From kragen@dnaco.net Thu Aug 20 12:09:46 1998
Date: Thu, 20 Aug 1998 12:09:45 -0400 (EDT)
From: Kragen <kragen@dnaco.net>
Reply-To: Kragen <kragen@dnaco.net>
To: vix@vix.com
cc: systalk@ml.org, talkback@threepoint.com, srctran@world.std.com
Subject: "push" patent covers BIND
Message-ID: <Pine.SUN.3.96.980820120008.21889U-100000@picard.dnaco.net>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-Keywords:
X-UID: 1295
Status: O
X-Status: 

Patent 5,790,793, which was granted a couple of weeks ago, apparently
covers most "push" technologies.  It also appears that its first claim
covers BIND.  It's available at
<URL:http://www.patents.ibm.com/details?patent_number=5790793>. It
says:

What is claimed as new and desired to be secured by Letters Patent of the United States is: 
    1. A method of communicating between computers, comprising the steps of: 

    - creating a message at a first computer, said message including a
    reference to a predetermined location;
    - transmitting, by the first computer, said message to a second
    location; and 
    - receiving said message by a computer at the second location;
    - decoding said message by the computer at the second location by
    retrieving data from the predetermined location, automatically by a
    single application, without requiring user interaction, into the
    computer at the second location.

Doesn't this cover DNS lookups by referral?  When I've sent a request
with no recursion requested to the "wrong" DNS server, the DNS server
creates a message including a reference to a predetermined location
(the NS record), transmits that message to my nameserver, which
receives it, and then decodes it and retrieves data from the
predetermined location automatically and without requiring user
interaction.

Interestingly, the references include the MIME RFCs.

Several of the other claims don't appear so obviously invalid to me.
Claim 2 specifically covers the case where the message is HTML, claim 3
specifically covers the case where the message decodes to HTML, etc.
If one claim is bad, does that destroy the whole patent?

Kragen



