1) Under the current patent laws (US/EU/wherever) are new ICA algorithms the discovery of mathematical truths or the invention of new procedures to identify independent components of a mixed signal set?
2) Would patenting something in a saturated market be economical anyway?
3) If JADE or a FastICA variant were conceived/discovered today, could it be patented?
This is a more specific variation on : Are algorithm-based inventions patentable? - In which the answer is given is a general - Abstract algorithms can't be patented, but applications can be.
How abstract is too abstract?