Decided September 8, 2008
(Appeal 2008-3860, Appl. No. 09/972,821, Tech. Center 2600)
This case relates to interactive television. The limitation at issue was:
removing the viewer discernable notification of the detected event, in response to detecting said priority is relatively low and a timeout for acknowledgement of said event has expired.The Board interpreted this claim as requiring that removal be conditional on detecting both conditions (priority and timeout). The Examiner used a §103 combination of three references, though only one of the references was used for teaching the removing action, the priority condition, and the timeout condition. (The other references were used for teaching other elements not at issue.)
The Board agreed with the Examiner's finding that the relied-upon reference taught the removal action and the two conditions. However, the Board found that the reference taught removal upon either of the two conditions, not both. The Board then found that:
The Examiner has not articulated a rationale to explain how Kikinis teaches or suggests that the viewer discernable [sic] notification of the detected event is removed in response to detecting both: (1) the priority is relatively low; and (2) a timeout for acknowledgement of the event has expired. Instead, the Examiner merely addresses how Kikinis teaches removal of the viewer discernable notification in response to detecting each of the two conditions separately (Ans. 5, 12). But the Examiner has not addressed how Kikinis teaches or suggests removal of the viewer discernable notification in response to detecting both of the two conditions.KSR requires "articulated reasoning with some rational underpinning to support the legal conclusion of obviousness," and because the Examiner did not provide such reasoning the Board reversed the §103 rejection.
(Decision, p. 7.)
What I noticed about this case is that the Board's grounds for reversal were no-reason-to-combine rather than does-not-teach. This highlights the danger of not addressing the combination in a §103 argument.
Note that the "combination" at issue for this particular element was not a combination of two references, but two features from the same reference. That might lead a practitioner to argue does-not-teach, or more particularly, does-not-teach-elements-as-arranged-in-the-claims. For an anticipation rejection, that would be enough: you can argue that the reference must teach the action based on both conditions, not simply action-on-condition-1 and same-action-on-condition-2.
Obviousness allows the Examiner to combine action-on-condition-1 with same-action-on-condition-2 to result in action-on-both-conditions — regardless of whether the conditions come from the same reference or different references. (Thus, I find it helpful to think of "combination" in terms of combining features rather than references.)
The catch is, though, is that the Examiner must provide a rationale for combining the features. When the features (here, the two conditions) come from different references, this point is so central to the obviousness analysis that it doesn't need stating. Yet I think it's easy for Examiners and practitioners both to overlook this point in a case like this one, when the two features come from the same reference.