Re: Features as interfaces (was Re: Signaling MEPs)

David Hull writes:

> >I view MEPs as somewhat like Java interfaces. 
> >
> My understanding is that features in general are analogous to
> interfaces.  MEPs are features, of course.
> 
> Is that about right?

>From [1]:

"The SOAP extensibility model provides two mechanisms through which 
features can be expressed: the SOAP Processing Model and the SOAP Protocol 
Binding Framework (see 2. SOAP Processing Model and 4. SOAP Protocol 
Binding Framework). "

I suppose your statement is generally true viewed from that perspective. 
Note, however, that the above frames features pretty broadly, as 
encompassing function that can be realized entirely in SOAP headers 
(modules), entirely in the binding (e.g. MEPs), or in combination. 

MEPs happen to be features that are designed specifically to be core 
characteristics of bindings in particular.  So, while I think it's fair to 
say that MEPs act a lot like interfaces for the bindings, some other 
features may be independent of the binding.   I suppose that all features 
are in some sense interfaces available in one or another SOAP 
implementation.

So, I mostly agree with what you've suggested.  I'm not sure it changes 
the conclusion about MEPs and bindings, but I think it's true as far as it 
goes.  Thank you.

Noah

[1] http://www.w3.org/TR/soap12-part1/#soapfeature

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------

Received on Tuesday, 28 March 2006 21:25:26 UTC