Re: Improving the handling of conversations in Internet mail

["Steven D. Majewski" <[e-mail]>]

On Dec 9, 16:01, Jacob Palme DSV wrote:
> 
> Do you mean that if "Followup-To: foo" where foo is an e-mail
> address would cause problems if it occurred in Netnews?
> What kind of problems would it cause? What would happen?
> 
> [ ... ]
> 
> In fact, I will test and see what happens. I hope I will
> not cause havoc by sending a test message.
> 

I'll volunteer ONE data point:

>@(#)$Id: trn.c,v 3.0 1991/11/22 00:14:59 davison Trn $
>Version: 3.4 
>Send bug reports to [e-mail]

Using Pnews to post a test article with my e-mail address in the
Followup-To: field, and using trn to respond with a F[ollowup]
command, I get the response:

>The original author has requested that messages be sent back via mail
>rather than posting to news.  Do you want to jump out of this and
>mail your reply instead? [yn] 

A response of "y" goes back to newsreader, i.e. you are expected to try
R[eply] instead of F[ollowup] command.

A negative response to this questions gives the following:

>The Cc: line should have the poster's requested return address.
>This program posts news to machines throughout the organization.
>Are you absolutely sure that you want to do this? [ny] 

And the reply does indeed have my address in the CC: field. 
It replies to both the newsgroup and my mail address.

So at least trn follows the "be liberal in what you accept" rule:
it doesn't do anything awful when it encounters an E-mail address
in that field. It asks for a clarification and trys to do something
sensible. I would not be too suprised to find that other newsreaders
try to "do the right thing" . If that IS tha case, then I don't think
it would be a bad thing to clarify the rules and state exactly WHAT
the right thing ought to be.  And in fact, a lot of internet standards
have been extended, limited or clarified by paragraphs in Assigned-
Numbers, Requirements-for-Hosts, and other RFC's, so I don't think 
the proposal is unreasonable ( *IF* it doesn't, in fact, "break" a 
lot of existing applications. ) 

My question about the proposal, though, is: how should it be
interpreted with respect to mail/news gateways ? Some programs
like trn are capable of responding by mail or news. Some programs
like Pine can read both mail and news ( locally, or via NNTP or IMAP ).
Some gateways are bi-directional, and some are one-directional. 
I'm not sure that overlaying the meaning of "Followup-To:" doesn't
create MORE ambiquity. ( I'm not sure that it *does*, either -- I just
haven't figured out all of the possible cases and interactions here. )


- Steve Majewski       (804-982-0831)      <[e-mail]>
- UVA Department of Molecular Physiology and Biological Physics