Re: CVS commit: src/games/factor

[Joerg Sonnenberger <joerg%britannica.bec.de@localhost>]

On Sun, May 16, 2010 at 02:10:35PM +0400, Aleksej Saushev wrote:
> Following the logic Joerg uses, one should reject all arguments to sqrt,
> asin, acos, atan, clog, casinh, cacosh, and other inverse functions just
> because they have more than one branch. In "fundamental theory of mathematics"
> be it geometry, real or complex analysis, or anything else, this approach
> is found counterproductive.

Please check the definition of the functions you are using. You are
confusing basic mathematic properties. sqrt does not have more than one
branch. If you want to solve a quadrativ equation, you have to check the
different (complex) roots. Similar for the inverse functions -- they are
defined to return the principle value. This restriction is a simple
result of the trigonometric functions not etc not being injective.

Joerg