Hi,

The expression a*(vectorize(λ* b )†c) displays correctly, if directly written to a math region, also if it is wrapped in Maxima(). Maxima doesn't do a lot here, all is done on the SMath side.

However, if it is used in a definition, then SMath complains about wrong arguments for the cross product.

How can I avoid this?


Plugin Maxima Test.sm (450 KiB) downloaded 36 time(s).

Assuredly, the 1rst expansion is correct in the Smath native vector algebra.
It reminds me Tom G. [Mathsoft Collab] reported Mthacad 11 x product was
different than in previous Mathcad 2000.

As for my point of view, the wrong thing is the result given in (vectorize(λ*a))†b; since λ is unknown (and there fore type of λ), vectorize() it should be hold in the formula until will be defined.
Just think if I'd like to use a vector as λ:

I think the given symbolic engine of SMath assumes scalars for all variables unless they have a non-skalar value. It would be convenient to have some type declaration objects (along with some symbol decoration), but as long as we don't have them, we can't expect symbolic vector equations to be handled correctly.

Section matrix operations symbolic.sm (16 KiB) downloaded 32 time(s).

Wrote

I think the given symbolic engine of SMath assumes scalars for all variables unless they have a non-skalar value.

Maybe, but here is some side effect of the evaluation operator (=) RHS evaluation logic (just a guess, I have to check it with the debugger).
A normal assignments works as I'd expect:

Wrote

It would be convenient to have some type declaration objects (along with some symbol decoration), but as long as we don't have them, we can't expect symbolic vector equations to be handled correctly.

I agree with you, maybe a dedicated multiplication operator for matrices (standard multiplication = 1 or both the arguments are assumed as scalars; matrix multiplication = both the arguments are matrices/vectors).