Re: Notes - Treasure Hunt - Nov 7th 2006

I'll ask a different question.. Is it a "lie" : ?

<wsp:Policy>
    <wsp:ExactlyOnce>
        <!-- Alt1 -->
        <wsp:All>
             <iona:hightlyThrouputtable signedAndVerifiedBy=".."/>
        </wsp:All>
        <!-- Alt2 -->
        <wsp:All>
              <ws:Reliable/>
        </wsp:All> 
    </wsp:ExactlyOnce>
</wsp:Policy>

No, its' not. Alt1 is what a client1 got. Alt2 is what a client2 got. Imagine there's no two alternatives.
A client is asking a broker : find me a service with 'Alt1', another one is asking for a service matching alt2;
Is this a 'lie' ? : 

Client1 gets this. The service happens to be reliable, but we don't see it here, did provider 'cheated' :

<wsp:Policy>
    <wsp:ExactlyOnce>
        <!-- Alt1 -->
        <wsp:All>
             <iona:hightlyThrouputtable signedAndVerifiedBy=".."/>
        </wsp:All>
    </wsp:ExactlyOnce>
</wsp:Policy>

Client2 gets this. The service happens to be highly-throuputtable, but we don't see it here, did provider 'cheated' :
 
<wsp:Policy>
    <wsp:ExactlyOnce>
        <!-- Alt2 -->
        <wsp:All>
              <ws:Reliable/>
        </wsp:All> 
    </wsp:ExactlyOnce>
</wsp:Policy>

No I don't think so. 

The key is understanding what an alternative is IMHO. Alternative is a self-contained, unrelated to other alternatives, piece of vocabulary a client has to understand and do something about, be it a no-op, send a notification to a peer that it's found a highly-throuputtable service, marked a GUI box that a highly-throuputtable service has been found etc... 
Treating wsp:optional in a current way makes alternatives more then just pieces of vocabularies.

Thanks, Sergey

  ----- Original Message ----- 
  From: Sergey Beryozkin 
  To: Sergey Beryozkin ; Daniel Roth ; public-ws-policy@w3.org 
  Sent: Wednesday, November 08, 2006 2:56 PM
  Subject: Re: Notes - Treasure Hunt - Nov 7th 2006


  Sorry, forgot to add. The statement that a provider is "lying" by saying 
  <iona:hightlyThrouputtable wsp:optional="true" signedAndVerifiedBy=".."/>
  is totally wrong IMHO...
  The provider is saying that requesters may ignore this statement and this is what the Proposal1 was all about. Adding new attribute to differentiate between different shades of optionality will only complicate the issue.

  Thanks, Sergey
    ----- Original Message ----- 
    From: Sergey Beryozkin 
    To: Daniel Roth ; public-ws-policy@w3.org 
    Sent: Wednesday, November 08, 2006 2:52 PM
    Subject: Re: Notes - Treasure Hunt - Nov 7th 2006


    Hi

    I've read the text...

    It appears we'll have 

    wsp:optional="true"/false
    wsp:ignorable="true"/false

    Can someone please explain me what difference, from a requester's perspective these statements make :

    <MTOM wsp:optional="true">
    <MTOM wsp:ignorable="true">
    <MTOM wsp:ignorable="true" wsp:optional="false">
    <MTOM wsp:ignorable="false" wsp:optional="true">

    Thanks, Sergey


    ----- Original Message ----- 
      From: Daniel Roth 
      To: public-ws-policy@w3.org 
      Sent: Wednesday, November 08, 2006 2:04 PM
      Subject: Notes - Treasure Hunt - Nov 7th 2006


      Asir is having network problems, so I'm sending this on his behalf.

       

      Daniel Roth