I just learned this rule from a math book yesterday :-)
Delta must be something like differentiation or something more general. My example was
mathematically meaningless. Also, in hyperdoc, there is no description about the
mathematical properties delta should have, so I think adding some words to hyperdoc about
"what delta is" is preferable.
It's great to see the patch, quick fix, thanks ! You mentioned the testcase before, but I'm sorry
it is not written yet. I tried the command in my axiom console, like this
)expose UnittestCount UnittestAux Unittest
but it comes out errors
UnittestCount is not a known constructor. You can make the
constructor known to the system by loading it.
UnittestAux is not a known constructor. You can make the constructor
known to the system by loading it.
Unittest is not a known constructor. You can make the constructor
known to the system by loading it.
Or maybe I should write a .input file to put all the test code in ? ...
Another question is: why we need to test "each" of the exported functions?
In my opinion, test the exponentiation related function is enough, is it ok?
PS: I know FriCAS is a fork of AXIOM. Is this patch will soon applied to
AXIOM? What's the difference between Axiom and FriCAS? I'm thinking about
which one is better for use.
Best Regards,
Xiaojun
2008/8/31 Martin Rubey
<address@hidden>
> (1)-> f: INT->INT:=x+->x+1
>
> (2) -> K := OREUP ( x, INT, 1, f)
Note that this violates the assumptions on sigma and delta - sigma has to be a
Ring homomorphism and delta has to be additive and in a certain way compatible
with sigma! I guess you wanted something like
f: UP(n, INT) -> UP(n, INT) := p +-> p.(n+1)
g: UP(n, INT) -> UP(n, INT) := p +-> 0
K := OREUP(E, UP(n, INT), morphism f, g)
M := E::K+n
Martin