July 27, 2009

Ex parte Cherian – and the misguided software per se rejection

By Gregory A. Stobbs, Principal

How do you check compliance with a standard without making reference to the standard? The applicants in Ex parte Cherian purportedly figured that out, but were denied a patent by the BPAI because their system claim was deemed non-statutory software per se. Ex parte Cherian, (BPAI 2009).

In this case, Cherian and his co-inventor Damodaran-Kamal were doing conformance testing of XML messages. An XML message is basically a way of communicating information so that a computer system can figure out what data goes where. Pretend that you are the computer and see if you can figure out this XML message:

<Person age=”32” height=”172cm”>

<Name>Fred Jones</Name>

</Person>

If you, the computer, had to put this information into a spreadsheet, chances are you would do something like this:

Name

age

height

Fred Jones

32

172cm

Of course, computers are not yet as crafty as we humans. If the sending party happened to mess up one of the XML syntax brackets in the above message, we humans could probably still figure out the sending party’s intent, whereas a digital computer will fail miserably. That is why the W3C organization publishes detailed XML standards that dictate how messages must be formulated for computers to correctly interpret them.

Cherian and Damodaran-Kamal had devised a way to test XML messages for compliance to the XML standards, without using a reference implementation of the language standard. At first blush this seems rather impossible, but what Cherian and Damodaran-Kamal did was to test conformance by executing a set of verification rules expressed as protocol verification graphs. These protocol verification graphs, however, were not expressly stated in the exemplary claim on appeal. Instead their claim made reference to an enigmatic sequence verifier:

1. A conformance testing system comprising:

a sequence verifier to receive a plurality of messages exchanged between two or more participants and to test conformance of each of the plurality of messages to a specific markup language standard without using a reference implementation of the markup language standard;

wherein the plurality of messages comprises a first message sent from a first participant of the two or more participants to a second participant of the two or more participants and a second message sent from the second participant of the two or more participants to the first participant of the two or more participants.

Note that this is a system claim, not a method claim. Yet the examiner and the BPAI applied In re Bilski to reject the claim as non-statutory under §101. In re Bilski held that a process must either be tied to a particular machine or must transform a particular article into a different state or thing, if it is to be patent-eligible. The intent was to weed out patent-ineligible processes that preempt abstract ideas.

Here the BPAI applied Bilski to applicants’ system claim using its tried and true technique: construe the claims giving the “broadest reasonable interpretation,” and then apply In re Bilski if at least one such interpretation can be construed as a process. Here the “system” preamble and the enigmatic sequence verifier element were not enough to dissuade the Board from reading this as a process claim.

So, was this “process” claim attempting to preempt the abstract idea of using testing messages without reference to the standard? Read the claim again and imagine a human checker, wearing a paper hat labeled “Chief Sequence Verifier.” In such an embodiment, the above claim might indeed lie in the domain of abstract idea and off limits for patent eligibility.

But in this case, the BPAI was further constrained by the applications’ specification, which did not teach the paper-hat embodiment, but did teach that a programmed computer could be used. No significant details of the programmed computer were described. Thus the Board had to assess, under the In re Bilski test, whether this “sequence verifier” was a particular machine that would render the claim statutory.

Of course, the Board must draw the line somewhere. Claims that arguably distinguish over human-with-paper-hat thought processes merely because the specification mentions a programmed computer need to be weeded out if the current Bilski precedent is to be followed. But in this case the Board used rather unfortunate reasoning that is likely to lead patent examiners astray. Instead of applying the machine branch of the Bilski test by asking, “is the claimed invention tied to a particular machine?” the Board made the presence of computer software the focus:

“The first issue before us is whether claims 1 and 5-19 claim software per se.”

Starting from this software per se vantage point, the Board found that the examiner had been right in concluding that the claims “are nominally directed to a machine/apparatus yet they only claim software not embodied on a computer readable medium.”

This focus on software per se misapplies the law. Indeed, it has been the law since In re Alappat (1994) that where software converts a general purpose computer into a special purpose machine, such programmed computer will qualify as a particular machine. The Patent Office is now acting as if In re Alappat has been overruled, but that is not what the In re Bilski court said:

“As a result, issues specific to the machine implementation part of the test are not before us today. We leave to future cases the elaboration of the precise contours of machine implementation, as well as the answers to particular questions, such as whether or when recitation of a computer suffices to tie a process claim to a particular machine.” [majority opinion]

“… We aren’t told when, or if, software instructions implemented on a general purpose computer are deemed “tied” to a “particular machine,” for if Alappat’s guidance that software converts a general purpose computer into a special purpose machine remains applicable, there is no need for the present ruling. For the thousands of inventors who obtained patents under the court’s now-discarded criteria, their property rights are now vulnerable.” [dissent by Judge Newman]

Until the Supreme Court sets us straight on the proper metes and bounds of the patent-eligible process, the Patent Office should stick to the script of In re Bilski, In re Alappat and indeed, Diamond v Diehr. With hundreds of pending patent applications awaiting decision by the Supreme Court on this issue, it is inappropriate that the Patent Office should prefer such confusing variants as the software per se test or the software not embodied on a computer readable medium test.