Appeal 2007-2127
Reexamination Control No. 90/006,621
as that term is defined in the '604 patent and in the computer art, to require
that all program threads are interruptible at the end of their timeslice. There
is no support for Patent Owner's position that "preemptive multithreading"
only requires some program threads to be interruptible.
b. Ligler and Reiffin declarations
It is argued that the declaration by Patent Owner, Mr. Reiffin, and the
declarations by Dr. George T. Ligler establish that a person of ordinary skill
in the art would understand that "multithreading" does not require any thread
to be interrupted. We address each of the arguments separately.
It is argued that one skilled in the art would understand "[t]hat the
clock-activated interrupt embodiment of the original 1982 application is an
example of 'multithreading' as that term is defined at column 1, lines 24-38
of the '603 patent (Ligler, Tab A, ¶¶ 33-35)" (Br. 32).
Dr. Ligler's declaration of January 7, 2002, is based on a definition in
the '603 patent which states "with at least one thread invoked by a periodic
clock-activated interrupt service routine" (col. 1, lines 30-32), i.e., it only
requires one thread to be invoked by an interrupt, but does not preclude
more than one thread from being invoked by an interrupt. As experts do,
Dr. Ligler carefully confined his declaration narrowly to the definition of
"multithreading" in the '603 patent. He did not state whether he believed
that definition to be the ordinary definition in the art or whether the
disclosed embodiment would be considered "multithreading" as that term is
ordinarily defined in the art. In any case, however, the definition at issue is
the definition of "multithreading" in the '604 patent and in the art. Those
definitions require all (or at least a plurality of) threads to be interruptible.
55
Page: Previous 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 Next
Last modified: September 9, 2013