[Axiom-math] Re: [Axiom-mail] Dynamically constructed return types

[Martin Rubey]

Dear Marcus,

Marcus Better writes:

 > Unfortunately it still won't compile, but I hope this is just a matter of
 > correct syntax. Here is the code, I would be grateful if someone can give me
 > a hint:

Advise: 

* When I start programming a function, I always start with a nearly empty
procedure body, which returns some arbitrary value (for example, 0) in the
goal-domain. Then I add functionality bit by bit, such that it always
compiles. Otherwise I experienced unintelligible compiler errors, which didn't
make any sense.

* never use tabs. If you do, please untabify before you send it per email!

* unfortunately, I don't understand the code below too well. What would you want
  to return, if there were no compiler problems?

* I'm rather weak in Algebra. Could you do a very simple example for me, just
  that I realize what the resulting domain would be? I understand that 

K:=SAE(FRAC INT, UP(x,FRAC INT), x^2+1::UP(x,FRAC INT))

  should be the domain of complex integers, shouldn't it? If I extend this
  domain again,

L:=SAE(K, UP(y, K), eq::UP(y, K))

  what do I get? Isn't this again a SAE of FRAC INT, only with a different
  defining polynomial? If this is correct, Axiom doesn't know this, does it?

Martin