RE : Re: simplification of radicals



----Razafy Paul <paulrazafi at yahoo.fr> wrote: -----

> radcan will simplify them to zero I think!

Yes, radcan will simplify sqrt(3) * sqrt(17) - sqrt(51) to zero. But
radcan has to be used cautiously:

 (%i75) eq : sqrt(a*b)= sqrt(a)*sqrt(b)$

 (%i76) is(eq);
 (%o76) false

 (%i77) is(radcan(eq));
 (%o77) true

 (%i78) subst([a=-1,b=-1],eq);
 (%o78) 1=-1

So radcan should not be used for general simplification (on constant
expressions such as sqrt(3) * sqrt(17) - sqrt(51), radcan is ok, I think).

Barton