W3C XML Protocol Working Group teleconference, 8 August 2001

1. Roll call

Present 32/25 Excused Regrets Absent

2. Agenda review and AOB


DF: Will postpone agenda item number 7 until next week due to
unavailability of list archives

AOB

Henrik: Perhaps we can have a regular report from the task forces at our
telcons?

3. Approval of minutes from Aug 1 telcon

Minutes approved as submitted.

4. Review action items


-- PaulC to send e-mail on issue 30
PC: can't send right now, will try to send as soon as the mail servers come
up.

-- GlenD & Noah to prepare draft text for the spec clarifying what the spec
does/does not mandate for mU/final recipient errors, and point to

types of mechanism that could generate such error.
DF: done, AI closed.

-- GlenD to propose a mustHappen mechanism (in loose terms, i.e. not as
tight as specification language)
GD: have drafted but not sent out due to listserve down.  For the record,
feels that WG should not go there.

-- Editors to make change 107
Done, remove this AI from list.

-- Eric Jenkins to propose a resolution of differences between "SOAP
application" and "SOAP node"
Resolution was posted although discussion postpone until next week due to
list servers down.  Mark action item as done.

-- Chris Ferris to draft spec language for handling application-defined
attributes vis a vis the infoset description
Done.  Will get to agenda item 7 next week.

-- Marc Hadley to send an email to Eric Jenkins regarding the disapearance
of the sentence referred to in issue 107
Done.

-- Editors to come up with a documentation structure for the different
pieces produced: set of documents, with a table of contents for each

document.
No work has been done yet.  Action Item still pending.
DF: Can we have something done by next week?
Henrik: perhaps Monday telcon to discuss this.
Marc Hadley, Nilo: agree to Monday telcon.
Jean-Jacques should be available for Monday telcon.
Editors agree to have something done by Tuesday.

-- Frank DeRose to send summary of the discussions on the proposals and
clarification (including the "kludge" proposal) to xml-dist-app.
Frank: haven't talked with Noah, yet.  Sent out summary of original
proposal to dist-app this morning, addresses both issue 78 and 16.
PC: do the W3c staff know whether mail being sent to dist-app is saved?
Yves:  Yes, it is being saved to server in France, it will be sent out when
the list serve comes back.
Frank: if my email falls through the cracks, I will repost it.
DF:  Yves sent out a mailing list through email that posters can use to
send messages to the WG.

-- Nilo Mitra to make a proposal about how to handle the examples in the
spec regarding the primer.
Done.
DF:  Requested that Nilo roll his proposal into whatever the Editors come
up with on Monday.
Nilo: Agreed

5. RPC task force issues

- issue 42, the task force recommends the use of "service" and "procedure"
instead of "object" and "method" respectively.

PC: suggests routine instead of procedure.  In SQL, the generic term is
routine.  Method and procedure are loaded words.
Jacek:  RPC has always been about procedures and thinks it would be useful
to retain this word to follow the RPC convention.
Henrik:  Is it used in other parts of the spec?
DF: Used in section 7.
Henrik:  Section 7 uses word "method" as well.

PC:  issue 42 wasn't scoped to just section 7.
DF:  on the other hand it references requirement 200, which addresses
conventions covered by RPC.
Jacek:  I think that proposal was aimed at section 7.

MH:  May no longer be relevant against version 1.2 spec.

DF: Proposes that we take back to originator and see if latest version
addresses the concern.

Frank:  Proposes that RPC task force review latest working draft for
terminology and look into a glossary definition.

Jacek: Will go back to the originator to see if latest version addresses
the concern.

- issue 45, proposal for new faultcodes in an RPC namespace.
Proposals:
new rpc namespace
clarification when existing server faults are returned
new faultcode: procedure not present
new faultcode: bad arguments
another faultcode called : rpc application fault

Jacek: The purpose of 4a and 4b was to handle older rpc implementations
that do not view application errors as rpc faults.  We have to let

applications indicate errors however they want.

Henrik:  concerned about the separation of the soap application from the
soap node.
Henrik:  would like to remove rpc application fault
Jacek:  this was added to allow languages that support exceptions to send
back exception information.

