bessel_i quadpack bug 5.26.0



On Feb. 29, 2012, Raymond Toy wrote:
-------------------------
>This could probably be fixed by changing how bessel_i(1,%i*x) is computed.
>bessel_i(1,3*%i) has a small real part >even though we know it must be
>purely imaginary.
>
>For the other bessel functions at %i*x the result is complex, except
>bessel_j.  For bessel_j, it looks like we are >careful and return a purely
>imaginary result.
>
>Or maybe tell maxima that bessel_i(n,%i*x) = bessel_j(n,x).  Maxima doesn't
>seem to know that property.  Maybe the >simplifier should honor %iargs to
>do this transformation.

the relation Mathematica has on page

http://functions.wolfram.com/Bessel-TypeFunctions/BesselI/16/01/01/

is

bessel_i(nu,%i*z) = (%i*z)^nu * bessel_j(nu,z) / z^nu

but I haven't checked this numerically.

Ted