W3C XML Protocol teleconference, 30 May 2001

1. Roll call

Present 44/38 Excused Regrets Absent

2. Agenda review:

     Approval of minutes postponed to f2f

     Discussion on keeping issues in one list or having a separate list of
       that come from XML Schema members. Decided to take off-line and that
       should be easy to merge the lists and thus only manage one list.

4. Agenda item no. 4: Action items:

   1) still pending
   2) left as pending
      waiting for feedback
   3) complete
   4) left pending waiting for external response
   5) complete with no indication of any responses from the lists
   6) done
   7) done
   8) pending
   9) pending
   10) pending
   11) done

5. Agenda item no. 5: Spec clarification issues

   i23/4: SOAP forbids PIs and DTDs
     MarcH: SOAP processors should discard such messages
     wg agreed that it should fault
     MarcH: will send a proposal for clarification of the spec language

   i30: SOAP spec is unclear
     DavidF will ping PaulC to elaborate on the issue, we'll leave i30

   i97: Base64
     Henrik: two issues: (1) SOAP spec is clear that the line length does
   not apply
                         (2) XMLSchema spec allows some translation
       the issue might be this discrepancy
     DavidF: I read the original issues differently - that SOAP may not be
   clear --
          and so there may need to be further clarification of the spec
     DavidF: in addition, we need to coordinate with XML Schema
     MartinG volunteered to coordinate with XML Schema group

6. Agenda item 6: Issues

   i95 and i22:
     MarkN: SOAPAction identifies a message as SOAP
            it also identifies the "intent" of the message
            I don't see need for such identification, propose to deprecate
     A proposal for an attribute identifying the very target of the message
       The purpose for this is streaming
     DavidF: two issues: 1) use of SOAPAction in http
       2) need for targetting
     Henrik reiterated the reasons for SOAPAction
       Said that it would be wrong to incorporate the intent identification
       the request URI, which Jacek proposed before
     Henrik stressed that HTTP is an application protocol, not a simple
       transport protocol; SOAPAction is trying to bind the HTTP language
       SOAP language
     ChrisF: what use of SOAPAction is for a pure HTTP processor
     Henrik: it will be ignored
     ChrisF: SOAPAction doesn't make an ignorant target return a SOAP fault
     DavidF interrupted and proposed a f2f presentation:
       both sides of the issue, all of the viewpoints
     John Ibbotson: Microsoft proposed a SOAP header with the "action"
     MarkN volunteered to do the f2f presentation and a proposal

     Noah: there are number of related things
       SOAP has mustUnderstand, a question came about whether you can
   safely use
       mU to extend SOAP, showed a concrete Glen's counterexample (with an
       ordering mU header). SOAP might imply that you have to inspect all
   the mU
       headers for you before processing them.
       Proposal for steps in analysing, see pointer from agenda.
     Noah: we need to see all the mU problems _before_ processing anything
       this conflicts with streaming
     More discussion came and DavidF interrupted to say that the question
   to be
       decided on the call is whether or not to accept Noah's framework for
        answering the question
     DavidF proposes: let's go with these steps into the f2f; nobody

     Infosetting-the-spec, WG expressed interest and discussion will be put
   on the f2f agenda

7. Agenda item 7: Draft f2f agenda

     DavidF: we will start discussion of issues with the most advanced
       We might want to break up into smaller groups
       Next publication must be a technical specification
       Schedule might be set from the end to the present day
       We might have to make some hard decisions (cutting issues)
       Everybody should read the current SOAP/XMLP spec and generate
         clarification issues
       Presentations on issues 11 and 13 go to Mark Nottingham
         12 to Stuart Williams
         14 and 66 to Martin Gudgin

     No questions nor comments on draft f2f agenda
       end of telcon