Erratum: Strangeness in <RespondWith> behavior

["Eve L. Maler" <eve.maler@sun.com>]

In core 1.0 lines 971-973 (so far unchanged by our 1.1 decisions), it says:

"To specify extension types, the <RespondWith> element MUST contain 
exactly the extension element type as specified in the xsi:type 
attribute on the corresponding element."

There are two things about this that I find troublesome:

- There is a tiny bit of ambiguity in the sentence as it stands.  The 
phrase "element type", to XML DTD old-timers, means roughly an element 
declaration -- it's a model for element instances.  With the advent of 
XML Schema and its OO-inspired design, we now have real "types" to which 
element declarations are bound.  The xsi:type reference makes clear that 
what's meant is the type name, not the element name, but it threw me off.

- Given this, we have a seemingly inconsistent situation.  When the 
statement is a native SAML element, the content of <RespondWith> is a 
qualified element name.  But when the statement is a foreign extension 
element, the qualified type name has to be supplied instead.

I would like to suggest that we fix the almost-ambiguity in V1.1 by 
saying "element*'s* type" rather than "element type", and will treat 
this as an editorial correction.  (But Jahan may want to record it as an 
erratum anyway to keep track of it, and I can back it out if you don't 
like it.)

And perhaps, for V2.0, we should discuss whether this element-vs.-type 
inconsistency is a problem that we want to fix.

(Sigh.  There's really nothing much about <RespondWith> that I like...)

	Eve
-- 
Eve Maler                                        +1 781 442 3190
Sun Microsystems                            cell +1 781 354 9441
Web Technologies and Standards               eve.maler @ sun.com