Ex parte Gilmour
Appeal 2009-003143; Appl. No. 10/135,2541; Tech. Center 2100
Decided August 8, 2011
The technology on appeal related to profiling electronic documents. A representative claim on appeal read:
178. A machine-implemented method comprising:The Examiner rejected the above independent claim as being obvious over a combination of two references, Robertson and Barrett.
automatically constructing a profile of a first entity based on a plurality of sets of text, the plurality of sets of text including at least one set of text that was authored not for any purpose related to creating or accessing the profile;
storing the profile in a machine-accessible storage facility;
in response to an action of a second entity, identifying the first entity as meeting one or more criteria, based on the profile of the first entity; and
outputting an indication to the second entity that the first entity meets the one or more criteria.
On appeal, the Applicant argued that the combination did not teach the limitation emphasized above. The Applicant argued that Barrett disclosed prompts which were authored specifically for the purpose of creating a viewer profile, "which is directly contrary to the claim limitation in question." The Applicant then noted that the final Office Action did not explain how Barrett taught the limitation, but instead only made the naked assertion that:
Barrett teaches monitoring a user's viewing preferences to create a profile. Preferences, which are viewed by a user, are included in a profile. Preferences, which are not viewed by a user, are not included in a profile, are not purposed [sic] for creating a profile (abstract, col. 2, lines 17-67).
In the Answer, the Examiner first noted that the Applicant's specification did not "support" the limitation at issue, and "thus, Examiner is entitled to give claim limitation its broadest reasonable interpretation in light of the specification."
The Examiner's Answer then explained how Barrett taught the limitation at issue, as follows:
The above information shows that only a single set of programs is selected and the other sets of programs from the programs in the viewer profile template are not selected to create the profile. In this case, the other set is program names, which are not selected, are authored not for any purpose related to creating or accessing the profile. The other set program names are represented as at least one set of text that authored not for any purpose related to creating or accessing the profile.
In a Reply Brief, the Applicant first responded to the Examiner's comment that the limitation at issue was not "supported' by the specification. The Reply Brief also rebutted the Examiner's explanation of Barrett
Finally, the Applicant argued that the Examiner had improperly dissected the claim limitation:
Furthermore, to properly evaluate the invention as a whole, when applying the prior art, it is improper to dissect a single claim limitation into parts and then find disclosure of each part of the limitation in different references (or different places in the same reference), as the Examiner has done. Hence, with regard to claims 178 and 202, one cannot separate the concept of "a set of text [used to create a profile]" from the concept of "authored not for any purpose related to creating or accessing the profile". These concepts are inextricably linked to form a single limitation in claims 178 and 202, which cannot be dissected (i.e., a set of text that
is used to create a profile of an entity and authored not for any purpose related to creating or accessing the profile). To dissect this limitation into component parts when applying prior art vitiates its meaning.
Accordingly, the Examiner cannot simply use some disclosure in Barrett of the general notion of "not for any purpose related to creating or accessing a profile," where that notion is unrelated in Barrett to any set of text that is used to create a profile, and then rely upon Robertson's general disclosure of "sets of text" to provide the rest of the claim limitation. To find the present invention prima facie obvious, the Examiner must first find some disclosure of a set of text that is used to create a profile of an entity but which is authored not for any
purpose related to creating or accessing the profile.
The Board affirmed the obviousness rejection by finding that the limitation at issue was not to be given patentable weight. First, the Board noted that the limitation appeared to be intended use, as it "added no structure to the claim nor functionally changes the 'at least one set of text' that it modifies."
Furthermore, according to the Board, "authored not for any purpose related to creating or accessing a profile" was a mental step and as such, it was not entitled to patentable weight. Although the mental steps doctrine is typically used to analyze for compliance with 101, the Board found it also appropriate in assessing patentable weight.
