I think you guys will want to see this extraordinarily interesting talk by Carl Hewitt on YouTube. He talks about the future of IP software -- future as in when computers start filing patent applications. "If there is no principled way to distinguish computation from human thinking" -- then what? Human thinking isn't patentable. And if the meaning is a mathematical denotation, and mathematics is also unpatentable, what happens when computers are filing thousands and thousands of patent applications on all the patentable IP in the universe?
A subtheme is that Mark Lemley, who thinks software patents are acceptable as long as we tweak the USPTO process to screen out claiming the function instead of the way to get to the goal, is suggesting a halfway measure, because the future for the US patent system is a patent quagmire, because of the future capabilities of computers to overwhelm the USPTO with patent applications, none of which should be patentable but which under the current regime would be.
Here's a bit of the section on why he thinks Mark Lemley is wrong, from the transcript on YouTube:
7:13... there's some halfway measures being proposed That should whet your appetite.
7:16a good friend and colleague a colleague and good friend Mark Lemley has proposed
7:20that software patent abilities okay
7:23if the computational processes are specified precisely
7:27right what he says is these guys they're they're claiming the goal
7:31not actually at actually explaining how
7:37and I think that mark is wrong I'll try to explain
7:40why the reason is the technology has moved
7:44on since the days square root and sort
7:49and we now have highly precise operational constructs in which
7:54computers on their own
7:56set goals have strategies for achieving in assessing goals
8:01dealin' conjectures and metaphors and analogies