Same here(though not a mathematician). Patenting algorithms is so horribly wrong on every level, it's still beyond my comprehension how this could ever be allowed. Some of the patents are ridiculously absurd and luckily rarely enforced. I remember the guy on sci.crypt who had a patent on "cascading ciphers", which covered almost any cipher build out of smaller cryptographic building blocks. It basically concerned almost any stream cipher mode and any attempt of combining more than one building block with XOR somewhere. Luckily, the guy never had the money to enforce it although he might have earned some royalties from doing nothing.

Heck, Siemens used to have a German patent on the Internet, except that it wasn't called Internet but something like "making available of structured textual data representations via long-range data transmission", etc.