I happened to look at Arnold-Trivium-1991.pdf again last night, and
tried the limit therein in maxima to see how well it would do.
It took about 20 minutes of CPU time using sbcl.
(A url to the pdf was posted some time ago in, I think, math-fun.)
(%i1) f : (sin(tan(x))-tan(sin(x)))/(asin(atan(x))-atan(asin(x)));
(%i2) limit(f,x,0);
It returned 1.
Which looks right, given plot2d(f,[x,-1,1]).
Is it expected that maxima would take so long to calculate such a limit?
-JimC
--
James Cloos <cloos at jhcloos.com> OpenPGP: 1024D/ED7DAEA6