PyCalCal implements all calendrica 3.0 algorithms.
They are tested against some dates mentioned in
the book and more importantly against the tables
in Appendix C.
All tests pass.


The failing tests were:
 testDynamicalFromUniversal          clarified by Profs's emails 2009/08/06
 observational_hebrew_from_fixed     FIXED 2009/11/15: trig funcs values!
 fixed_from_observational_islamic    FIXED 2009/11/15: trig funcs values!
 astro_hindu_lunar_from_fixed        FIXED wrong reference data array
 fixed_from_astro_hindu_lunar        FIXED 2009/11/17 wrong grouping in 's'
                                         value correct is (mod.. - deg...)
It has been quite difficult to find why observational Hebrew convertions
were failing. I had to check the full chain of calls and compare results
of pycalcal with calendrica 3.0 ones.
I was put off track by the statement in the Errata (Jun 2009) about
principal values of trigonometric funtions: Prof Dershowitz confirmed
that calendrical 3.0 expects/returns angles in interval [0,360).


Todo list
=========
see http://code.google.com/p/pycalcal/wiki/ToDoList
