I disagree with Greg's characterization of the problem.
The key architectural advantage of XML is that you separate syntax
definition and data structure definition. This makes for a vastly better
result, it leads to much better consistency than you get with the ad hoc
approach that has been encouraged by yacc.
Regardless of the syntactic representation we should define SenderID in
terms of the XML Infoset. SPF has a slight amount of convoluted syntax, but
on the whole it reasonably clean - at least at present. The problem lies in
the fact that the only extension mechanism provided in SPF is a list of
attribute/value pairs. This allows a lot to be done but forces you into
convoluted dances when you have an extension that requires values that have
some structure.
Do not forget here that our objective here is to gain acceptance by the
Internet community at large. The endorsement of the IESG is significant but
so is the endorsement of the stakeholders.
What we definitely should not do is try to predict the disposition of the
IESG through clairvoyance as some are suggesting. If the IESG want to make a
point to the group then they should send us an email. I do not accept the
claim that we should change course here and make decisions on what have been
flagged as showstopper issues by key stakeholders on the basis of reading
tea leaves. If this is a showstopper issue for the IESG then it is Ted's
responsibility to tell us.
We have a completely principled architecture here. We are recognizing the
importance of both a clean architecture with clear separation of syntax and
data structures, we equally recognize the need to provide support for legacy
systems that use a less architecturally pure syntax.
If this proves indigestible to the IESG my guess is that the XML RFC will be
accepted as standard and the SPF RFC as informational. That suits me fine
since I don't think anyone in the Internet world takes the slightest piece
of notice of which status is applied. It will not change any product plans I
have and I don't imagine it will change the product or project plans of any
other party.