% From: Jarmo Hietarinta <hietarin@utu.fi>
% Date: Sunday, February 04, 2001 11:26 PM

% The following is not reduced to the simplest form.

Comment from Eberhard Schruefer:  This a longstanding problem.  To get
a low complexity algorithm isn't easy (even if only domain elements are
involved).  James and I tried something (for domains) almost ten years
ago along the line of Schinzel's theorem which has promisse of being
efficient;

(e^((3x)/4)+2*e^(x/4)+e^(x/2) - 3)/(e^((3x)/4) - e^(x/4) - e^(x/2) + 1);

% on combineexpt;

num df(ws,x) - 6*e^x -4*e^(x/2) - e^(x/4);

end;

