Message-Id: <199603271915.OAA23315@wilma.cs.utk.edu>
From: Keith Moore <moore@cs.utk.edu>
To: Dave Crocker <dcrocker@brandenburg.com>
cc: drums@cs.utk.edu, moore@cs.utk.edu
Subject: Re: Headers and agents 
In-reply-to: Your message of "Wed, 27 Mar 1996 07:28:57 PST."
             <v03005b04ad7f0ed24db9@[205.214.160.111]> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Wed, 27 Mar 1996 14:15:35 -0500
Sender: moore@cs.utk.edu

> At 8:54 PM 3/26/96, Keith Moore wrote:
> >Mailing lists are their own beasts, probably deserving their own section
> >in the abstract model document instead of trying to label them as gateways.
> 
> 	Keith, the enhanced architecture diagram that I showed in L.A.
> distinguished "personal" user agents from independent "service" user
> agents.  You indicated that you would like to defer such model enhancements
> for later.  It now sounds as if we might want to fold in some of them now.

Well, since we're talking about an abstract model, I'd like to keep the number 
of abstractions small and the definitions of the categories somewhat imprecise,
noting that these are just abstractions that don't necessarily correspond 
exactly to things in the real world.

In my mind, mailing lists are an important abstraction, and personal user 
agents and service agents are not so important.  I suppose "mailing lists" fit 
into the "service agents" category, but lots of people who would know what a 
"mailing list" is wouldn't have any idea what a "service agent" is.

I suppose I think "mailing lists" are their own beasts, because unlike other 
"service agents", mailing lists tend to accept messages from the Internet 
community (as opposed to an enclave) and make those messages available to 
subscribers from the Internet community (again, as opposed to some enclave).
Thus they have *two* interfaces that are widely visible, and there are some 
constraints on mailing lists that don't exist for "service agents" in general.

People can/do whatever they like within enclaves (and this includes 
mailing-list-like-thingys within enclaves), but when they start building 
widely visible interfaces and/or using standard software packages, it's 
helpful if the standards documents give rules and/or advice to keep them 
working well.


Keith