Smell test for the validity of software patents

While HN seems primarily negative on software patents, all of the recent patent-related stories here have brought out more patent supporters than I would have expected. Some of them are legally or business minded folk rather than technical, others seem to be caught up in a sort of 'fanboi' mentality of defending their favorite company.

To those who are pro-software patent (and understand at least the basics of software development), I would ask if they honestly think one or more people could sit down to write a complete software application for any device that performed at least one practically useful function without (knowingly or unknowingly) violating at least one existing patent (valid in the USA)? Because I honestly don't think it is possible to do that at this point in time and would bet a significant (to me) amount of money that it couldn't be done even if you were specifically trying to do it.

Despite all the talk of "innovation", so much of what is core to the fundamentals of computer science has been swept up in currently active patents that it doesn't matter if you innovate or not, you're absolutely going to violate other people's patents when you write any significant amount of code. It is just a matter or whether they notice and persue the matter legally or not.

With that in mind, I don't see how any rational person could conclude anything other than the fact that software patents are insanely out of control right now and need a seriously overhaul if not a complete tear-down.