From Peter Hohmann
Stephen Ornes raises the vexed issue of software patents (16 March, p 26). Software relies on mathematics, usually considered a discovery rather than an invention, and so is often unpatentable, stifling innovation.
The dilemma could be solved this way: mathematical truths pre-exist informally and are discovered by mathematicians. However, the particular way in which a mathematical truth is formally expressed is an invention, since there are many ways in which such things can be expressed. Hence a particular software implementation of an algorithm is patentable.
West Perth, Western Australia
