On Thu, 30 Sep 2004, Patrick Wallace wrote:
> Does Java provide floating point and trig at the usual ~16dp?
Yes. Doubles are 64-bit IEEE 754 and the trig operations are defined
to be within 1 ulp of the true value, where a ulp is a 'unit in the
last place'. Not certain whether that's in decimal or binary, but
either way you should get about 16 sig figs.
--
Mark Taylor Starlink Programmer Physics, Bristol University, UK
[log in to unmask] +44-117-928-8776 http://www.star.bris.ac.uk/~mbt/
|