pep proposal : A date object for the standard library

[Robin Becker]

It seems odd to define dates as a duration in days from a fixed time and
then define durations in a different scale ie seconds. Wouldn't it be
better to adopt the integers to represent days and the fraction to
represent fractions of a day which is what various proprietary date
formats do already (although with different origin). Dates and durations
could then have a common underlying representation.

Of course in the modern world we need nano-seconds and luckily
1+nano != 1 in ieee arithmetic

so the nano second expressed in days is not effectively zero, but the
pico second is off the scale and cannot be added to days with any
accuracy.


No matter what format is used eventually all compact date
representations run out of accuracy in one direction or another so
wouldn't it be wise to establish the representation limits in years and
fractions of a second and make these a required part of implementations
of AbstractDate derived classes.
-- 
Robin Becker