> ratsimp(specint(t^(-1/2)*%e^(-2*a^(1/2)*t^(1/2))*%e^(-p*t),t));
Actually, general-purpose integrate gets this one, too... This points
out a problem for new users of Maxima: having to know that if general
integration doesn't work, they might want to try specint; same thing
for limit/tlimit, simpsum/nusum, etc.
It turns out, by the way, that the return values of specint and
integrate are different, but numerical evaluation shows that they
differ by a constant (phew!).
-s