An approach to doing all of this may be influenced by generic arithmetic packages in other languages, e.g. the generic stuff I wrote in Lisp that interfaces to, among other things, Lisp numbers (e.g. arbitrary precision integers, rationals, double- floats, complex =[pairs of previous]) various constructed number algebras (e.g. intervals [various kinds], quad-precision, bigfloats, and perhaps other things that I don't recall. maybe polynomials, functions expanded in Chebyshev coefficients, ..). In particular you get into trouble with allowing all kinds of types to interact -- what type is the result? You also need to figure out what to do with items like zero. Which type of zero? And you may have other objects, like infinities, undefined, and impossible.
While I assume that whatever I've written can be written in Python, perhaps somewhat stilted because of the intrusion of the limited Python native numbers and operators, but written somehow. This does not mean that any of the serious issues (like coercions, infinities, etc.) will be magically solved. If you do come up with a nice solution (minimally including intervals), I hope you write it up nicely. RJF On Sep 1, 2:02 am, Jean-Pierre Flori <jpfl...@gmail.com> wrote: > Dear all, > > During the coding sprints taking place at ECC2011 summer school, one > of our project might be to build interfaces to the three libraries > available at multiprecision.org (by Andreas Enge and others): > - MPC (complex numbers with arbitrary precision - an optional spkg > already exists, so that would only be an update - see trac #4446) > - MPFRCX (polynomials using the above lib and MPFR) > - CM (construction of class polynomials by floating-point > approximation using the above libs) > > I'd like to know if anyone has any suggestion or objection about/ > against that idea. > (A list of other potential topics can be found > athttp://www.loria.fr/~zimmerma/ecc.html) > > Best regards, > JP -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
