Board affirms obviousness even though benefit described in reference not linked to claim element

Takeaway: The Examiner rejected claims to a method for context aware computing as obvious over a combination of three references. The Applicant attacked the Examiner's combination rationale, arguing that the benefit used by the Examiner as a reason to combine was not actually produced by the claimed feature. The Board did not discuss the benefit disclosed in the reference, but instead merely stated that the Examiner had found a motivation to combine in the purported benefit. (Ex parte Cutlip , PTAB 2013.)

Details:

Ex parte Cutlip
Appeal 2010007365; Appl. No. 11/847,298; Tech. Center 2100
Decided:  June 13, 2013

The application on appeal was directed to context aware computing. A representative claim on appeal read: 

     1. A method for ontologically driving context mediation in a computing system, the method comprising:
     collecting a plurality of events arising from a solution in a computing environment;
     loading operational meta-data for the solution;
     contextually mediating the collected events with the operational meta-data to produce context sensitive events; and,
     correlating the context sensitive events with corresponding symptoms in a display to an end user in the computing environment,
     wherein the events comprising both base events and also splunked events.

The Examiner rejected the claim as obvious over a combination of three references. The third reference, Chandrasekaran, was relied upon for the "base events and also splunked events" limitation. The Examiner cited the following statement as a reason for utilizing the base/splunked events feature in Chandrasekaran: "Based on the foregoing, techniques are clearly needed for efficiently processing sequences of events that may occur at different points in time." (Background of Chandrasekaran.)

In the Appeal Brief, the Applicant attacked the Examiner's reason to combine:

     The "technique" taught by Chandrasekaran is separate from the base events and compound events. Specifically, the techniques of Chandrasekaran are used for processing and evaluating base events and compound events. Therefore, to obtain the alleged benefit of "efficiently processing sequences of events that may occur at different points in time," classifying events as base events and compound events is not enough since the benefits are attributable to the technique and not just classifying events as base events and compound events to which the technique is applied. ...  Whether or not modifying the events of the applied prior art (i.e., the combination of Baudino and Zaner-Godsey) to include base events and compound events would result in the proposed benefit has not been established by the Examiner.
(Emphasis added.)

The Examiner did not address argument this in the Answer.

The Board affirmed the rejection. The Board briefly addressed the Applicant's rationale argument, but only to say that the Examiner was right, without explaining why:

     Appellant also contends that the combination of Baudino, Zaner-Godsey, and Chandrasekaran is improper. Reply Br. 6-7. In particular, the Appellant contends that “[w]hether or not modifying the events of the applied prior art (i.e., the combination of Baudino and Zaner-Godsey) to include base events and compound events would result in the proposed benefit has not been established by the Examiner.” App. Br. 18 (emphasis omitted). We disagree with the Appellant.
     The Examiner found that a person with ordinary skill in the art would have been motivated to combine Baudino, Zaner-Godsey, and Chandrasekaran in order to efficiently process sequence of events that may occur at different points in time. Ans. 8. As such, we find that the Examiner has provided an articulated reasoning with some rational underpinning to support the legal conclusion of obviousness.

My two cents: I agree that the Examiner articulated his reasoning. But I say that the Applicant showed that the reasoning was faulty, i.e., that it lacks a "rational underpinning." As a reason to combine, the Examiner proferred the benefit "efficient processing." But the reference which mentioned efficient processing did not disclose that this efficiency was a result of the feature relied upon in the combination.

The statement relied upon by the Examiner – "Based on the foregoing, techniques are clearly needed for efficiently processing sequences of events that may occur at different points in time" – was in the Background section of Chandrasekaran, following a list of shortcomings of the prior art. At most, this sentence shows that some combination of techniques disclosed in Chandrasekaran leads to efficient processing of sequences of events at different points in time. It does not show that efficient processing is produced by the feature which the Examiner took from Chandrasekaran – the use of base events and splunked/compound events.

In fact, other teachings in Chandrasekaran suggest that efficiency is a result of specific techniques such as the use of array operations to compare attribute values with constant values ([0100]) and splitting base event indices into parallel tracks based on a maximal composite partitioning key ([0114]).

So it's not the base and splunked/compound events that result in efficient processing. Then why would a POSITA be motivated to use these events to modify the primary and secondary references to produce Applicant's claims?

This rationale just doesn't stand up to scrutiny, which is why I'm convinced the Board got it wrong. And even if they didn't, I'm disappointed the Board didn't explain why the Examiner's rationale was rational, instead of just stating that the Examiner "found" that the reference taught a benefit. A "rational underpinning" requires more than merely putting "findings" on paper – it doesn't become rational until it's based on findings that are also established to be correct.

The Applicant clearly identified the issue – "whether or not modifying the events of the [asserted combination] to include base events and compound events would result in the proposed benefit." The Applicant also showed that Chandrasekaran did not link the benefit to the claimed feature. That should be enough. The Applicant shouldn't be required to spell out what other features in Chandrasekaran actually did produce that benefit (array operations, special keys, etc.).