Re: [ws-rx] Question about resolution of issue i090

[Matthew Lovett <MLOVETT@uk.ibm.com>]

Hi Anish,

I think that the sender needs *at least*
one EPR for protocol messages, and if we make this element optional then
there could be cases where we don't have one. That leads into a debate
about which addresses are valid here, in particular the WS-A None and Anon...
which I need to think about some more!

Cheers,

Matt

Anish Karmarkar <Anish.Karmarkar@oracle.com>
wrote on 22/03/2006 19:40:34:

> Matt,
> 
> That would fix it.
> 
> Do you think it makes sense to make it optional, given that it is
a 'may'?
> 
> -Anish
> --
> 
> Matthew Lovett wrote:
> > 
> > Hi Anish,
> > 
> > The EPR added by issue 90 gives a place where protocol messages
can be 
> > sent, but does not restrict the endpoints that can be used for
> > application messages. You might be right that we should have
been even 
> > looser - perhaps we should have said something more like:
> > 
> > This REQUIRED element, of type wsa:EndpointReferenceType as specified
by 
> > WS-Addressing [WSAddressing] specifies *an* endpoint reference
to which 
> > WS-RM protocol messages related to the offered Sequence *may*
be sent.
> > 
> > Cheers,
> > 
> > Matt
> > 
> > 
> > Anish Karmarkar <Anish.Karmarkar@oracle.com> wrote on 22/03/2006
16:51:31:
> > 
> >  > Paul Fremantle wrote:
> >  > > I'm not sure I agree that this restricts the
sequence to a single
> >  > > endpoint. It simply associates the endpoint with
the sequence (in the
> >  > > same way that there is a CS endpoint associated
with the initial 
> > sequence).
> >  > >
> >  >
> >  > I'm not sure I follow (very likely missing your point)
-- specifically I
> >  > don't know what you mean by '... CS endpoint associated
with the
> >  > *initial sequence*'.
> >  >
> >  > The resolution of i090 says this about the specified
endpoint:
> >  >
> >  > "This REQUIRED element, of type wsa:EndpointReferenceType
as specified
> >  > by WS-Addressing [WSAddressing] specifies the endpoint
reference to
> >  > which WS-RM protocol messages related to the offered
Sequence are to be
> >  > sent."
> >  >
> >  > In the non-offered sequence, RMD/RMS are fuzzy concepts
not necessarily
> >  > associated with a single EPR. For example, we talk
about sending
> >  > terminate message to the "RMD" not to a
particular EPR. Doesn't this
> >  > resolution change that for offered sequences? I.e.,
doesn't this
> >  > resolution restrict the RMD to the specified EPR (for
the offered 
> > sequence)?
> >  >
> >  > Thanks for the response.
> >  >
> >  > -Anish
> >  > --
> >  >
> >  >
> >  > > Paul
> >  > >
> >  > > Anish Karmarkar wrote:
> >  > >
> >  > >> I have a question on the resolution of issue
i090 as recorded at
> >  > >> http://www.oasis-open.org/apps/org/workgroup/ws-rx/download.
> >  > php/17304/MinutesWSRX-031606.html
> >  > >>
> >  > >>
> >  > >> The resolution restricts the offered sequence
to a single EPR. This
> >  > >> constraints/conflicts with the resolution
of an issue (can't recall
> >  > >> the issue number) that introduced that following
text at line 139-140
> >  > >> (in CD-03):
> >  > >>
> >  > >> "Note that this specification makes
no restriction on the scope of 
> > the
> >  > >> RM Source or RM Destination entities. For
example, either may span
> >  > >> multiple WSDL Ports or endpoints."
> >  > >>
> >  > >> The intent of that resolution was to allow
RMD and RMS to span
> >  > >> multiple WSDL endpoints, EPRs etc. Was this
conflict discussed during
> >  > >> the concall? At least in the minutes this
is not captured. If it was
> >  > >> discussed, can someone post the rationale
for reverting the previous
> >  > >> decision? Was there any discussion on making
> >  > >> /wsrm:CreateSequence/wsrm:Offer/wsrm:Endpoint
optional?
> >  > >>
> >  > >> Thx!
> >  > >>
> >  > >> -Anish
> >  > >> --
> >  > >
> >  > >