W3C home > Mailing lists > Public > public-ws-addressing@w3.org > November 2004

RE: WS-Addr issues

From: Francisco Curbera <curbera@us.ibm.com>
Date: Sat, 6 Nov 2004 18:42:38 -0500
To: Doug Davis <dug@us.ibm.com>
Cc: "Jim Webber" <Jim.Webber@newcastle.ac.uk>, "Marc Hadley" <Marc.Hadley@Sun.COM>, "Mark Little" <mark.little@arjuna.com>, "Martin Gudgin" <mgudgin@microsoft.com>, public-ws-addressing@w3.org, public-ws-addressing-request@w3.org, "Savas Parastatidis" <Savas.Parastatidis@newcastle.ac.uk>
Message-ID: <OF9A8BBE45.9DE39BE0-ON85256F44.0081653A-85256F44.00823F4D@us.ibm.com>

Gudge is right. I don't know exactly what part of the spec would make one
think that reference properties/params are intended to carry arbitrary
headers. I guess the argument is that any mechanism can be extended beyond
its intended use but that doesn't really help. Reference properties and
parameters are part of the address of the endpoint in that every message
directed to the endpoint must include them or will not reach it. This is
not the case for action; for example note that the same endpoint accepts
multiple messages with different action values (the different messages a
port type accepts).


                      Davis/Raleigh/IBM@IBMUS         To:       "Martin Gudgin" <mgudgin@microsoft.com>                                        
                      Sent by:                        cc:       Francisco Curbera/Watson/IBM@IBMUS, "Jim Webber" <Jim.Webber@newcastle.ac.uk>, 
                      public-ws-addressing-req         "Marc Hadley" <Marc.Hadley@Sun.COM>, "Mark Little" <mark.little@arjuna.com>,            
                      uest@w3.org                      public-ws-addressing@w3.org, public-ws-addressing-request@w3.org, "Savas Parastatidis"  
                                                      Subject:  RE: WS-Addr issues                                                             
                      11/06/2004 07:40 AM                                                                                                      

Gudge wrote on 11/06/2004 04:07:10 AM:
> >
> > So - in addition to seeing off wsa:action I would also like to see
> > refprops/refparams removed. Certainly people will want to populate the
> > header space with particular header blocks, but bodging this
> > through an
> > addressing mechanism seems a poor factoring.
> RefProps/Params are NOT a generic way for specifying what SOAP headers a
> service requires. They *ARE* part of the address information.

I don't really see the difference.  If when they are used they become
SOAP headers how can they be view any other way?  I believe you said you
wanted to use the SOAP processing model to process them and that processing

model is (excluding the Body stuff) all about how/when to process headers.

Personally, I think reading the text for wsa:Action and interpretting it
to mean "dispatch data" and then going even further to meaning that it can
become "the payload" to the point where the Body is empty (ala WS-MDEX) is
reading far more into wsa:Action than reading "ref.props. are headers" from

the ref.props defintion.

In both cases the text is sufficiently vague to allow all sorts of readings

and I'm hoping the WG will, after the votes and hanging chads are counted,
make the text explicitly clear.

Received on Sunday, 7 November 2004 03:40:15 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:28:21 UTC