I agree the title is somewhat misleading as it relates to the case that is being discussed in the article and NOT at all for software patents in general (which are not granted in many countries). The supereme court already went past the machine-or-transformation test in the well-known Bilski vs. Kappos / State Street bank case as others also quoted below. It left the tough part to the patent applicant and the examiners to exercise rigorously the patentability clauses under sections 101, 102, 103 & 112 in USPTO's MPEP. The arguments made from each side in the interpretation of these sections to the patent application in question holds the key whether that patent gets approved or not. It seems to me that this process has not always been objective with USPTO when you see some patents that get away with outrageous claims!
The title of this article, as well as at ther very least the opening paragraph are seriously misleading, if not outright incorrect. The ruling by the Supremes pertains only to the particular software covered by the Alice Corp. patent. In particular former USPTO Commissioner of Patents Davis Kappos states "From the perspective of the parties involved, this week’s Alice Corp. v. CLS Bank decision held that a process that lessens settlement risk for trades of financial instruments is too abstract for patenting. However, to the leagues of interested onlookers holding their collective breath across our country and indeed around the world, the Supreme Court’s unanimous ruling subtly conveyed a much more significant judgment: software, as a class, is every bit as worthy of patent protection as any other medium in which innovation can be practiced." See http://patentlyo.com/ for further discussions and comments.
Operating Systems are not abstract ideas like a push button that represents a way to interact with the system (remember abstract window toollkit), that's we missunderstood of a patent, applications are abstract ways to represent a property of an artifact that consumes resources or allocates them, software creates an interface with hardware in the most pure form, which is patentable because it's a physical artifact, software systems are complex tools that represent not an abstract idea, but a ***physical***description of a processor, that can be a digital signal processor, or a logical inference machine system like a type system that analyzes an abstract machine (the program) in a logical inference machine from the point of view of how a real machine must be executed under program control, so it gives the control steps for a program, and so it's part of the operating system. Unfourtunately typeless programming languages such as assembly or other languages can't control at any point of it's execution that and depend on human control which is not something patentable, so computers in general sense are not patentable because they don't represent a useful trustable artifact for society unless they are correctly defined and automatically checked, As systems depends in all cases of a power control unit, such device is patentable and it's biggest headache for system designers
IMO the most intelligent decision on the "software patents" issue was Diamond v. Diehr, which as I understand it (IANAL) stated that software could be patented as part of an invention that passed the "machine or transformation test". That is, the invention as a whole had to be a particular machine (not a generic computer, though it could have a generic computer as a component) or it had to transform matter in a tangible way. In Diehr, the patent was for a rubber curing process so it involved a tangible transformation.
Unfortunately, Bilsky and now Alice have muddied the waters, replacing the fairly simple and concrete "machine or transformation" test with whatever "abstract" means to the courts.
Personally (again, IANAL), I see that the US Constitution states that the purpose of patents is "to promote Progress of Science and Useful Arts". IMO each patent awarded should clearly show that its award in fact promotes Progress and not the opposite. IMO too many patently ridiculous patents are awarded, and their presence gives their owners too much opportunity to retard Progress of Science and Useful Arts.
IMO, the only time USPTO should award a patent is if the invention is something that took a great deal of time and/or money to develop, but once discovered is cheap and quick to copy. This is never the case with software -- if it takes a lot of time to create the original software, creating a duplicate from scratch would take approximately the same time. Sure, someone could copy the binary, but then they'd be infringing copyright. For software, IMO copyright is plenty of protection.
RTL-based inventions are an interesting question. Here's my IANAL opinion: while in some sense an FPGA is like a general-purpose computer, in most cases it's part of a specific machine so the overall machine can be patented under Diehr, and this would allow the RTL to be patented for that specific machine. OTOH, if the FPGA is a generic component of a generic computer, such as the FPGA in Bunnie Huang's open-source Novena laptop, then IMO it's not a specific machine and RTL written for it would be like software and not patent-worthy.
>> Don't we in the tech industry think of hardware and software distinctions as blurred these days? A chip is several lines of RTL code.
For digital systems, Yes. For analog systems, NO. However with time the software will eat more of the hardware. The fact remains some of the coolest ideas are evolving out of software and in that case, that is the best place to be. You can have the best hardware but only the best software will make people connect and see the best in it
>> I think that patents are necessary to prove the product of an individual or an organization.
No, patents do not prove anything. It is more of a commercial vehicle than a technology tool. It offers exclusivity which provides incentives to innovate. I support patents but I hate the way it works at the moment because if Pythogoas, Euclides etc had all patented their equations, we may not have modern science as we do today
This will start the debate on what can be patented and what not. I guess its a tricky situation for software and hardware systems. But whatever is the case patent lawyers and engineers will always remain busy.
A Book For All Reasons Bernard Cole1 Comment Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...