- From: Jim Gettys <jg@pa.dec.com>
- Date: Fri, 5 Jun 1998 09:07:23 -0700
- To: Josh Cohen <joshco@microsoft.com>
- Cc: Benjamin "Snowhare" Franz <snowhare@xmission.xmission.com>, HTTP Working Group <http-wg@cuckoo.hpl.hp.com>, luotonen@netscape.com
>From a quick look at the patent, I expect that previous prior work is widespread; I think this may be an example of overeager patent applications vs. patent examiners that don't know enough computing history. But I haven't read the patent carefully, so take this comment with a block of salt. Examples: 1) The tunneling of the X Window System protocol through various compression servers, which know how to translate that protocol to a more compact representation and back again. This was done by a Stanford prof (name slips my mind, but I can regenerate it) years ago (late 80's is in my mind), and a product version appeared as the Serial Line X work that came out of NCD and the X Consortium well before the application date of the IBM patent. The patent claims of "Differencing, caching or protocol reduction techniques increase performance over the external communication link. " are certainly present in this system, both differencing and caching. 2) the tunneling of the IP protocol through chaosnet protocols, done widely at MIT in the 1983 time frame. Here TCP's behavior in the face of packet loss (particularly at the time) was mitigated somewhat by Chaos Net's more agressive retransmission strategy. 3) Both SLIP and PPP do differencing and caching on the TCP/IP headers to increase performance; these have been around for many years. The idea, as I understand it, has been around for a long time. Examples 1 and 3 above apply most closely, that occur to me without much thought, and long predate the patent application. - Jim Gettys
Received on Friday, 5 June 1998 09:09:11 UTC