As I've written too many times, software patents make no sense for lots of reasons. Although rather more circumspect than me in its phraseology, the Hargreaves Review essentially agreed:
In this case, the Review believes the balance of evidence lies in continuing to withhold patent recognition of non-technical computer programs as part of a sustained effort to deal with the growing and dangerous problem of thickets
But there still remains a grey area where pseudo-software patents are being granted because of legalistic trickery that succeeds in dressing up software as if it were something else - notably the “computer-implemented invention” (CII):
On Open Enterprise blog.