Thursday, June 21, 2018

The history of software patents explained on Ars Technica




Timothy B. Lee has a very long (booklet-length) article in Ars Technica, “Why the Supreme Court’s Software Patent Ban Didn’t Last”.
  
This would also fit well on Electronic Frontier Foundation’s series on stupid patents.
  
The article stresses a couple of historical decisions, Flook in 1978, and Alice in 2014.  The Federal Circuit has generally been more friendly to questionable patents than the Supreme Court itself.
  
True, a mathematical theorem (Pythagorean, or the 4-color problem, or Liouville's) or formula (quadratic, like in Algebra 1) can’t be patented. Nor can a well established computer programming method today (I would wonder right off whether methods in an OOP java library can be patented, unclear) 

  
Somewhere, when solving a problem in the real world with real objects uses math, it can be. Jack Andraka’s patent app (2014), documented on Google (based originally on a science fair project that won a contest in 2013) depends on real world biology and various other objects (nanotubes), however “simple”.  But the purported “process” has to work (here, medically) and get approved by the FDA.

No comments: