is there a deep reason ratsimp and trigsimp act differently
in the effective interactive ev(...) examples:
(%i1) display2d:false$
(%i2) integrate(x^2*erf(x),x);
(%o2) (%pi*x^3*erf(x)+(sqrt(%pi)*x^2+sqrt(%pi))*%e^-x^2)/(3*%pi)
(%i3) diff(%,x),ratsimp;
(%o3) x^2*erf(x)
(%i4) integrate(sin(x)^3,x);
(%o4) cos(x)^3/3-cos(x)
(%i5) diff(%,x), trigsimp;
(%o5) sin(x)-cos(x)^2*sin(x)
(%i6) trigsimp(%);
(%o6) sin(x)^3
Based on the ratsimp use in the first example
I would have expected the trigsimp use in
the second to have done trigsimp(diff(...) ).
Ted Woollett