I've finished checking in the Bessel routines and stuff, and I've
added some preliminary support for symbolic Bessel functions.
Here's a short transcript:
(C1) bessel_y[1](3.0);
(D1) 0.3246744247918
(C2) bessel_i[1](3.0);
(D2) 3.953370217402609
(C3) bessel_y[3/2](z);
COS(z) SIN(z)
SQRT(2) SQRT(z) (------ - ------)
z 2
z
(D3) ---------------------------------
SQRT(%PI)
(C4) bessel_j[3/2](z);
COS(z) SIN(z)
SQRT(2) SQRT(z) (------ - ------)
z 2
z
(D4) - ---------------------------------
SQRT(%PI)
It doesn't know about derivatives and such, but it should be easy to
add since maxima knows how to do that for %j.
Currently, Bessel functions of half-integral order are always expanded
out in terms of elementary functions. Should there be a variable
controlling this?
Ray