RE: [VANILLA-LIST:3019] technical question

[Trent Piepho <xyzzy@speakeasy.org>]

On Wed, 22 Mar 2000, Chawla, Jay wrote:
> Hi everyone.
> 
> Thanks for all of your great responses!  I still have some questions based
> on what you've all said and some reading I've done, so I'll intersperse them
> between what you guys wrote, along with any comments I have regarding this
> patent
> or patents in general.  My new questions/comments are in ALL CAPS.

Do you think you could not use all caps?  It is normal email style to quote
others' messages with a > character at the beginning of each line, and then
write your response in legible text with proper capitalization.  You might
also want to try to wrap your lines at 80 columns wide, rather than making
some lines short, 
because it looks really bad when you do this and it is again hard to read.

> > If its plain text messaging, you're describing 'irc,' or Internet Relay
> > Chat, that has been around for quite some time.
> >  
> I READ THE IRC SPEC (RFC 1459) AND IT MAKES NO MENTION OF 
> ANY AGGREGATION -- THE SERVERS JUST FORWARD OUT MESSAGES 
> SEPARATELY TO EVERY GROUP MEMBER.  IT'S BEEN SUGGESTED
> TO ME THAT PLAIN OLD SOCKET OUTPUT BUFFERING DOES THE 
> AGGREGATION DESCRIBED IN THE PATENT, BUT I DON'T SEE IT.

What do you mean by, "I DON'T SEE IT"?  When small pieces of data are
transmitted by TCP, they are usually aggregated into larger packets, to make
more efficient use of the network.  See the "NAGLE Algorithm", which describes
the most common way for a TCP/IP networking stack to aggregate data into
larger packets.  You could easily show a network trace, which had a single TCP
packet with IRC messages from more than one group member in it.

> I THINK THE PATENT CLAIMS AGGREGATION OF PACKETS, AND
> FURTHERMORE THE AGGREGATION IS DONE BEFORE THE
> PACKETS ARE ENQUEUED TO GO TO ANY SPECIFIC GROUP
> MEMBERS -- THE AGGREGATION IS DONE FOR THE WHOLE 

I don't seem to remember this from the claims.  It appears that the payloads
must be aggregated before the message is transmitted, but that seems obvious,
as the server can't very aggregate a message that has already been transmitted. 
You can't add more items to a package once you've mailed it.  Does any claim
state that all members of the group must transmit a packet to the server
during the aggregation time interval, or can only *some*?  As well, must the
server send the aggregated data to *all* or only *some* members of the group? 
Must the data the server sends to each group member be identical, or can
different group members receive different data.

> > and sendmail server that sends out group messages.  An aggregated message
> > is a digest message consisting of list email collected over a period of
> > time.
> 
> MY UNDERSTANDING FROM READING RFCs 821(SMTP) 1429 (LISTSERV DISTRIBUTE
> PROTOCOL)
> AND A COUPLE OF OTHERS, IS THAT MAIL TO A GROUP IS SENT SEPARATELY TO EACH
> SERVER THAT HAS MEMBERS IN THE GROUP.  ALSO, I COULDN'T
> FIND MENTION OF ANY AGGREGATED MESSAGE OR DIGEST MESSAGE IN
> THOSE DOCUMENTS.  COULD YOU PLEASE POINT ME TO A SOURCE
> THAT DESCRIBES THIS OR TELL ME WHAT I MISSED?  

A "digest" is a basic concept of publishing, dating back who knows how many
hundreds of years.  In a digest, a number of pieces of information that were
published separately are aggregated together into a single publication.  For
example, a company might piece together AP items and other news stories about
a specific topic, and then publish them in a monthly trade digest.  Perhaps
you even subscribe to some of these?

This same concept applies to an email list.  In a normal mailing list, email
messages are sent to members of the group as they are received by the server. 
An individual subscribed to the list may get many emails.  Some people don't
want to receive a few emails every day from their list, and would rather all
the messages were aggregated into a single email, which they receive at the
end of a time period, for example one week.  It's a common thing for mailing
lists that carry announcements.  Rather than receive each individual
announcement, a subscriber receives a periodic digest with several
announcements in it.

Try looking a digest message from a mailing list:

http://newlug.org/mailArchive/gblug-discussion/9903.March_1999/15.html

> THANKS!  I WON'T REPEAT YOUR EXPLANATION HERE, BUT I THINK IT 
> MIGHT MAKE OUR CASE!  OF COURSE, I'M NOT GOING TO STOP LOOKING 
> FOR MORE STUFF. ;-)

It appear to me that the bulletin board systems people set up on their home
computers in the 70s, as well as the PLATO notes system created in 1973, would
fall under this patent.

Really, what you have here is a problem in information sharing that is as old
as human civilization.  You you have a plurality of people, many of which have
some information they want all the others to know about.  Each person can't
send his information to all the others simultaneously, he has to tell them one
at a time in person, give them each something written down, or send them a
unicast network packet.

The most naive and unorganized way for each person to share his information
with all the others, would be for each individual to simply go to every other
individual, one by one, and give them the information.  This is very
inefficient!  

Say we are a group of 100 people communicating by normal postal mail.  We want
to share our ages each other.  Each person in the group would send their age
to the 99 other people in in the group via 99 separate letters.  The total
number of letters sent is 990, as there are 100 people times 99 letters per
person.

There is a much better way to share information than this.  Instead, have each
member of the group send their information to a single person.  This person
collects and aggregates all the information he receives.  When he has received
the information of each person (as in claim 3), or gets tired of waiting
(claim 1), or a set time-limit expires (claim 2), he sends the aggregated
information to each person in the group.

In the 100 person mail example, we would have agreed upon a coordinator when
we established our group.  All 100 people would send their age to the
coordinator, who would aggregate the information, i.e. make a list of
everyone's age, and then send that list to every group member.  The total
number of letters sent is now only 200!  Much better than the 990 sent in the
unorganized way.

This same system is used all over the place.  Consider personal ads in the
newspaper.  A number of people send their ads to the newspaper, who collects
them over a time period.  When that period expires, the newspaper aggregates
the ads onto a few pages, and sends out newspaper to everyone.  Change
"computer" to "newspaper subscriber", "wide area communication network" to
"paper-boy network and the postal system", and "group messaging server" to
"offices", and that patent describes the newspaper personal ad system
perfectly.  I'll even translate claim 1:

A method for providing group messages to a plurality of [newspaper
subscribers] connected over [the paper-boy and postal systems], comprising the
steps of:

 * providing [an] [office] coupled to said [paper-boy and postal system], said
   [office] communicating with said plurality of [newspaper subscribers] using
   said unicast [paper-boy and postal system] and maintaining a list of message
   groups, each message group containing a list of at least one [newspaper
   subscriber].

 * sending, by plurality of [newspaper subscribers] belonging to a first
   message group, messages to said [office] via said unicast [postal system],
   said messages containing a payload portion (the ad) and a portion for
   identifying said first message group (the address).

 * aggregating, by said [office] in a time interval determined in accordance
   with a predefined criterion, said payload portions (the ads) of said
   messages to create an aggregate payload (the personal pages).

 * forming an aggregated message (the newspaper) using said a aggregated
   payload; and

 * transmitting, by said [office] via said unicast [paper-boy system], said
   aggregated message to a recipient [newspaper subscriber] belonging to said
   first message group.

When you translate the paper to a non-computer system, it's ridiculousness
because obvious.  It's only because patent examiners have no concept of
computers or networks, that they let things like this pass.  Patent examiners
should be held accountable for letting bogus patents pass by just because they
don't understand them.