At 2:46 PM -0800 3/15/06, Michael Bernstein wrote:
The 'what should the validator be testing' discussion (at least the part
I'm participating in) is concerned with figuring out how to
deterministically tell whether an implementation is conformant without
giving false negatives for more elaborate implementations, and seems to
be shaking out ambiguities in the spec.
Is that not appropriate for this list?
The latter does not rely on the former. That is, figuring out how to
tell if an implementation is conformant is certainly relevant here.
Whether that should not be put into a particular validator, and how,
is not.
--Paul Hoffman, Director
--Internet Mail Consortium
