On Sun, 20 Jan 2002, [iso-8859-1] Anselmo P?rez Serrada wrote: > We all know that we need just two points to draw a stright line (like > the solar time ones), five points to draw an hyperbola (like the > datelines) but... > > Which is the minimum number of points (ie., of dates) we need to > draw an analemma (in a flat dial) within a worst-case precision of, > say, 1 minute of time? > > I am trying to solve this question using Bezier curves to approximate > the analemmae and brute-force to find the best solution, but it is not > so easy to find a general one (for any position of the dial)... > > Any suggestion? Or maybe my question is useless because the classical > twelve zodiacal dates make a good balance 'generality vs. precision'? > > [ 41.63 N 4.73 W]
I don't have a good answer for the analemma, but a feeling that this complex multiple-sinusoidal curve may require more than a few points! On your first point though, I thought any quadratic curve (including the hyperbola) could be defined through three points? Dave 37.29N 121.97W
