Ballmer is probably correct, since the executive director of the Public Patent Foundation and senior counsel to the Free Software Foundation already admitted back in 2004 that Linux appears to violate many software patents, he claimed 283 likely infringements in the Linux kernel alone.
http://www.infoworld.com/article/04/08/02/HNmspatentsthreat_1.html
Yeah, yeah, yeah, we know. Also notice he said nothing about the validity of those patents. From the software patents I've seen, probably over 90% of software patents are junk (and that's not counting that software patents shouldn't even exist according to the USPTO).
As I said before, my program, conceived, designed and written by me using no "inspiration" from other sources aside from standard programming techniques I've learned ("here's how you normally save an XML file"), probably violates at least one software patent. I doubt there is even one large piece of software out there that doesn't violate several software patents.
Did you know there's a guy out there who got a patent in '98 based on what others were already doing using free software? Yep, then as they improve their free products he amends the '98 patent to include what they've done. Now he's suing them for patent infringement on things they did recently, saying it violates a patent issued in 1998.
Truth is that unlike other items, there is absolutely no standards (other than there not being an existing patent) when it comes to software.
In theory, I could copy a piece of code from the Internet that displays "Hello World" and get a patent on it (assuming somebody else hasn't done that yet).
Is it possible that the Linux kernel could have infringements? Possibly. Even antiRepublicrat mentioned that his clean-room program could very well be violating at least one patent--so it's a possibility.
OTOH Iggle, what has MS come up with that's actually innovative besides Microsoft Bob and Clippy?