>>>>> "Edwin" == Edwin Woollett <woollett at charter.net> writes:
Ted> Maxima 5.28.0-2 http://maxima.sourceforge.net
Ted> using Lisp GNU Common Lisp (GCL) GCL 2.6.8 (a.k.a. GCL)
Ted> /* This first erf only needs one bfloat followed by one expand: */
Ted> (%i1) bfloat(erf((sqrt(2)*%i+sqrt(2))/2));
Ted> (%o1) erf(5.0b-1*(1.414213562373095b0*%i+1.414213562373095b0))
Ted> (%i2) expand(%);
Ted> (%o2) 4.741476366409942b-1*%i+9.692642119442159b-1
Ted> /* but this erf needs bfloat, expand, bfloat to do the job: */
Ted> (%i3) bfloat(erf((sqrt(2)*%i-sqrt(2))/2));
Ted> (%o3) erf(5.0b-1*(1.414213562373095b0*%i-1.414213562373095b0))
Ted> (%i4) expand(%);
Ted> (%o4) 1.020646306179612b0*%i*sin(2.5b-1*%pi)
Ted> -3.50100287874768b-1*sin(2.5b-1*%pi)-3.50100287874768b-1*%i*cos(2.5b-1*%pi)
Ted> -1.020646306179612b0*cos(2.5b-1*%pi)
I think this is an artifact of the implementation of erf for complex
bfloats. I'll look into it an fix it soon. (Especially since the
current git version generates an error in this latter example!)
Ray