Lists Home |
Date Index |
Thanks Rich. Why can't SOAP simply point to
its Infoset-based specification to normatively
denote what the allowable syntax productions
will be for a SOAP document instance/aka, message?
That is what they do, yes? Why do they need
a syntax-subset on top of that since they do
not validate what they bind?
I didn't intend to start an infoset vs syntax
war. Oh well... obviously that concerns folks,
but as specs go, there seems to be a disconnect
here between what SOAP specifies and the means
of it, and what it needs. If I understand you,
they want to cite a subset for binding instead
of simply denoting via the mapping of their
information items which parts of XML they
can accept usefully, and by that, tightly
couple back to their infoset implementation.
Ok within their application, but it makes
very little sense in terms of the overall
XML specification family to force that
tight coupling back out to the generalized
XML processor. If the "subsetting has begun",
then it has begun awkwardly and without much
thought with regards to the environment.
From: Rich Salz [mailto:firstname.lastname@example.org]
> If the SOAP specification is built on the infoset,
> is a syntax-subset necessary?
The SOAP processing model (i.e., what you do with a message, what its
semantics are, etc) is described in terms of the XML Infoset.
The only defined binding of a SOAP message is XML over HTTP. There are
other sample bindings, but they all do XML over other transports.
So yes, even the SOAP WG recognizes that XML is necessary for interop.