WIT

jeremy@ai.mit.edu (Jeremy Wertheimer)
Errors-To: listmaster@www0.cern.ch
Date: Tue, 14 Jun 1994 02:36:26 +0200
Errors-To: listmaster@www0.cern.ch
Message-id: <9406132300.AA11763@nature>
Errors-To: listmaster@www0.cern.ch
Reply-To: jeremy@ai.mit.edu
Originator: www-talk@info.cern.ch
Sender: www-talk@www0.cern.ch
Precedence: bulk
From: jeremy@ai.mit.edu (Jeremy Wertheimer)
To: Multiple recipients of list <www-talk@www0.cern.ch>
Subject: WIT
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
 > 	WIT looks really intersting.   But can we keep the 
 > development discussion in ordinary e-mail for a while?   Thanks! 

 > Rick Troth <troth@rice.edu>, Rice University, Information Systems 

Caught between the Scylla of not using the tools ourselves, and the
Charybdis of limiting the discussion to the users of the new tools,
let me risk the redundancy of cross-posting. 

I've appended below the text of a proposal that I recently posted on
WIT. Perhaps, for now, we could cross-post WIT proposals here, and
those interested could continue the discussions on WIT. I apologize in
advance if the subject of my proposal has already been discussed on
this list. In its relatively unstructured current form, I'm having
trouble keeping up with this list.

Jeremy

P.S. <A HREF="http://info.cern.ch/wit/hypertext/WWW/Topic1001/Proposal1012">
Here is a pointer to the proposal on WIT.</A>
-----------------------------------------------------------------------------
<ADDRESS>From: <A HREF="http://www.ai.mit.edu/people/jeremy/jeremy.html">Jeremy Wertheimer</A>, <A HREF="mailto:jeremy@ai.mit.edu">&lt;jeremy@ai.mit.edu&gt;</A></ADDRESS>
<ADDRESS>Date: Mon Jun 13 23:38:40 MET DST 1994</ADDRESS>
<ADDRESS>Topic: WIT - W3 Interactive Talk</ADDRESS>
<H1>We need fine-grained proposals (points).Messages can be sets of points.</H1>
<P>
I suggest that we need very fine-grained proposals (I'll call them
<I>points</I> for now), if we want to support the creation of 
tightly-structured discussions. A typical message would contain a
set of related points, or a set of related arguments to earlier
points. This would satisfy the twin goals of maintaining 
a tightly-structured discussion, and of allowing people to communicate
their ideas in reasonably large units (i.e., it won't force us to
limit our conversational moves to a few lines each).<p>

As evidence for the desirability of this structure, consider how we 
currently implement structured discussions in email and news. We yank
the original message to which we are responding, divide it up into 
small blocks, and attach our point-by-point responses to these blocks.
<p>

I suggest that we formalize this structure in WIT. At a minimum, all
that is required of the implementation is to allow a user to submit
a set of related proposals or arguments. The grouping of these points
into a message should be maintained in the database. This grouping is
orthogonal to the proposal/argument structure.
All that is required of the users is super-human discipline.<p>

P.S. Another facility that I would like to suggest is the
<I>raporteur</I> (summarizer). Although that will be the topic of
another proposal, I mention it here because, as an experiment, I'll
try to maintain 
<A HREF="http://www.ai.mit.edu/people/jeremy/wit/points-proposal.html">
a summary of the discussion of this proposal.</A>.