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

[Stavros M. <mac...@gm...>]

Robert, I was surprised that the noun-operators work just fine....
But '"=" seems to be the one bad case:

' "*"(a,' "+"(b,c),d)  =>  a (b + c) d    << parens necessary
' "+"(a,' "*"(b,c),d)  =>  a + b c + d    << parens not necessary


' "."(a,' "+"(b,c),d) => a . (b + c) . d
' "+"(a,' "."(b,c),d) => a + b . c + d

"="(a,b+1) => a = b + 1       << correct rendering with regular "="
' "="(a,b+1) =>  a = (b + 1)   << BAD, parens are not necessary


On Sun, Aug 10, 2025 at 6:41 PM Robert Dodier <rob...@gm...>
wrote:

> 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
>
>
> _______________________________________________
> Maxima-discuss mailing list
> Max...@li...
> https://lists.sourceforge.net/lists/listinfo/maxima-discuss
>