I don't know where the NaN is being manipulated, there are lots of design issues that could be addressed using it.
Note that there are huge numbers of NaNs -- a NaN has a reserved exponent; the fraction part can be used to encode info. You might also see how infinities are handled. A NaN can be thought of as resembling "0/0" if you want to put it into the rational number almost-field. Then you have only one of them. There are papers that describe what you might do. Here's one, about 20 years old. http://www.cs.berkeley.edu/~fateman/papers/retrodiag.pdf On Mar 22, 9:14 am, Robert Bradshaw <rober...@math.washington.edu> wrote: > On Wed, Mar 21, 2012 at 7:39 PM, kcrisman <kcris...@gmail.com> wrote: > > > On Mar 21, 10:00 pm, Gonzalo Tornaria <torna...@math.utexas.edu> > > wrote: > >> In sage 4.8: > > >> sage: NaN - NaN > >> sage: 0 > >> sage: NaN + NaN > >> 2*NaN > >> sage: NaN * NaN > >> NaN^2 > > > Naturally! Since NaN is a symbolic expression, apparently: > > > sage: type(NaN) > > <type 'sage.symbolic.expression.Expression'> > > > See sage/symbolic/constants.py, e.g. > >http://hg.sagemath.org/sage-main/file/c239be1054e0/sage/symbolic/cons... > > Based on the code in that file, I think that means this is the Pynac > > NaN. > > This is horrendous. Better to not have this "constant" (at least not > in the global namespace) than behavior like this. The top level NaN, > if we want one, could be in RR (precision doesn't really matter for > this value). > > - Robert -- 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
