The PTO may still be accepting them, but I imagine the first time one goes to court they will be overruled. The decision is very specific: to be patentable, the subject of the patent must either "[be] tied to a particular machine" or "transform[...] an article". A storage medium is not a machine itself, and is not as far as I can see tied to a particular machine; nore does it transform an article. I note that the article you link to cites a federal court decision that such items are patentable, which obviously means that until such decisions are overruled explicitly by a court of appeals or the supreme court the PTO should continue as if it were still valid, but I would think that such an appeal court would, in light of the recent decision, look again and see that such claims are invalid.

In re Bilski only deals with the patentability of certain process claims. Patents can be granted to "[w]hoever invents or discovers any new and useful process, machine, manufacture, or composition of matter." 35 U.S.C. Sec. 101. Beauregard claims do not really claim a process, they claim an article of manufacture. They really do not fit within In re Bilski at all.

The scope of In re Bilski is very important. Lower courts are limited in how far they can stretch the decision. Even the Court of Appeals for the Federal Circuit (the appeals court that decided In re Bilski) is limited in what they can do with this decision. Individual three-judge panels on U.S. appeals courts are bound by precedents established by any panel of the given appeals court. The Court of Appeals for the Federal Circuit has to hear a case en banc (all the judges at one time) to overturn its own precedents. Since the Federal Circuit has exclusive appellate jurisdiction over patent cases, they are the ones who would need to hear an appeal.

This means that any precedents that were in place before In re Bilski, and were not affected by the decision, are still binding on everyone except the Court of Appeals for the Federal Circuit sitting en banc or the Supreme Court of the United States. The Court of Appeals for the Federal Circuit noted that:

It is undisputed that Applicants' claims are not directed to a machine, manufacture, or composition of matter. Thus, the issue before us involves what the term "process" in 101 means, and how to determine whether a given claim...is a "new and useful process."

This suggests that the holding of In re Bilski is limited to process claims and that many other types of claims that are used in software patents are unaffected.

Bilski was about business method patents not tied to any machine. The Federal Circuit tried to make this clear in the In re Bilski opinion itself (page 21):

We further reject calls for categorical exclusions beyond those for fundamental principles already identified by the Supreme Court. We rejected just such an exclusion in State Street, noting that the so-called "business method exception" was unlawful and that business method claims (and indeed all process claims) are "subject to the same legal requirements for patentability as applied to any other process or method." 149 F.3d at 1375-76.[Fn23 Therefore, although invited to do so by several amici, we decline to adopt a broad exclusion over software or any other such category of subject matter beyond the exclusion of claims drawn to fundamental principles set forth by the Supreme Court.]

It is true that the validity of many broadly drafted claims may be at issue, but many software claims just do not make sense unless the claims are understood to be tied to computational devices. For example, Beauregard claims, which are claims on a computer readable media adapted to implement a method or system, are considered patentable by the PTO. These kind of claims are very popular because they allow patent holders to go after the software distributors rather than end-users.

It will be harder to enforce software patents, now that the defense lawyers can wield Sec. 101 with more power. But it is a mistake to declare victory against software patents based on a case where all the PTO wanted was for the patent applicant to add "computer implemented" to the claim language.

Someone correct me if I'm wrong, but I don't think the patent examiner's job is to check for prior art. Scanning the entire patent library for prior art is a non-trivial task which cannot be automated or done properly in a hurry.

The patent examiner's job is to check for prior art. They use keyword searches over multiple databases. They also identify related classes of invention and do a quick check (e.g., checking out the abstracts and drawings) of dozens, if not hundreds, of patents, to try to find relevant prior art. They actually do a decent job in most cases, which is why virtually all issues patents were initially rejected.

That being said, prior art can be in any language, any time before invention, and just about anywhere "public" (e.g., a German masters thesis in a school library, a brochure distributed at a trade show, or source code in a public CVS repository). No patent examiner is capable of doing an exhaustive prior art search, especially since patent examiners, like everyone else, have limited time.

