Ted,
While I have not been following this issue particularly closely, this
appears to be a case where two experiments are using the same codepoint
to enode data with (allegedly) different meaning.
Specifically (quoting from Frank's message):
| Sender ID implementations SHOULD interpret the version prefix
"v=spf1"
| as equivalent to "spf2.0/mfrom,pra", provided no record starting with
| "spf2.0" exists.
This is known to cause havoc. There are literally hundreds of
articles pointing this out again and again since 2005-08.
I have not been able to find a concise description of exactly what havoc
will ensue, but if we take them on their word it seems inappropriate to
conduct an experiment where the same codepoint published in the DNS will
be interpreted differently by the competing experiments.
It seems far more appropriate to amend the Sender ID spec to say that
implementations of Sender ID SHOULD NOT interpret v=spf1 lines so that
both experiments can run in parallel without interference.
- Bill
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf