From: Greg Broiles <gbroiles@netbox.com>
Date: Sun, 26 Jan 1997 08:55:52 -0800 (PST)
To: cypherpunks@toad.com
Subject: Toto's database
Message-ID: <199701261655.IAA18570@toad.com>
MIME-Version: 1.0
Content-Type: text/plain


Toto writes:

>My database seems to indicate a direct co-relation between the
>number of postings criticizing moderation and the number of
>UCE/Spam postings showing up on the list.

An interesting assertion. Is your database available for others to draw
their own conclusions from? I'd appreciate seeing at least a first-level
abstraction of the data, e.g., the number of messages posted in opposition
to moderation and the number of "spam" messages posted, listed by date. The
possibility that other factors caused the correlation you mention suggests
that other categorizations (e.g., total # of messages, # of messages which
are not spam nor related to moderation, and # of pro-moderation messages)
would also be useful, although I notice you do not mention those statistics.

Of course, correlation alone doesn't tell us much. If it really exists, it
suggests that the two phenomena observed are related in some way, but it
doesn't tell us the nature or the cause of the relationship. (All serial
killers drink water; but are all water-drinkers serial killers? Does
drinking water cause a person to become a serial killer?)

One explanation would be the link you suggested, that a moderation
proponent is sending (or causing to be sent) the spams as a way to make
moderation seem more attractive, or to detract attention from the
anti-moderation arguments. (If that really worked - e.g., if spam had the
effect of reducing the effectiveness of certain arguments, or of being
harmfully "diversionary" - wouldn't that suggest that moderation was, in
fact, necessary or at least useful? Hmm.)

Another explanation would be that a moderation opponent is sending the
spams as a way to punish the proponents of moderation.

It's further conceivable that you don't have a big enough data set to draw
meaningful conclusions from. The moderation decision, implementation, and
associated arguing have taken place within three weeks (John Gilmore's
original announcement was sent Sunday 1/5/97); it's hard to see how you're
going to be able to distinguish signal from noise in such a small data set.
I think this is especially true given the relatively inexact nature of the
"spam" weapon; as far as I can tell, apart from Dmitri's ASCII
art/cocksucker messages, the spams we've received have been from real live
spammers (not pissed-off listmembers) who harvested the
"cypherpunks@toad.com" address from messages sent to Usenet as spam bait.
Inherent in the "spam bait" attack is a lack of precise control over when
messages will be sent, or how many will be sent; so I'm curious about the
reliability of a correlation you've found between the actions of
distributed third parties (who are pursuing their own, cypherpunk-ignorant,
goals) and messages sent to the list.

Yet another explanation might identify external factors (like the recent
storms in the American midwest) which caused people to send more messages
independent of their ideology or content. My ISP, io.com, recorded much
heavier mail flow recently (to the point that it created problems for
delivery) during the storm; speculation is that people were stuck at home
for several days and turned to the Net to amuse themselves. Such an event
might cause otherwise independent-minded groups (spammers and moderation
opponents) to act in similar ways, like sending more messages than usual.
Of course, one would expect to see a rise in all message traffic, were this
the case.

Perhaps you'll be good enough to make your database available so that we
may each draw our own conclusions. I'm looking forward to hearing more
about how you derived this correlation. (If you have trouble finding a
server from which to make the data available, just say so; if nobody else
has a handy server, I'll add it to my web page.)

>Has anyone actually 'seen' John Gilmore, lately?

He was at C2Net's party last Friday night (1/24), being interviewed by some
folks from a Japanese TV program.


--
Greg Broiles                | US crypto export control policy in a nutshell:
gbroiles@netbox.com         | 
http://www.io.com/~gbroiles | Export jobs, not crypto.
                            |