Hi all
It seems I found a bug. The current cvs maxima is unable to give a correct
answer to integrate(t^2*sqrt(1-t^2)/(1+b^2*t^2),t,-1,1), where b is real. I
belive the mistake occurs on simplification step. I have this impression
because by default the integral is expressed through some radicals and
logariphms with %I inside and maxima asks if b is pos or neg.
The correct answer is %pi/b^2*(1-1/sqrt(1+b^2))
rgds,
v