DF:  are you suggesting to modify the proposal such that the rpc
application fault (4a) would be removed?
Henrik:  yes this would be acceptable
Jacek:  do we allow other application-specific extensions to the faults?
Henrik: yes.

Jacek:  propose that remove 4a and call out the extensibility mechanism for
faults.

DF: Proposal to the WG that it should direct jacek and editors to come up
with modifications to the spec based on original proposal with 4a

removed and some additional positive text added.

WG: Agreed, no dissent rcorded.

6. Issue 100: mU reformulation, including handling of mU headers that

survive past their intended point of processing

Glen and Noah came up with a 3-part proposal
-reformulation of text
-overlap issue mentioned in preamble to proposal
-boolean attribute issue (which appears to be a schema issue which might be
answered by Gudge)

MH:  Proposal should have been based on the infoset rewrite.  Need a merge
of proposal with infoset rewrite.
GD: No problem.

MH:  Move one sentence from mustUnderstand proposal to general section on
attributes. [Scribe missed which sentence, upon asking for

clarification, WG indicated that the sentence did not need to be moved.]

GD:  Henrik requested that remove the last sentence from proposal and GD
does not have a problem with that.

MH:  Proposed rewording: "Absence of this attribute information item is the
same as giving it the value of false."

DF:  Will proposed changes change spirit of the text or just rewording?
MH:  Just a rewording.

DF:  If we agree, then we should feel comfortable about letting Glen and
Marc do the merge.
Henrik: agree as long as my amendments are included.
Chris:  agree as long as last paragraph in Glen's proposal should be
explicitly identified as non-normative.
Henrik: I also support this.
GD:  I'm fine with making this non-normative.

Henrik: my amendment is that the text should indicate that the attribute
has to be in the message and has to be in a specific place
MH:  this is addressed elsewhere in the spec.
Henrik: okay

Amendments to proposal:
merging with the infoset
remove the very last sentence of the last paragraph
provide some mechanism to make the last paragraph non-normative

DF:  can the WG accept this (modified) proposal?  Bear in mind, the spec
will be reviewed en-masse.

WG accepts this, no dissent recorded.

7. Handling application defined attributes in Infoset description, review

proposed wording

Postponed until next telcon.

8. Issue 109: Inclusion of HTTP Extension framwork in SOAP 1.2

This issue was highlighted on dist-app, and responses sought on the
following 3 options:
(i) removal of all references to HTTP Extension Framework binding
(ii) preservation of status quo
(iii) relocate the references to the HTTP Extension Framework binding to a
non-normative appendix or a separately published document
All but one of the responses to date have all been in favour of (i). One of
the votes for (i) is conditional on evidence of lack of use of the

framework, in which case the voter is in favour of (iii). The one other
vote was for (iii).
The WG should decide whether to accept option (i) or option (iii).

Henrik:  suggest that we move all references to HTTP extension framework
out of the normative part of the spec, and move them to a

non-normative appendix. In addition, we provide notice in the spec that we
will remove the references from the next version unless we hear

requests to retain them.

Hugo: suggestst we remove references from the spec with a note that, if
requested, we will bring them back.

A non-binding vote was taken for Henrik and Hugo's proposals:
9 votes cast for Hugo's.
4 votes for Henrik's.
No "cannot live with" votes for Hugo's proposal.

DF: Chair rules on these results, and directs the Editors to remove all
references from spec, add a "bring back on request" note, and add a

note to the change log (i.e. follow Hugo's suggestion).

9. Proposal for response to XML Schema WG on base64 encoding.


Postponed.

10. The XML Encryption WG has asked for a volunteer for an informal task

force to help clarify how XML Encryption and SOAP work together
The initial work would be to create some scenarios/examples. Decisions:
a) Does the XMLP WG agree to such an arrangement, and if so (b) who would
volunteer?

WG thinks that this is a good idea.  David Orchard volunteered for the task
force.

Henrik proposed that all communications of the task force should be made on
a public list.

Meeting adjourned.