Congress never passed a law allowing software patents, though. It was the courts (but only the lower courts, never the Supreme Court) which historically interpreted the old patent laws (written before the age of computers) to apply to software -- I don't know of any parts of Title 35 which cover software specifically. Congress could modify the law to explicitly allow software patents, but they haven't.
1. Is it good policy to allow software to be patentable, and
2. Does existing law allow for software to be patentable?
I am fairly agnostic on question one, and about three Juris Doctorates short of being able to answer question two, but it's the Supreme Court's job to answer question two without any regard whatsoever to question one.
This article appears to be arguing about question two based on his opinions on question one, which is like arguing that a cat is a handbag because you want to be able to take it as carry-on luggage.
There is extensive literature (a subset of "legal realism" which accounts strategically for judicial action) which claims that judges do in fact incorporate policy considerations in decision-making. Secondly, it's widely acknowledged that the court frequently and with considerable influence advises Congress to change the law. In that context, it might be fine & useful for this article to speak in policy terms to the Court.
This isn't unusual. Constitutional concepts like "reasonable search," "cruel and unusual punishment" similarly invite courts to make policy judgments.