Hi Steve:

Steve Loughran wrote:
> > The SOAP spec itself can't (and arguably shouldn't) handle this kind
> > of thing itself.
>
> Well, then -why didnt WS-A to have some guidelines for the problem.
>
> Given that the WS-RF 1.0 spec depends on two different drafts of WS-A,
> surely it should have been obvious that there were going to be legacy
> addresses out there.

I agree, and I would suggest that feedback/errata be offered to the WSA
working group. I haven't had time to scan the archives, but I'm sure
this came up during the course of the WG discussions (not that you'd
know it by reading the specs). It would be nice to see some mention of
how to deal with multiple versions in there.

> Life is so much simpler with URLs. You cannot post a request to three
> different URLs, and if you use the element in the message body to
> select your action, its inherently impossible to hae duplicate
> payloads.

I don't think this is a particularly valid comparison. It would be more
like adding multiple versioned HTTP headers rather than posting to
different URLs. The point of SOAP headers is that they modify, not
replace, the (potentially non-extensible) content of a message.

--Glen