RE: operation name uniqueness draft available

Well, the operations that a service provides are defined by the port
types it supports. Those port types are determined (syntactically
speaking) by inspecting the bindings via the ports. So given port type
A, B, C where C extends A and B and bingings of the same name if you
have the following service:

<w:service name='Foo' >
  <w:port name='portForA' binding='A' ... />
</w:service>

then the service only provides A. Alternatively:

<w:service name='Bar' >
  <w:port name='portForC' binding='C' ... />
</w:service>

provides C ( and hence A and B ). It could also do:

<w:service name='Bar' >
  <w:port name='portForA' binding='A' ... />
  <w:port name='portForB' binding='B' ... />
  <w:port name='portForC' binding='C' ... />
/w:service>

And provide explcit ports for A and B

Hope this helps,

Gudge

> -----Original Message-----
> From: Steve Graham [mailto:sggraham@us.ibm.com] 
> Sent: 23 January 2003 18:24
> To: Martin Gudgin
> Cc: www-ws-desc@w3.org; www-ws-desc-request@w3.org
> Subject: RE: operation name uniqueness draft available
> 
> 
> 
> Ok, thanks for the explanation Gudge.
> 
> As a follow up question.  It is ok then, for a port element 
> to provide "the particulars of a specific end-point" at which 
> a *subset* of the service's operations are available.  In 
> other words, it is perfectly legal for only a subset of the 
> operations to be available at the end point address specified 
> by any of the service element's port children?
> 
> sgg
> ++++++++
> Steve Graham
> sggraham@us.ibm.com
> (919)254-0615 (T/L 444)
> Emerging Technologies
> ++++++++
> 
> 
> 
>                                                               
>                                                               
>            
>                       "Martin Gudgin"                         
>                                                               
>            
>                       <mgudgin@microsof        To:       
> Steve Graham/Raleigh/IBM@IBMUS                                
>                 
>                       t.com>                   cc:       
> <www-ws-desc@w3.org>                                          
>                 
>                       Sent by:                 Subject:  RE: 
> operation name uniqueness draft available                     
>             
>                       www-ws-desc-reque                       
>                                                               
>            
>                       st@w3.org                               
>                                                               
>            
>                                                               
>                                                               
>            
>                                                               
>                                                               
>            
>                       01/23/2003 01:03                        
>                                                               
>            
>                       PM                                      
>                                                               
>            
>                                                               
>                                                               
>            
>                                                               
>                                                               
>            
> 
> 
> 
> 
> 
> [inline]
> 
> > -----Original Message-----
> > From: Steve Graham [mailto:sggraham@us.ibm.com]
> > Sent: 22 January 2003 16:19
> > To: Martin Gudgin
> > Subject: Re: operation name uniqueness draft available
> >
> >
> >
> > ok more interpretation help.  If appropriate, you can 
> respond to the 
> > list, if you think others would be interested in this clarification.
> >
> > If I have a usual portType hierarchy:
> >
> > ptA
> >
> > ptB
> >
> > ptC extends ptA and ptB
> >
> > Question 1: Can I have a constellation of 3 bindings, 
> assume here they 
> > are all soap/http bindings.
> >
> > bindingA type=ptA
> > bindingB type=ptB
> > bindingC type=ptC
> >
> > or am I forced to have a single binding?  (and thereby 
> require a most 
> > derived portType for any set of porttypes that I want to associate 
> > with a service.
> 
> Multiple bindings is perfectly legal.
> 
> >
> > Question 2: Can I declare a service that declares ports for 
> different 
> > subsets of the service's portTypes?
> >
> > <service name="foo" implements "ptA ptB ptC">
> >   <port name="ptAport" binding="bindingA" > ... </>
> >   <port name="ptBport" binding="bindingB" > ... </>
> >   <port name="ptCport" binding="bindingC" > ... </> </service>
> >
> > or am I forced to have a single port that describes the soap/http 
> > endpoint to the entire set of operations on the service?
> 
> You can do what you have above ( although I would note that 
> the status quo does not have an implements attribute, look at 
> the mapping section 2.10.3 to see how the port types property 
> gets populated. )
> 
> Gudge
> 
> >
> > sgg
> >
> > ++++++++
> > Steve Graham
> > sggraham@us.ibm.com
> > (919)254-0615 (T/L 444)
> > Emerging Technologies
> > ++++++++
> >
> >
> >
> >
> >
> >
> >                       "Martin Gudgin"
> >
> >
> >                       <mgudgin@microsof        To:
> > <www-ws-desc@w3.org>
> >
> >                       t.com>                   cc:
> >
> >
> >                       Sent by:                 Subject:
> > operation name uniqueness draft available
> >
> >                       www-ws-desc-reque
> >
> >
> >                       st@w3.org
> >
> >
> >
> >
> >
> >
> >
> >
> >                       01/22/2003 03:09
> >
> >
> >                       PM
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > An initial draft that deals with the operation name 
> uniqueness issue 
> > we identified this week in AZ is at[1]
> >
> > Draft contains diff markup so it should be easy to spot the changes.
> >
> > Best practice note is not in the draft yet.
> >
> > Gudge
> >
> > [1] http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/wsdl12
> .xml?rev=1.37.2.1
> &content-type=text/xml
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
>

Received on Friday, 24 January 2003 07:10:08 UTC