The Homomorphic Encryption patent land rush(blogs.teamb.com)
blogs.teamb.com
The Homomorphic Encryption patent land rush
http://blogs.teamb.com/craigstuntz/2012/04/04/38707/
4 comments
Lets get together and fool the patent lawyers into thinking that quantum teleportation of matter is just around the corner. Perhaps they'll spin their wheels adding "...but with teleporters!" onto the end of every conceivable human activity and let us makers get some work done.
Even better, if we can get speculative patents now on all sorts of future technology that will take at least 20 years to develop later, the patents will all be expired by the time the tech is ready.
The thing is, companies often keep patents alive by coming up with little variations or extensions of existing patents.
The fact that people can speculatively patent anything to begin with is a glaring sign something is severely broken. (Do a patent database search for Internet and Video. Note how many and how vague they are and how far the dates predate Youtube.)
The fact that people can speculatively patent anything to begin with is a glaring sign something is severely broken. (Do a patent database search for Internet and Video. Note how many and how vague they are and how far the dates predate Youtube.)
The functions necessary for some of these "inventions" might not even exist. So this might end up being almost as good in the end.
Granted, I am neither a patent expert nor a homomorphic encryption expert, but how is this patentable since homomorphic encryption is just a mathematical formula?
From Wikipedia [1]: Homomorphic encryption is a form of encryption where a specific algebraic operation performed on the plaintext is equivalent to another (possibly different) algebraic operation performed on the ciphertext
[1]: http://en.wikipedia.org/wiki/Homomorphic_encryption
From Wikipedia [1]: Homomorphic encryption is a form of encryption where a specific algebraic operation performed on the plaintext is equivalent to another (possibly different) algebraic operation performed on the ciphertext
[1]: http://en.wikipedia.org/wiki/Homomorphic_encryption
In theory, a mathematical formula is not patentable. In theory, a patent that precludes every practical use of a mathematical formula is invalid. In theory, a patent is required to be specific enough that a person of ordinary skill in the field could use it to reduce the invention to practice (a requirement that no homomorphic encryption patent could possibly meet.)
By "theory" I mean statute and Supreme Court precedent. These are just theory, not law. Law is not what the legislature says or what the Supreme Court says. Law is what happens to you in court.
By "theory" I mean statute and Supreme Court precedent. These are just theory, not law. Law is not what the legislature says or what the Supreme Court says. Law is what happens to you in court.
The general rule in the US patent system is that mathematical formulas are not patentable, and so neither are algorithms. But a device or system that uses an algorithm is patentable. So they don't patent the algorithm itself, but "a method and system". In other words they're patenting the idea of running the algorithm on a computer.
Insane, yes, but upheld by current precedent.
Insane, yes, but upheld by current precedent.
In the Bilski case, the Supreme Court reminded us that Benson, Flook and Diehr are still good precedent. So patenting the idea of running the algorithm on a computer is also not upheld by current precedent.
Oh, they also said to ignore State Street, the precedent that upholds software patents. But they're just the Supreme Court. Nobody listens to them in this area anyway. I know I sound sarcastic, but I'm serious.
Oh, they also said to ignore State Street, the precedent that upholds software patents. But they're just the Supreme Court. Nobody listens to them in this area anyway. I know I sound sarcastic, but I'm serious.
I had forgotten about Bilski. Still, if the courts don't listen to a precedent, is it really a precedent?
Yes, all you have to do is hire more lawyers than Apple/IBM/Microsoft or whoever.
Go to court in East Texas and persuade a jury that a bunch of damn Yankee judges up there in Washington know better than an honest American corporation.
Go to court in East Texas and persuade a jury that a bunch of damn Yankee judges up there in Washington know better than an honest American corporation.
As I understand it from talking to patent lawyers: they're patenting the specific use of the algorithm, not the algorithm itself. It's "use of such-and-such formula to accomplish security goal X in an electronic computer system" that's at play, not the fundamental math.
Just for fun, next time a patent lawyer tells you that, ask them to explain how such a patent is distinguished from Flook's patent. http://en.wikipedia.org/wiki/Parker_v._Flook
Will do!
