RE: [Axiom-developer] Exact Number Domain

[Bertfried Fauser]

Dear Tim,

a very cute question indeed. The approach proposed by Bill Page is quite
interesting in that way, that the completition of a number field can be
done in different ways using ideas from topos theory. I don't even see how
to manage the wealth of possibilities here. However what about truncating
computations in a sort of p-adic way along with a stream like
reprsentations?

To pick up Ralf Hemmeke's example, if you want to represent pi, you might
want it to expand in powers of some prime number, eg 3

pi= a1*3^1+a0*3^0+ .....

a stram like object could represent the n-th power truncated version of
such expansions. The inductive limit of these objects would actually (but
infinitely) represent the number.

Over all it seems to be very complicated and I would not expect such a
representation to be helpful for numeric calculations (speedwise) at all.

ciao
BF.

% PD Dr Bertfried Fauser
%     Institution: Max Planck Institute for Math, Leipzig 
<http://www.mis.mpg.de>
%   Privat Docent: University of Konstanz, Phys Dept 
<http://www.uni-konstanz.de>
%  contact|->URL : http://clifford.physik.uni-konstanz.de/~fauser/
%          Phone : Leipzig +49 341 9959 735  Konstanz +49 7531 693491