Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 alpha 5/22/85; site cbosgd.UUCP
Path: utzoo!watmath!clyde!cbosgd!mark
From: mark@cbosgd.UUCP (Mark Horton)
Newsgroups: net.mail,net.news
Subject: Re: Data compression to lower phone
Message-ID: <2203@cbosgd.UUCP>
Date: Thu, 5-Jun-86 15:03:57 EDT
Article-I.D.: cbosgd.2203
Posted: Thu Jun  5 15:03:57 1986
Date-Received: Sat, 7-Jun-86 05:32:35 EDT
References: <327@spdcc.UUCP> <8200002@nucsrl> <2369@phri.UUCP> <9875@ucsfcgl.ucsfcgl.UUCP>
Organization: AT&T Bell Laboratories, Columbus, Oh
Lines: 23
Xref: watmath net.mail:1657 net.news:4953

In article <9875@ucsfcgl.ucsfcgl.UUCP> arnold@ucsfcgl.UUCP (Ken Arnold%CGL) writes:
>Well, you could just compress the message body and leave the envelope
>(the part with the address) uncompressed.  What are the other problems?

We're missing two things here.

(1) The average netnews message has as much header as it does body, so
    there would be great savings in compressing the header too.  (This
    isn't hard to fix - you compress all the header lines that don't
    change.)

(2) Compress is a learning program - it works well on very large files
    by noting patterns that occur over and over and compressing them.
    For short messages, it won't have learned much by the end of the
    message.  Perhaps this could be solved by somehow getting compress
    to save context between messages, or having a standard context it
    uses to start, with common sequences.  But you'd have to make sure
    the sender and reciever were in sync at all times.

This sounds hard but do-able.  Anyone want to write it, preferably as
a subroutine so it can be embedded in news senders and readers?

	Mark