Appeal 2007-2127
Reexamination Control No. 90/006,621
It is argued that one skilled in the art would understand "[t]hat the
clock-activated embodiment of the 1982 application teaches '[f]or most
applications, clock interrupts at intervals of about every 10 to 30
milliseconds should be frequent enough to keep up with keys stroked at the
keyboard' (A0064) (Ligler, Tab A, ¶ 25)" (Br. 32) and "[t]hat the editor will
usually, if the time period between clock-activated interrupts is 10 to 30
milliseconds, have completed its processing before the next clock
interrupt (although the editor should indeed be interrupted if that
processing happens not to be complete) (Ligler, Tab A, ¶ 26)" (Br. 32).
We assume these statements are correct. Importantly, as stated in
Dr. Ligler's second quotation, if the editor does not finish it must be capable
of being interrupted.
It is argued the one skilled in the art would understand that: "every
timesliced multithreaded system must provide threads with the ability to
voluntarily relinquish control of the CPU before the end of the timeslice;
otherwise a thread that requires only one millisecond to complete its
immediate task would idle and waste CPU time for the remaining 29
milliseconds of each timeslice (Reiffin, ¶ 34)" (Br. 32-33); "allowing such
tasks to idle would make the system several orders of magnitude slower
when processing brief tasks such as inserting single characters into an editor
buffer; such a slow inefficient computer would be worthless (Reiffin, ¶ 34)"
(Br. 33); and, "as provided in Microsoft-published treatises, the solution to
this idling problem is the voluntary release of the unused remainder of a
timeslice (see, e.g., RAY DUNCAN, ADVANCED OS/2 PROGRAMMING
237, Microsoft Press (1989) (providing one alternative for an idle thread as
'simply to give up the remainder of its timeslice') (Ex. 6); C. HUGHES, T.
56
Page: Previous 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 Next
Last modified: September 9, 2013