Friday, February 26, 2016

PTAB decisions involving computer vs. user distinctions

I reviewed quite a few 2015 ex parte appeal decisions in which the distinction made by the Applicant related to actions performed by a computer rather than a user. I'll discuss a few of these in today's post, with more to follow in a future post.

In Ex parte Furlong (Appl. No. 10/769,117, PTAB 2015), the Board found that "automatically" in a computer claim covered user input that led eventually to computer action. The Furlong application was directed to software for generating expense reports. The user tracked daily activities and tasks using the software, and the software associated these daily activities and tasks with expenses, then generated an expense report. One of the limitations at issue on appeal read: "in response to the identification of the activity item being expensable, automatically associating, by the processor, at least one expense item with the activity item." The Applicant argued that in the reference, the user does the association, not the computer:
Vance's new trip is not (and will not be) associated with a car rental until Vance's user associates the new trip with the car rental. Since Vance's act of associating is not automatic, Vance's act of associating fails to teach the claimed act of automatically associating at least one expense item with the activity item. 
The Board found that "automatically associating" was broader than the Applicant's reading:
No implementation for such automation is recited or narrowed, and no degree of automation is recited either. Thus, even were Vance to do no more than automatically record an association after a user matched an expense item with an activity item, this creation of an association by the computer following the user’s input would be within the scope of the claim.
Similarly, in Ex parte Ren (Appl. No. 12/582,142, PTAB 2015) the Board found that a particular claim limitation covered computer input originating from a user. The application was directed to a vehicle navigation system. One of the claims on appeal ("a method of displaying a navigation map") involved rendering a nonphotorealistic image based on three-dimensional data associated with buildings surrounding the vehicle. The limitation argued on appeal was "modifying a viewing angle of the nonphotorealistic image based on feedback from the user." The Applicant argued that in the reference, the viewing angle was modified based on feedback from the system/vehicle, rather than user feedback as claimed. The Board was not persuaded, and adopted the Examiner's finding that "because a user operates the system/vehicle, Herbst’s feedback is based on the user."

In Ex parte Blythe (Appl. No. 11/965,946, PTAB 2015), the Board found that the claim language did not exclude human action. The claims were directed to detecting unauthorized changes to ATMs by comparing the ATM's electromagnetic "fingerprint" to a stored reference fingerprint. The Applicant appealed a § 103 rejection and argued the limitation "receiving, with the detection device, an electromagnetic profile emitted from the financial terminal, said received electromagnetic profile including radio wave activity if present." The Examiner relied on Yuzik ("Surveillance of Suspects of ATM Fraud") for this, pointing to an "RF listening device" in the surveillance system that "listened for suspicious RF signals transmitted from the vicinity of the ATM to a nearby receiver operated by a suspect." The Applicant argued that a combination using Yuzik's RF listening device required a person to actually hear radio activity, which was "a different approach from that provided by the subject invention." In the appeal Decision, the Board noted that "nothing in the claim excludes a person from actually hearing radio activity for detecting as well."

Similarly, in Ex parte Schimpf (Appl. No. 12/647,965, PTAB 2015), the Board found that the claims did not exclude manual entry of database fields. The application was directed to a system that dynamically constructed a new database query, allowing the user to directly specify which field contents in the current record were important. The Applicant argued the reference required a user to "manually enter desired values into fields when generating a search query." In contrast, the Applicant argued that the claims required the search query to be populated from database record field values. The Board disagreed and said the Applicant's manual field entry vs. computer populated distinction wasn't present in the claim:
Those claims recite a database record contains field values (“fields in said currently active database record contain values retrieved from said database”) and a search query uses field values (“a search query requesting all records having values in said at least one user selected field”), but the claims do not recite, and therefore do not require, using database record field values as search query field values. Moreover, the claims do not restrict the source of search query field values or the manner by which search query field values are obtained.
Ex parte Granda (Appl. No. 12/651439, PTAB 2015) is an instance where the Board found that the claims did not require a particular method step to be performed by a computer. The application was directed to a resolving cell phone connectivity problems "while avoiding audible ringing and or inadvertent user pick-up during error diagnosis and correction." The claimed method on appeal involved: receiving an error message at an autonomous probe in response to a data call placed by the probe; mapping the error message to an array of corrective actions. The method also included "executing at least one corrective action from the array of corrective actions." In arguing this limitations on appeal of an obviousness rejection, the Applicant asserted that the feature relied on by the Examiner was carried out by a user:
Shabalin merely discloses directing users to articles containing instructions that a user may follow in order to resolve an application error and fails to describe executing at least one correction action as recited by Appellants’ claim 1.
In construing the claim, the Board found that the Applicant had read too narrowly:
[C]laim 1 does not require that the executing at least one corrective action be performed by any particular unit, but merely recites executing at least one corrective action from the array of corrective actions (see claim 1). Thus, We find that Appellants’ contention that Shabalin"s executing is carried out by the user, (i.e., suggesting that it must be performed by the autonomous probe) is not commensurate with the scope of the claims.
Finally, Ex parte Rys (Appl. No. 12/111,877, PTAB 2015) is an example where the Board found that the claim language allowed for intermediate steps performed by a user. The application described techniques for renaming multiple files. The limitation at issue on appeal in the "computer-implemented method" read: "on detecting that the name is changed, automatically placing a next file in the specified collection in the rename state; [and] displaying a computer-generated name for the next file, the computer generated name based on the changed name of the single file." The Examiner mapped this claim limitation to the "Rename and Copy function with a serialization option" described in Kaplan. Kaplan's software presented the user with a series of dialog boxes, where the user provides and initial destination filename. The filename was then automatically incremented as additional files are copied, but for each filename the user was required to take action to either accept or change the filename. One of the distinctions made on appeal by the Applicant was this user interaction with Kaplan's dialog box. The Board was not persuaded.
With regard to those additional features which Appellant argues are explicitly required, we note Appellant’s claim utilizes the transitional word “comprising. We therefore find that the possible inclusion of certain windows or dialog boxes within the process described by Kaplan does not preclude the Examiner’s reliance on Kaplan’s description of an automated serialized renaming process.
(Emphasis added.)


  1. Karen,

    Thank you for sharing these decisions for your readers. It seems that most of these applicants have fallen into the trap you've described before: appealing an issue based on a colorable claim construction instead of amending the claims to make that claim construction explicit.

    In fact, the only clear exception in my mind is Furlong. Sure, the applicant could have amended the claim to state that the processor determined whether the item was "expensable" and then performed the association in response. However, I'm not convinced the Board's interpretation was reasonable, as it completely read "automatically" out of the claim.

    My guess is that the Board used a strained BRI as a stand-in for their gut dislike of the claim. I think if they had wanted to use actual law, they could have expressed their dislike as a written description issue if the spec did not disclose how a processor could determine whether an item is expensable. If there were disclosure in the spec, then there might instead be a scope-of-enablement issue, since the claim encompassed any approach to automatically determining the expensability of the claim, and not the disclosed approach.

    What are your thoughts?

    1. Michael,

      Thanks for your input. Let me review Furlong and come back with my thoughts.

    2. >Board completely read "automatically"
      >out of the claim.

      I think the Board would disagree. Remember, according to Board the claim read on nothing more than "automatically record an association after user matched expense with activity."

      My case summary didn't give all the details, but in Furlong the Board went further to find that the reference taught "autofilling expense details from receipts in the report’s activity
      sections," which met the claim "automatically associating." Finally, the Board cited to Leapfrog to say automating manual processes is generally predictable and thus likely obvious. So the outcome wasn't predicated on the broad interpretation.