From: Mel Bartels 
cc: atm{at}shore.net
Reply-To: Mel Bartels 


Back in the late 1980's I started with spherical trig conversion functions.
 When 286 or AT class computers became available, I tried Taki's matrix
conversion routine using directional cosines.  For several years I had both
conversions running side by side in software, instantly selectable. 
Formula included conversions to equatorial coordinates, altazimuth
coordinates, field rotation, axis misalignment, and so forth.  I abandoned
the spher trig functions as too brittle.

In the real world I found the small inexactitudes hard on the spher trig
functions.  They would blow up, or show larger deviations from ideal than
then matrix conversion routines.

Polynomial correcting terms to spher trig is an interesting idea.  I am not
sure that they would not end up more complicated and esoteric than matrix
conversions.

Mel Bartels

--- BBBS/NT v4.01 Flag-4