That is not quite correct. You cannot patent something that precludes the use of a mathematical concept but you can patent a method for applying a mathematical concept. For example, you cannot patent "sorting" but you can patent a "bubble sort algorithm" because the latter is one of an infinite number of methods in which the mathematical concept of sorting can be implemented. The RSA patent did not prevent people from using public key cryptography (either in theory or in practice), only one particular algorithm for implementing such a concept.
Algorithm patents are allowed in most countries, contrary to popular belief; when most people talk about "software patents" they are actually talking about business process patents which rarely exist outside the US. Actual algorithms are recognized as a direct abstraction of electronic circuit designs and electronic circuit designs are widely recognized as patentable subject matter.
(Any consistent application of an "algorithms are mathematics" argument would exclude all currently patentable subject matter, for better or worse. I do not have a strong opinion one way or the other but it is worth pointing out because this fact is lost in many discussions of patents.)
Algorithm patents are allowed in most countries, contrary to popular belief; when most people talk about "software patents" they are actually talking about business process patents which rarely exist outside the US. Actual algorithms are recognized as a direct abstraction of electronic circuit designs and electronic circuit designs are widely recognized as patentable subject matter.
(Any consistent application of an "algorithms are mathematics" argument would exclude all currently patentable subject matter, for better or worse. I do not have a strong opinion one way or the other but it is worth pointing out because this fact is lost in many discussions of patents.)
Benson's patent was for a particular algorithm, not an abstract concept, yet it was found invalid: http://en.wikipedia.org/wiki/Gottschalk_v._Benson
The RSA patent didn't make it through the patent system by claiming a particular algorithm; it made it through by including the hardware: http://en.wikipedia.org/wiki/RSA_(algorithm)
See also http://en.wikipedia.org/wiki/Freeman-Walter-Abele_Test
The RSA patent didn't make it through the patent system by claiming a particular algorithm; it made it through by including the hardware: http://en.wikipedia.org/wiki/RSA_(algorithm)
See also http://en.wikipedia.org/wiki/Freeman-Walter-Abele_Test
Long multiplication isn't patentable
The use of a device (such as a pen) to write down the values as you do the multiplication is patentable
I wonder if this prevents this area being researched and discovered?
Not necessarily. There are an unlimited number of algorithms available to implement any particular concept so people are always looking for faster, more efficient algorithms for implementing the same mathematical concept.
However, the marginal value of doing the R&D to find additional algorithms diminishes as more are described even if they all are patented. This is not a side effect of patents per se but a side effect of how algorithm research works.
Companies often spend millions of dollars on the computer science R&D required to generate a new algorithm so there is a return on investment calculation. For the R&D to be worthwhile, they either need to produce an equivalent new algorithm for less money than licensing it or believe that they can find a materially better algorithm than the current licensable algorithms such that they have a market advantage. It is very similar to chemical process R&D.
However, the marginal value of doing the R&D to find additional algorithms diminishes as more are described even if they all are patented. This is not a side effect of patents per se but a side effect of how algorithm research works.
Companies often spend millions of dollars on the computer science R&D required to generate a new algorithm so there is a return on investment calculation. For the R&D to be worthwhile, they either need to produce an equivalent new algorithm for less money than licensing it or believe that they can find a materially better algorithm than the current licensable algorithms such that they have a market advantage. It is very similar to chemical process R&D.
Not so much in academia. But yes, this would hamper or delay practical/field deployment. See EC crypto etc.
ECC is widely used in industry. DJB seems skeptical of its patent encumbrances:
http://cr.yp.to/patents.html
I do agree there was a time when ECC patents were an issue (particularly Certicom's), but it was certainly not as big a deal as the RSA patent.
http://cr.yp.to/patents.html
I do agree there was a time when ECC patents were an issue (particularly Certicom's), but it was certainly not as big a deal as the RSA patent.
Well, if no solution exists, I am going to patent the business process of finding a solution to homomorphic encryption.
My process largely involves hiring mathematicians and buying lots and lots of coffee.
My process largely involves hiring mathematicians and buying lots and lots of coffee.