From: [EMAIL PROTECTED] on behalf of Eric Rescorla
>More accurately, we want the charter not to foreclose the option
>of doing something better, on the grounds that it's incompatible.
I think that what we are actually arguing about is over language that indicates what level of 'better' is sufficient to allow a change that breaks compatibility with the legacy base and the type of incompatibility involved.
For example changing to using S/MIME would render the entire installed base incompatible and all the code used to support the implementations. I don't think any level of 'better' would be acceptable, it would be a back to the drawing board exercise.
On the other hand fixing the canonicalization scheme is almost certainly going to be required, that will render legacy verifiers inoperative but it will be relatively easy to upgrade the code base, it is an acceptable change at a much lower level of 'better'.
_______________________________________________ ietf-dkim mailing list http://dkim.org
