Appeal No. 95-4096 Application No. 08/188,660 If one were to substitute, for whatever reason, a packet- switched bus for the MBUS of Dashiell, Dashiell’s operation would appear to be inoperable. The data consistency sought by Dashiell would not be achieved by the use of a packet-switched bus since, as explained by Baxter, at column 5, lines 28-31, “a packet-switched bus is a bus whose bandwidth is allocated on a demand basis, as opposed to a circuit-switched bus, whose bandwidth is allocated for the duration of the connection.” Since the bus of Dashiell is not a packet-switched bus, extensive modification of Dashiell would be necessary in order to derive any operable system, such modification constituting invention itself, wherein data consistency is maintained while employing a packet-switched bus. This, of course, is appellants’ invention. Further, since Baxter is not concerned at all with the use of cache memories or the maintenance of data consistency, there seems to be no reason for the artisan to have been led to employ the packet-switched bus of Baxter in a shared memory multiprocessor system for maintaining data consistency. Where is the suggestion to so employ Baxter’s packet-switched bus? A mere reference, by Baxter [column 1, lines 16-17], to such a 8Page: Previous 1 2 3 4 5 6 7 8 9 10 11 NextLast modified: November 3, 2007