There's this one: http://www.google.com/patents/US20100332285?dq=US+2010033228...

describing a computerized system for managing IP; its descriptive matter includes sentences like this: "The patent creation component 430 defines the process for enabling the client organization to develop and protect inventions that are new, useful, and non-obvious." and one of the elements in its claim 5 is "a fourth electronic IP creation component for filing and prosecuting patent applications, trademark applications, and copyright applications".

I don't think so, but that's a really challenging thing to find via search engines.

Judges are more familiar with law than they are with other fields. Making software look like statutory material by careful drafting of a patent application may be possible, but I imagine making a legal maneuver look like statutory material for a patent is much harder.

I have often thought about trying to patent the patent troll process.

Yeah, sorry but there is prior art.

But I thought prior art doesn't matter anymore? In 2013 the US patent system changed from first-to-invest to first-to-file.

I am not a lawyer, this is not legal advice, if you're actually in a situation where it matters you should check with a real lawyer, etc.

This is a common misconception. First-to-file does not mean that prior art doesn't matter. First-to-file means that, if I invent something today, you invent it tomorrow, you try to patent it the next day, and I try to patent it three days from now, nobody gets the patent. In first-to-invent, I would get the patent.

The 2013 law explicitly still gives weight to prior art.

Prior art still invalidates a patent. First to file only comes up when you have to people inventing the same thing at similar times.

You can invent something without it being prior art if you don't publish how to do it, file a patent, or use/sell it in commerce.

So if you are testing a new invention in secret and then the University of Wisconsin files a patent on the same thing, they get the patent. If someone else had already sold the products or described it publicly, nobody could get a patent on it (except for the person who sold it and then only for a year after they sold it).

That is not at all what first-to-file means. Depressingly common misconception though.

That doesn't stop software patents on literally millions of long established practices.

Can we get a patent on prior art itself?

Only if there is no prior art.

So it's prior art all the way down?

Also, does the patent on all prior art cover itself?

IANAL, but I doubt you can patent recursion either.

"A method for defining an computational operation, performed by a computer system or a mechanical equivalent of thereof, in terms of the same computational operation performed by the identical system. Such computational operation is defined in such a way it consists of two separate computational suboperations - one and only one of them is defined completely in terms of the composed computational operation itself."

Sounds like it could fly with USPTO.

(And if it doesn't, you can try and confuse people by patenting Y combinator.)

Halliburton at one point attempted to patent it. http://www.google.com/patents/US20080270152