From: hughes@ah.com (Eric Hughes)
Date: Tue, 30 Nov 93 10:47:42 PST
To: cypherpunks@toad.com
Subject: Banning any subscriber
In-Reply-To: <199311280747.XAA15002@mail.netcom.com>
Message-ID: <9311301837.AA09077@ah.com>
MIME-Version: 1.0
Content-Type: text/plain


>> [...] none have ever
>> been implemented in software, except for killfiles, which are not
>> effective against disruption in an anonymous environment.

>Actually, I disagree. The Extropians list has an "::include" command
>that can be used to specifically include only certain thread or
>certain users (or any combination). I know for a fact that Dean
>Tribble and Paul Baclace are doing an "::exclude all" and then a
>selective "::include foo" to include certain threads and/or authors.

>I would call this a classic example of a positive reputation system.

It's a positive reputation system (+RS), albeit primitive, but the
reputation system (RS) as such is not in software but rather in the
minds of those who must explicitly include what they want to see.
What the extropians list sofware (ELS) is in this case, as software,
is an information system that can support a +RS, but not that system
itself.  The distinction is fine, and not always easy to see.

Now I was careful not to claim that RS's had never been implemented,
but rather never implemented in software.  The ELS is almost a +RS,
but not completely so.  A +RS must have a database of objects (people,
threads, topics, lists, etc.)  to be sure, and some sort of statement
about preferences about these objects, but database is not per se the
+RS.

The key that distinguishes an information system from a RS are the
rules of inference which connect the _preferences_ in the database to
_actions_ on the objects of the data.  The ELS does not contain
preferences at all but rather directly stores the actions on the
objects.  The connections between the preferences and the actions are
in the minds of the users of the ELS.

One can argue that the actions themselves represent the preferences,
but this is an argument to justify an existing design.  Ontologically
("what it is") preferences about objects and actions on objects are
different things; my attitude toward something is different than what
action I take toward it, although these may have been less
distinguishable when I was, say, fifteen.

Eric