Wired: Software Writers Patently Enraged. In a nutshell, software company A is awarded a patent for a specific method of doing something that lots of other software companies do (in this case, encrypting and decrypting documents). Company A then goes after Software Companies B and C to request license fees for infringement. Companies B and C argue that the patent should never have been issued, since their products considerably pre-date the patent application and since at least one has published “prior art” describing the technique in question.
There have been variations of this story written many times in recent memory, including such lulus as British Telecom claiming a patent on the hyperlink. In the Rambus case, the machinery of justice seems to be grinding against obviously fraudulent patents. But the more severe problem for most software companies seems to be these patents that many would argue never should have been issued in the first place.
Am I missing something? It seems like the PTO should be finding the prior art and refusing to issue the patents. Without this, all software writers would seem compelled to develop enormous patent application libraries on the off chance that someone will come along and patent something they wrote five years ago–a kind of IP Mutually Assured Destruction?