[Ietf-not43] CRISP @ RIP43

[Hollenbeck, Scott]

> If I take February 2001 as the starting point (first draft of EPP),
> then, yes, SOAP was quite immature (and, IMHO, still is) but XML-RPC
> was already an old technology, with no change in the specification for
> three years and many interoprable implementations.

Not that it differs much from February 2001, but the first EPP drafts were
published on 10 November 2000 as individual submissions.  Please remember
that _flexibility_ was one of the important design requirements: flexibility
in transport (not just HTTP), flexibility in extension, etc.  My assessment
of XML-RPC was that it didn't meet those requirements.

BTW, the provreg requirements draft was just recently published as RFC 3375:

http://www.ietf.org/rfc/rfc3375.txt

> > (there was nothing being done with them
> > in the IETF),
> 
> It is not the same thing: SOAP will never be an IETF protocol, which
> has nothing to do with its stability or maturity.

That's another issue.  Given the IETF role in Internet protocol development,
the role of the IETF in ICANN, and the interest of ICANN in developing a
standard registry-registrar protocol, I felt it best to do the
standardization work in the IETF instead of the W3C, OASIS, or someplace
else.  As you said, SOAP would probably have been dead on arrival in the
IETF in 2000.

> > that wasn't really necessary.  Plus, EPP _can_ be moved 
> around using SOAP or
> > XML-RPC instead of being streamed over TCP.
> 
> EPP XML elements can probably be embedded in SOAP, but not in XML-RPC,
> which does not allow "unknown" (to the protocol) XML elements.

Note that as another XML-RPC deficiency with respect to flexibility... ;-)

-Scott-