News for hackers. Stuff that matters.
Re:Judge Says Mathematical Algorithms Can’t Be Patented by JoshTriplett on Thursday March 28, 2013 @03:15PM (#5457031)
Because you can't directly patent an algorithm, software patents use all sorts of fascinating workarounds to pretend like they've patented something else instead. A typical software patent will include the same patent claims written several different ways, based on legal precedents used with software patents in the past. A patent on an algorithm for fooing bars via steps A, B, and C, with the additional claim of doing D while fooing, might have one set of claims:

1. A method for fooing bars, comprising A, B, and C.

2. A method in accordance with claim 1, wherein said fooing further comprises doing D.

3. A system for fooing bars, comprising A, B, and C.

4. A system in accordance with claim 3, wherein said fooing further comprises doing D.

5. A computer program product for fooing bars, comprising A, B, and C.

6. A computer program product in accordance with claim 5, wherein said fooing further comprises doing D.

This covers several different legal precedents; even if "method" gets invalidated as an approach, the patent may still work under the "system" or "computer program product" theories.

This gets taken to a ridiculous extreme; many such patents describe the computer program product as a physical object consisting of bits on a medium such as a CD or disk, and include as a figure an illustration of such a medium.

Re:Judge Says Mathematical Algorithms Can’t Be Patented by diminoten on Friday March 29, 2013 @02:57AM (#5459610)
It's not pretending, they actually are patenting the method rather than the code itself. It's a critical distinction, as code is actually copywritten.

There's a huge difference between design and implementation.