A SCOTUS case on the Docket this year relates to business method patents, which has caused a slew of amicus briefs from anti-software patent organizations. This article links to one of them.
I'll admit to not having given this question a thorough analysis but my inclination has always been that the anti-software patent crowd was too radical. There are obviously some serious problems with the patent system as it relates to the computer industry. Patents take too long to get and they last too long. Its too easy to get a patent on something that you aren't actually selling or even things you can't actually build, and so there is a big problem with patent trolls scooping up patents for futuristic concepts that haven't really been created yet and they certainly didn't create, later suing when someone else actually reduces them to practice. These are all policy issues that deserve attention.
On the other hand, patents play an important roll in the software startup space, where they serve as a proof point that a new company has a real technological innovation and investors have a real asset that they've purchased with their money. It is not fair to say that software patents only hinder innovation - they both hinder and help it - and furthermore I think a lot of that hinderance could be addressed if there was sufficient effort to do so.
A lot of time and advocacy that could go toward building a better patent system for software has been instead spent on eliminating software patents. To me, concluding that "software" as an embodiment of ideas is beyond patenting is to argue that patents should not exist at all in any domain, and I think that is the real goal of many of those who oppose software patents.
All engineering is the reduction of principals to practice. What difference does it make if that reduction occurs in a microprocessor or in a test tube? If we cannot see where the line is between science and engineering in the practice of computing it is because we haven't sought to find it. We deliberately mislabel most academic software engineering programs as "computer science" when they are doing nothing of the sort. If software is, indeed, truly special, it makes more sense to craft our policy in such a way that the concept of a patent works consistently across all domains of engineering, rather than to create an inconsistent policy environment by exempting software in order to uphold ancient distinctions that our advancing knowledge has blurred.
In any event, I would guess that the court will deliberately avoid engaging the question and narrowly limit their conclusion to business methods, which is the matter at hand. There will be a lot of hand wringing on Slashdot, and that will be the end of it.
Groklaw - SFLC files Bilski brief: Software should not be patentable and don't forget the 1st Amendment