dutcher@seas.gwu.edu (Sylvia Dutcher) (09/16/90)
In article <20293@orstcs.CS.ORST.EDU> holtt@nyssa.cs.orst.edu (Tim Holt) writes: >Patents for software ideas. It seems to come up very frequently >in this group. That's why I posted it here. > >Specifically: > 1) How is it done? I have been developing some computer. . . My husband, Phil Gagner, who is a lawyer as well as computer scientist, answers as follows: 1) Generally, it's not done. Computer programs or software ideas are only patent subject matter to the extent that they are part of the operation of some machine, like a rubber-curing machine; not a machine like a Turing machine. The theory from a very confused Supreme Court was that algorithms, being mathematical formulae, and therefore (?) reducible to truths about mathematics are therefore tautologies and therefore not patentable. 2) Some types of programs can be patentable subject matter. Read the very good analysis of this area in the recent (last month's) ACM journal. 3) See a patent lawyer who knows something about this area. Don't try to do it yourself--you won't succeed without more effort than it's worth. 4) Most patent lawyers know very little about this area, and don't under- stand what they think they know. Software patents are different, since half the trick is to disguise the fact that it's the algorithm that is being patented.