Message-Id: <v0214041bacf11c653952@[165.227.40.23]>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Sun, 10 Dec 1995 17:21:40 -0800
To: drums@cs.utk.edu
From: paulh@imc.org (Paul Hoffman)
Subject: Some possible decision choices

Post-Dallas greetings. As we all ran out of the room after the session on
Friday, I noticed that Keith left up his slide that was meant as a
guideline for what we do once we have identified a problem. I had stared at
that slide for much of the meeting, and decided that it might be a good
frame of reference for our future work. I snagged the slide, and I hope
Keith doesn't mind if I put it here for everyone who wasn't at the meeting.

As you can see, we have lots of things to think about!

Some possible decision choices:

- no change from 821/822/1123 etc.
- clarify the intent of the old language, add caveats, examples
- remove a feature/header/command (# address literals)
- deprecate a feature/header/command (actively discourage its use)
- recommend a particular interpretation over others
- chose one interpretation or behavior, and deprecate others
- remove old header or command, then provide old funtionality with
   new (better-defined) commands
- recommend new behavior (numeric timezones, updating in-reply-to
   in replies)
- impose new constraints
- express a choice as a tradeoff, and state pros and cons