Re: Something else to consider

I think I like this too! Basically it says an attribute is a
different set of messages that are grouped together; one to
<get> the attribute and one to (optionally) <set> the attribute.
An <operation> OTOH is a set of messages that are sent and (optionally)
received.

Let's see what Steve thinks .. he's back from vacation but
is prolly buried in email.

Sanjiva.

----- Original Message -----
From: "Jim Webber" <jim.webber@arjuna.com>
To: "'Savas Parastatidis'" <Savas.Parastatidis@newcastle.ac.uk>;
<public-ws-desc-state@w3.org>
Sent: Monday, July 14, 2003 3:13 PM
Subject: RE: Something else to consider


>
> Savas:
>
> This seems a nice solution. But what about headers? In our (offline) talk
we
> covered some of these issues, but could you follow through in this group
> too?
>
> Jim
>
> > -----Original Message-----
> > From: public-ws-desc-state-request@w3.org
> > [mailto:public-ws-desc-state-request@w3.org] On Behalf Of
> > Savas Parastatidis
> > Sent: 13 July 2003 23:36
> > To: public-ws-desc-state@w3.org
> > Subject: Something else to consider
> >
> >
> >
> > All,
> >
> > Drawing from Sanjiva's very good suggestion in the WSDL
> > mailing list for the removal of <message>, please allow me to
> > suggest another syntax for attributes for all of us to consider:
> >
> > <attribute name="ncname">
> >     <get [(body="qname") | (element="qname")]>
> >         [<xsd:complexType> ... </xsd:complexType>]
> >     </get>
> >     <set [(body="qname") | (element="qname")]
> >         [<xsd:complexType> ... </xsd:complexType>]
> >     </set>
> > </attribute >
> >
> > It's very similar to Sanjiva's syntax for the operation.
> > Absence of a get or a set will make the attribute write-only
> > or read-only respectively.
> >
> > It is still the case that binding will be required.
> >
> >
> > Going back to David's comments on the requirement for having
> > "static" attributes in a WSDL document... I would argue that
> > the only reason one would want to put a static value in an
> > interface document is because they want to make available
> > metadata information. I believe that the current draft of
> > WSDL already supports this through the <feature> and
> > <property> elements, if I haven't misunderstood the intention
> > for their introduction.
> >
> > Regards,
> > .savas.
> >
> >
> >
>

Received on Monday, 14 July 2003 10:01:36 UTC