It is possible for the patent prosecution process to be more exhaustive. Just hire more examiners and give them more time to search. Of course, when you hire more examiners, you need to pay more to attract talent. You also have to pay the additional workforce. Thus, a more in-depth patent prosecution process costs more, which means fees would need to be higher (discouraging small businesses and independent inventors from securing rights in their inventions) or the patent office would need to be subsidized by public funds (costing taxpayers money).

It's actually the responsibility of the person submitting the patent, and if they're wrong, they're vulnerable to expensive lawsuits from anyone who does hold prior art.

Not exactly (at least not in the United States). The person submitting the patent has to provide the patent office with a list of the prior art that the person is aware of. If the person is not aware of any prior art, then the list that gets sent is empty. If the person submitting the patent (or that person's representative patent practitioner) fails to disclose known prior art, or misrepresents submitted prior art, then the entire patent can be thrown out. This is actually a problem because it discourages patent practitioners from saying anything about why they believe the prior art they are submitting might be relevant, wasting the time of patent examiners.

It is actually a good idea to look for prior art when drafting the patent, even though there is no duty to conduct such a search. Drafting around prior art enables patent applicant to obtain broad scope without having to narrow their claims during patent prosecution (which, due to the Festo cases, can drastically narrow the claimed invention scope).

The third possibility, however, is that prior art exists but isn't patented -- meaning the patent examiner wouldn't be able to find it in their library anyway.

Patent examiners have access to many databases, but not all prior art can be found in their databases. Moreover, the patent databases are probably the only sources of prior art that are categorized at a fine enough granularity (400 classes further divided into 150,000 subclasses) to be useful to patent examiners trying to find prior art without relying on exact search terms (an issue given that terminology changes over time).

That means the patent submitter is not as likely to be charged with wrongdoing, but the patent will remain in force until someone takes the time to point it out.

If the patent submitter misses something, then the whole patent may not be lost if at least one of the claims is not anticipated by or obvious in light of the prior art. It is only if the patent submitter knew about the art, but withheld it from the patent office, that the whole patent would be lost due to wrongdoing.

Of course, with the quality of patents granted, the language and the overly broad claims means someone reading patents may very well conclude that there is no software they can possibly write without 'the possibility that they might be infringing a patent'.

Methinks that I should have phrased it as "often times, the possibility that you might be infringing a particular patent is pretty clear from a cursory reading of that patent." It does make a difference whether you are assessing the risk of infringing any patent versus the risk of infringing a particular patent. After all, if there are 1,000 patents that relate to your work, and if for each patent there is only a 0.07% chance that you infringe it, then there is a 50% chance that you infringe at least one of those patents.

The other issue is that, if you're creating a new product, and you know that its covered by a patent, you're willfully infringing on the patent, and are vulnerable to additional liabilities. If you can prove that you didn't know about the patent when you created the product, then you're not going to be hit as hard if you're found infringing.

That is indeed a core reason why engineers are often discouraged from looking at patents. Another reason is that they have a tendency to document their thoughts with emails that say things like "I think we infringe this patent," even when they have not properly analyzed the claims of the patent or its validity (you really need a qualified lawyer for that kind of work).

On the other hand, it is really stupid advice perpetuated by lawyers with little to no training in risk management. If reading a patent trebles the potential damages, but reduces the probability of infringement to less than one-third of the original probability of infringement, then reading the patent is better than ignoring it.

Often times, the possibility that you might infringe a patent is pretty clear from a cursory reading (which is all you can handle when wading through hundreds of possibly relevant patents). In many cases, you can at least tell whether a patent might be related to the work you are doing. Those that are too far out-of-field can be ignored while those that look suspiciously familiar may require referral to a qualified lawyer.

True, one of those out-of-field patents may come back to bite you because of broad claims. But, if they are not clearly related on their face and the most relevant patents were reviewed in more depth, then you have a pretty solid defense against claims of willful infringement. If, in the process of reviewing all those patents, you managed to engineer around or negotiate licenses for a few patents that appeared closely-related, you may have considerably decreased your risk of being sued. You just have to make sure no one sends around any emails that say "I think we infringe this patent."