Hi;
I have been trying to do these simplifications:
First, though, a little set up:
(%i1) kill(all);
(%o0) done(%i1)
(%i1) trigsign:false$
matchdeclare (aa,all);
(%o2) done
(%i3) :lisp (setf (get '$gam 'wxxmlword) "<g>%gamma</g>")
<g>%gamma</g>
so that
(%i3) 'gam;
(%o3) %gamma
is pretty in wxMaxima.
It happens that
(%i4) gam:2*%pi/3$
which allows us to see that
(%i5)is(expand(exponentialize(ev(cos(aa)-sqrt(3)*sin(aa)=-2*cos(aa-
ev('gam))))));
is(expand(exponentialize(ev(sqrt(3)*sin(aa)-cos(aa)=
2*cos(aa-ev('gam))))));
(%o5) true
(%o6) true
The difference between these is in the sequence of the terms, and the
sign.
Now
(%i7) tellsimpafter(cos(aa)-sqrt(3)*sin(aa),-2*cos(aa-'gam))$
tellsimpafter(sqrt(3)*sin(aa)-cos(aa), 2*cos(aa-'gam))$
One works:
(%i9) cos(aa)-sqrt(3)*sin(aa);
(%o9) -2*cos(aa-%gamma)
and the other does not:
(%i10) sqrt(3)*sin(aa)-cos(aa);
(%o10) sqrt(3)*sin(aa)-cos(aa)
(Using sqrt(3)/2 instead of the pretty gamma does not help.)
The question is, "Why not?" Is something re-organizing things after
tellsimpafter?
Thanks for looking,
Doug Y