Sur son blog Samsara le professeur Peter Junger nous signale la disponibilité son opus de 130 pages : You Can't Patent Software; Patenting Software is Wrong. Il passe en revue les jugements de la cour suprême et des cours fédérales. Sa conclusion :

When I was in law school back in the mid-fifties of the last century, the received - if somewhat oversimplified - wisdom, even among those of us who wanted nothing to do with patent law, was that all patent claims that have been reviewed by the Supreme Court are invalid and that all patent claims that have not been reviewed by the Supreme Court are valid. I gather that nothing has really changed since then, except that the Court had, until recently, pretty much given up reviewing patent claims, And the result, of course, has been that all sorts of patent claims - including those that cover software - have been upheld by the lower courts - in particular by the Federal Circuit that has of late had nearly exclusive jurisdition over appeals in patent cases, even though the Supreme Court had earlier held in the cases known as Benson, Flook, and Diehr that software, including algorithms for the processing of information and mathimatical formulae, is, like other abstact ideas, not patentable. The Supreme Court, however, has of late evidenced a willingness once again to review patent cases and has even raised sua sponte the question of whether a patent was invalid because one cannot patent "laws of nature, natural phenomena, and abstract ideas." And it is also no longer true that the Federal Circuit can hear appeals in patent cases where the issue of the validity of the patent was not initially raised in the plaintiff's complaint: such appeals in federal cases are now heard by the numbered or D.C. Circuits. It seems improbable that we will have to wait very long before the Court has, once again, the opportunity to consider the validity of a patent on a computer program. It also seems improbable that when that day of judgment comes the Court will refuse to follow its own holdings in Benson, Flook, and Diehr and will instead treat those cases as having been overruled by the judges of the Federal Circuit. The Court has, after all, never taken kindly in the past to efforts by the lower courts to openly refuse to follow its precedents. On the other hand, it seems equally certain that those members of the patent bar and legal academics, who have for years been advising their clients and students in reliance on the opinions of the Federal Circuit that software is patentable, will raise a mighty protest that the Supreme Court was wrong when it decided Benson, Flook, and Diehr and that the Federal Circuit was right in acting as if it had overruled those cases. The trouble with that position is that, as I argue in this article at what most of you will consider excessive length, the Supreme Court was right in holding that computer programs are no more patentable than are mathematical inventions like the calculus or logical truths like De Morgan's law that "NOT (A AND B )" equals "NOT A OR NOT B ". Computer programs are texts, not machines as some lawyers have confused themselves into believing, and thus they may be copyrighted and protected by the First Amendment, but they are not patentable as machines. Computer programs are indeed processes, but they are not patentable processes because what they process is information and what they produce is information, not some modification of material goods or articles of commerce. The simple fact is - though the reasons for it may be hard for most lawyers to grasp - that, as the title of this article puts it: "You can't patent software: patenting software is wrong."

Ajout 20061107 1938 : j'en profite aussi pour signaler le témoignagne de Dan Bricklin (bio sur wikipedia) dans un procès au Texas. Lire aussi l'article du NYT sur le sujet.