IP Law Daily Oracle prevails in Thought Inc.’s appeal of non-infringement finding in dispute over TopLink software
News
Monday, August 21, 2017

Oracle prevails in Thought Inc.’s appeal of non-infringement finding in dispute over TopLink software

By Mark Engstrom, J.D.

A federal district court did not err in granting summary judgment of non-infringement to Oracle Corp. on five claims of a Thought Inc. patent on a "System and Method for Accessing Data Stores as Objects," the U.S. Court of Appeals for the Federal Circuit has ruled. The appellate court rejected Thought’s only infringement theory on appeal—that the "find()" method of Oracle’s TopLink program met the claim limitation "extracting the object attributes and the object name from the object." The judgment of the district court was therefore affirmed (Thought, Inc. v. Oracle Corp., August 21, 2017, Moore, K.).

Thought sued Oracle for patent infringement. According to Thought, Oracle’s TopLink software for Java developers infringed five claims of U.S. Patent No. 5,857,197, which disclosed a system and method for object-oriented programs to access data in a relational database. Prior art techniques used customized code for each relational table, but the patent in suit used an abstraction layer with a set of interchangeable runtime adapters using the same application programming interface, thereby "effecting a consistent interface to the data store regardless of its underlying structure."

Oracle denied infringement and asserted a counterclaim of invalidity. The district court granted summary judgment of non-infringement, but dismissed the counterclaim without prejudice. On appeal, Thought argued that the find() method of Oracle’s TopLink program met the claim limitation "extracting the object attributes and the object name from the object" (the "extracting" limitation). Oracle argued that the find() method could not meet the extracting limitation because wholesale copying or passing a reference to an object was not part of the plain and ordinary meaning of "extracting the [data] from the object."

The Federal Circuit agreed with Oracle and the district court. The plain and ordinary meaning of "extracting … from the object" could not mean merely passing along or copying the entire object, including the container of the thing extracted. The full "extracting" clause at issue claimed a "first adapter extracting the object attributes and the object name from the object to effect packing of the object attributes and the object name as data." According to the appellate court, the plain language and context of that clause showed that the word "from" indicated the source from which the extracted thing was taken.

"Just as a child might ‘extract’ all of the cookies from the cookie jar and leave the cookie jar itself behind," the court explained, "the first adapter may extract all of the data contained within the object but must leave behind the data container of the object itself." In addition, the claim indicated that the extraction was performed to reorganize the object attributes and object name as "data" that was separate from the object. Finally, the use of the terms "extracting" and "obtaining" in other claims supported the conclusion that "extracting" did not include the container of the data.

Oracle also argued that the alleged software could not extract information from a single object, as required by the patent claims, because the find() method acquired the object name from one object (the entityClass) and the object acquired attributes from another object (the primaryKey). Thought disagreed, arguing that the set of parameters to the find() method was itself an object that contained references to two other objects, from which the object name and object attributes were extracted. Thought did not dispute that the claim language required the object name and object attributes to be extracted from a single object of a class; it argued that there was a single "set of parameters object" (passed to the find() method) that contained references to two other objects—one for the object name and the other for the object attributes.

The Federal Circuit agreed with Oracle and the district court that the accused method could not extract two pieces of data "from the object" when each piece of data was derived from a separate object. Significantly, Thought and its expert had admitted that, under Thought’s theory of infringement, the "object" from the object application was "an object with references to two other objects, not the two objects themselves." Moreover, the claim language required that object attributes and object names had to be extracted "from the object." Finally, the antecedent basis for "the object" indicated that the object "compris[ed] object attributes and an object name."

In the Federal Circuit’s view, a set of parameters containing nothing more than references to two separate objects, from which the object name and object attributes were separately extracted, could not meet the express claim language. Ultimately, the singular "object" required by the claim could not include the object name and object attributes if it was merely a set of references to separate objects that comprised the object name and object attributes.

The case is No. 2016-2369.

Attorneys: Mark Carlson (Hagens, Berman, Sobol, Shapiro LLP) and Kelly G. Hyndman (Sughrue Mion, PLLC) for Thought, Inc. Steven Moore (Kilpatrick, Townsend & Stockton, LLP) for Oracle Corp., Oracle International Corp., and Oracle America, Inc.

Companies: Thought, Inc.; Oracle Corp.; Oracle America, Inc.; Oracle International Corp.

MainStory: TopStory Patent TechnologyInternet FedCirNews

Back to Top

Interested in submitting an article?

Submit your information to us today!

Learn More