On Jun 16, 2004, at 2:05 PM, Andrew Newton wrote:
On Jun 16, 2004, at 1:19 PM, Greg Connor wrote:
Here is my take on this. During the jabber session, Jim Lyon
expressed that it is important to him that MARID be a part of a
larger context, and that "email policy document" is something that
the world needs anyway. This sounds "nice" BUT I don't accept this as
a "requirement".
And so this is the higher level question. Just how much extensibility
do we need?
We need to be able to extend the kinds of information conveyed in the
authentication record, without restriction; authentication is part of a
larger solution space which is currently poorly understood. Jim Lyons
provided some good examples.
We do not need to be able to change the result of the authentication
test.
From my point of view, backwards compatibility with SPF syntax with all
extensions in XML is a perfectly reasonable compromise: it provides
backwards compatibility for people who have already published and it
provides the level of extensibility we need without inventing a new
extensible language. Is it really unreasonable to expect the IESG to
understand this argument? Surely this kind of compromise is common in
engineering practical, deployable solutions.
Margaret.