[magellan]

Sometimes software patents are not even algorithms. They can be methods, as well as concepts. IBM has a patent on "ecommerce over the web", and there was a dispute between IBM and Amazon on this. Somebody else patented hyperlinking, which makes the entire Internet since the days of the Gopher protocol a giant patent violation.

I'm all for patenting new, real software inventions. But unfortunately, most software patents are B.S. They are not "inventions" per se.

Some software patents are bogused because they patent something which has already been invented by someone else. This is patent poaching. But the patent approval process collapsed under both ignorance of the software industry and the sheer number of patent applications.

Now everyone who has anything to do with creating software is patenting everything it can, so as to protect itself from patent poachers.

[BikerJoe]

“I’m all for patenting new, real software inventions. But unfortunately, most software patents are B.S. They are not “inventions” per se.”

Quite right. Patent law is quite specific about “prior art”. Unfortunately, it is well known that the Patent Office just dealt out software patents without doing much checking into the “prior art” argument. In fact, a lot of patents were given for things that 2nd year Computer Science students would independently design.

One famous example that made Windows perform poorly in the early ‘90’s was the AT&T patent on the “backing store”, which was simply copying the hidden portion behind a window to disk. Microsoft wouldn’t pay the royalty, which is their right, but Windows performance suffered because they didn’t come up with a better non-violating technique for a while.