factor simplifies it most of the way. Unfortunately, factorsum doesn't
handle rz*z+ry*y+rx*x+r*rz^2+r*ry^2+r*rx^2 =>
rz*z+ry*y+rx*x+r*(rz^2+ry^2+rx^2) because rx, ry, rz appear in both
the r*xxx term and the rest.
You could also try playing with ratfac and ratvars variable ordering.
> I notice that for (slightly) simpler integrals, maxima does simplify the
> result, so it seems it has hit a max simplification-search-depth somewhere,
> given up, and dumped the not-simplified result.
No, this is not the way it works. There is a standard simplification
algorithm which does no combinatorial search.
> Is there a variable I can set so that maxima will try harder to simplify?
> (at the moment, the results comes just about instantly).
No.
-s