Again it is to help prevent abuse I don't really see how this is a valid concern? Tokenization is actually fairly CPU intensive