │ 3  4 │     │ 2 │    │ 11 │

Here the issue is that the operator is %MNCTIMES which does not have a
right or left binding power (precedence) defined for it. Try this:

:lisp (setf (get '%mnctimes 'rbp) (get 'mnctimes 'rbp))
:lisp (setf (get '%mnctimes 'lbp) (get 'mnctimes 'lbp))

which gives %MNCTIMES the same precedence, and therefore parentheses
in the same places, as MNCTIMES.

>         (%i4) a = b . c;
>         (%o4)                              a = b . c
>         (%i5) '"="(a, b.c);
>         (%o5)                             a = (b . c)
>         (%i6) :lisp (defprop %mequal 80. lbp)
>         (%i6) :lisp (defprop %mequal 80. rbp)

Likewise I guess we would say

:lisp (setf (get '%mequal 'rbp) (get 'mequal 'rbp))
:lisp (setf (get '%mequal 'lbp) (get 'mequal 'lbp))

I don't see why the nounified operators should not have the same
binding powers as the un-nounified -- I will try to remember to update
grind.lisp.

Robert