Re: Support for application protocols

On Fri, Jan 24, 2003 at 11:33:05AM -0800, Jeffrey Schlimmer wrote:
> Mark, thanks for the additional explanation. FYI, the WG hasn't really
> begun to work on bindings with any significant earnest; when we do, your
> post below will undoubtedly be quite helpful. 

Thanks Jeffrey, but just to clarify, despite the impact of my proposal
on how application protocol bindings would be defined, the bulk of the
proposal targets the framework in which bindings are constructed (i.e.
the binding component).

If that's what you meant by "work on bindings", then nevermind. 8-)
Otherwise, I think it's something that the group would do well to
address long before it starts working on bindings.

BTW, upon further reading of the draft, I think some of the prose would
need changing too.  In particular;

"A binding component describes[sic] a concrete binding of a port type
component and associated operations to a particular concrete message
format and transmission protocol."

I gave some thought to how that could be modified to reflect the changes
in my proposal, but I couldn't come up with anything simple.  I think
this stems from the fact that an application protocol defines the
equivalent of a WSDL port type.

I hope you'd agree that this raises some significant architectural
issues for WSDL 1.2.


Mark Baker.   Ottawa, Ontario, CANADA.
Web architecture consulting, technical reports, evaluation & analysis

Received on Sunday, 26 January 2003 00:29:22 UTC