Re: request for explanatory wording for features

A feature is accepted if it's at the remote (server's) end of an 
operation. In my view, this would need to be indicated in an operation's 
<output> element. We don't have this level of granularity today; so we 
cannot know whether the feature applies to inbound or outbound messages; 
we must assume they apply to both. This sounds like a new issue.

Required is a little ambiguous; it could mean supported "at the client 
side", which we would need to indicate in an operation's <input> 
element, similarly to the above. Required in the current spec sense 
means tagged with the @required attribute: the feature MUST be supported 
by the service (or the client).

Jean-Jacques.

Jeffrey Schlimmer wrote:

> Glen, how would a WSDL processor know whether a feature was accepted or
> required?
> 
> --Jeff
> 
> 
>>-----Original Message-----
>>From: www-ws-desc-request@w3.org [mailto:www-ws-desc-request@w3.org]
> 
> On
> 
>>Behalf Of Glen Daniels
>>Sent: Sunday, September 21, 2003 8:08 AM
>>To: Sanjiva Weerawarana; www-ws-desc@w3.org
>>Subject: RE: request for explanatory wording for features
>>
>>
>>
>>(writing this on a plane - not sure if I'll be able to get it
>> sent before Sunday night...)
>>
>>Hi all:
>>
>>
>>>Feature proponents: Can someone please give some explanatory
>>>sentences that explains what a feature is? The current wording
>>>is, um, recursive:
>>>
>>>    "A Feature component describes a particular feature that
>>>     a Web service accepts or requires in particular interactions."
>>>
>>>While its cool to have a recursive definition, it doesn't help
>>>anyone understand what a feature is supposed to be. Maybe there's
>>>wording in the SOAP spec we can borrow (or refer to). Can someone
>>>take this on please? Glen?
>>
>>Sure.  How about this (paraphrased from the SOAP spec):
>>
>>	A feature component describes an abstract piece of
>>	functionality typically associated with the exchange of
>>	messages between communicating parties.  Although WSDL
>>	poses no constraints on the potential scope of such
>>	features, examples might include "reliability",
>>	"security", "correlation", and "routing".  The presence
>>	of a feature component in a WSDL description indicates
>>	that the feature is either accepted or required in
>>	particular interactions.
>>
>>This is a band-aid patch that clears up the particular wording you
>>noted had problems.  I do plan to take a swing at creating some
> 
> further
> 
>>explanatory text about features/properties in general, as discussed
>>a couple of F2F's ago.
>>
>>--Glen
>>
> 
> 
> 

Received on Monday, 22 September 2003 05:45:40 UTC