We find that this purpose is an undecipherable thought, fully within the mind of the author as he or she composed the text. We strongly question whether one can discern the true purpose of any text, such as an e-mail: was an author’s purpose to inform or misinform the sender or recipient; was the purpose the author’s self aggrandizement or humble flattery of the sender or recipient; was the purpose to acknowledge some debt or to seek relief from it or, as recited in this claim, to be part of a profile. No objective reader of a set of text knows for sure what the author’s purpose in authoring the text was, let alone knowing specifically whether its purpose was related to creating or accessing a profile. The purported limitation expresses a mere idea, a mental step taken by an author and unknown by any but that author, if indeed he can discern his real purpose in authoring the set of text.
Because this was the only limitation argued by the Applicant, and because the Board also found the references to be properly combinable, the Board affirmed the obviousness rejection.
One ALJ dissented from the majority's position. According to the dissent:
[T]he language of claim 178 does not require that someone discern the true purpose of the text. Instead, I find that the negative limitation contained therein is directed towards discriminating whether the text is related to creating/accessing a profile or not.Applying the broadest reasonable interpretation in light of the specification, the dissent found that the limitation encompassed "intercepting text, such as e-mails, that is generally not inputted into a profile template."
While Appellants’ choice of terminology, e.g., “authored” and “purpose”, could possibly draw one to conclude that an intended use may be at the horizon when read in isolation, I find that when the claim is read in its entirety and read in light of the specification, it becomes apparent that Appellant is merely attempting to discriminate one type of data from another and not trying to create some mind-reading phenomenon.
The dissent also found that the limitation did affect the claimed functionality, and was therefore not intended use:
When looking at devices, it is understood that the structure of a data file does not change with the type of data placed therein. However, I am not convinced that the same analysis applies to the method steps of the present invention. ... The step of including a set of text with a particular characteristic affects the functionality depending on the type of text intercepted. Accordingly, I disagree with the Majority that the claimed “at least one set of text” should not be given any patentable weight.
My two cents: The Board's use of the mental steps doctrine to determine patentable weight seems out of left field. I don't get it. If an "objective reader" could not figure out the meaning of the limitation at issue, an indefiniteness rejection would have been more appropriate.
Nor do I see how you come up with intended use, because the use was actually claimed.
Then, with its talk about the limitation not providing any "functional changes" to the claimed text, the Board sounded like it was trying to invoke the Non-Functional Descriptive Material doctrine. But, as the dissent noted, NFDM is not applicable here, since the claimed data does affect the functionality. Namely, it affects the claimed construction of a profile.
Seems to me that the Board just didn't like this claim, but didn't know how to reject it, so it tried a shotgun approach: intended use, quasi-Non-Functional-Descriptive-Material, mental steps ...
I feel strongly that the claim was badly written from an enforcement point of view. I don't want to put intent in my claim because that gives the accused infringer plenty to argue about. Plus, what sort of proof do you use to show the author's "purpose" in creating a text?
Let's assume the limitation has patentable weight, and look at the Examiner's position. Best I can tell, the Examiner's logic is as follows:
- Viewer is presented with information about multiple television programs
- Viewer's choices determine which programs are included in a viewer profile
- Those programs which are included read on "authored for a purpose related to creating or accessing the profile"
- Those programs which are not included read on "authored NOT for any purpose related to creating or accessing the profile."
Any program not used in creating the profile is ... by definition ... "NOT for any purpose related to creating the profile". On the other hand, that doesn't take into account the claim term "authored." How does the term "authored" fit with the reference's TV programs?
The Applicant characterized the claim language as "a set of text used to create a profile" where the set includes a text NOT authored for any purpose ... " If the claim really said that, I think the Examiner is wrong. Because the unselected programs the Examiner is relying on to meet the "not authored" limitation are not included in (are not "used to create") the profile.
But the claim doesn't say that. Instead, the claim says "automatically constructing a profile of a first entity BASED ON a plurality of sets of text". And the reference does base its profile construction on the entire group of programs: those that make the cut and are selected for inclusion in the profile; and those that are not selected for inclusion in the profile. So isn't the profile constructed BASED ON the entire set of programs?
In total, I think the Examiner's position isn't off base in quite the way the Applicant argued, but does have some holes in it.