On Sep 28, 2004, at 11:01 PM, Dave Crocker wrote:
"message contents authentication" is quite precise and does not
mean bounces or bounce address. it means the message, itself.
I was using "bounce vs. bounce address" as examples of differences
between functional requirements. So, does "message content
authentication" mean headers or body or both or either?
what i do not understand is how that would improve the charter?
It doesn't. This is an attempt to improve the process.
Most/all of the candidate proposals have specific details about
the use of DNS and/or another query service. So I guess you are
suggesting decomposing things into major pieces of design choice.
And I think that raises an interesting question: is the work
best done as an effort to do a design from the ground up,
deciding on individual components that are then assembled
together, or should the working group take one (or more) specific
proposals and work to refine them (minimally)?
There is a great deal of experience that very strongly suggests
that the latter is the only way to achieve any sort of timely
working group output.
If we know what our functional system requirement(s) are to be and we
know many of the design choices that can be made, I'm not sure either
process is faster than the other.
That being said, explicitly stating that one of the input proposals
will be picked to be minimally refined works for me too. It's just
that discussion of all those "correct" design choices will happen in
the picking of the one proposal making consensus harder to judge. We
already saw this with the MASS BoF in SD.
-andy