Nonsimplifying nounforms: abs, realpart, carg, etc. (maxima-Bugs-902290)

Dieter,

I agree that those are the three implementation options, which
correspond to three different kinds of behavior for the user.

If you agree with my arguments above, that the noun and the verb in
the case of realpart/etc. should be distinct, then only implementation
(2) will work.

As you point out, if we *do* decide to make it both a simplifying
function and a verb (which is what I propose), then we will have to
decide exactly what simplifications to incorporate.

>   ...What should be improved e.g.
>   'realpart(z+w) --> realpart(z+w) and not realpart(z)+realpart(w)?

I don't think that would not be an improvement.  In particular,
realpart(z+w)-realpart(z) would not simplify to realpart(w).  The
general simplifier works best when arithmetic operations are exposed
for further simplification.

              -s