Rather than create a vague definition of "existence" why not make a
precise definition but make it optional. A verify MAY do an existence
test and if they do so, it will be done as follows...
I think an _optional_ existence check actually creates the very problem
you're talking about - "a domain advertising ADSP to have no clue as to
what verifiers will do".
Surely a technical spec that guarantees a random result is not a good
spec.
If the existence check is rendered optional then a random result is
exactly what we'll get. You'll have two different implementations of
the same spec, one doing the existence check, and one not, both able to
legitimately claim spec compliance, yet each producing a different
algorithmic result for the same input.
Arvel
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html