bear wrote:
> So if we're going to have a NaN read/write syntax, we *MUST* require
> implementors to guarantee that, no matter what, their reader will
> never throw an error and return a NaN.

Okay, I'm comfortable with that.

Perhaps we need to defined a general invalid object which has (a) no
read syntax and (b) is returned by read whenever it has a problem with
the input (numbers too big, objects to big, bad syntax, etc.)

Regards,

Alan
--
Dr Alan Watson
Centro de Radioastronomía y Astrofísica
Universidad Astronómico Nacional de México