2

(I originally posted this on law.SE and was directed here.)

Patent US4309569, published in 1982 and expiring 20 years later, describes an algorithm for creating hash trees (known as Merkle trees, after its inventor), used for data validation. How is this patentable?

The US has granted "software patents" since the 70s, but my understanding is that this does not extend to obvious/trivial algorithms or mathematical formulae. In my view, hash trees -- and, thereupon, the application of the mathematical formulae presented in the patent -- are entirely obvious and non-novel. I believe this would be the view of any computer scientist or seasoned programmer; it therefore strikes me as unlikely that it was never independently reinvented (before or since), hence invoking a legal challenge.

  • If they are named merkel trees after the guy who filed this patent, it seems to me they can't have been that obvious at the time - else they wouldn't be named after him. Or am I missing something here? – DonQuiKong Jun 2 '17 at 23:25
  • Low-hanging fruit of early(ish) computer science. So much so that perhaps it never even crossed anyone else's mind that it was patentable. – Xophmeister Jun 2 '17 at 23:29
  • The obviousness of this must be judged as of Sept 1979. – George White Jun 7 '17 at 0:02
1

For patents issued before June 5, 1995 and which expired after that date, patent terms were either 17 years from the grant date, or 20 years from the application date whichever was later. This patent should have expired no later than September 5th, 1999. This is a pretty early software related patent. One of the issues with software patents is that patent examiners generally focus on earlier patents for prior art. That far back it wasn't as common for software patents to be generated and the examiners themselves were not necessarily experts in computer science. The threshold for non-obviousness was sometimes too low. Thus, fairly routine programming techniques were sometimes granted because there was no previous example in the patent art. I'm not a computer scientist so I can't comment on how obvious the algorithm presented in the cited patent would have been in 1979 when the patent was filed. The fact that they are referred to as Merkle Trees and the inventor on this patent is Ralph Merkle suggests it may have been novel. There is even a Wikipedia article devoted to them. At least this patent covers using a mathematical algorithm to accomplish a specific task (creating and authenticating a digital signature) rather than patenting the abstract mathematics themselves.

  • 1
    Minor nit: although you're right that the law in 1982 was for a term of issue date+17 years, the law was changed effective June 5, 1995 to application date+20 years; and as part of that 1995 law, patents applications that were on file prior to the law's effective date were given a term until the later of 17 years from issue or 20 years from application. I remember well the surge of applications being filed in May and early June to get in before the deadline. So in this case, the later of those two dates is twenty years from the application date, i.e. September 5, 1999; not January 5. – codingatty Jun 6 '17 at 19:56
  • @codingatty Thanks, I was aware of the change but didn't do the check. I'll edit the answer when I get a chance. – Eric Shain Jun 6 '17 at 22:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.