Scribe: Marc Hadley
How do we move forward – some responses, not everyone has responded yet.
AOB: None.
Henrik – Modify fig 2.1 – done
Henrik – Service Model – not done. Not clear if work would go into paper.
Stuart to remove section 7 – done.
Marc and Mark to own bits of the doc – done.
Stuart sent out a mail outlining major sections of document and suggesting ways to move forward.
Intermediary processing model – Mark has put some effort in here.
Bindings – Marc has put some effort into this.
David – Talked yesterday with Stuart about AM doc. Relation with main spec, publication. Wanted to push Stuart to agree to schedule for how we bring AM doc to WG and then publication. Should be WD in it’s own right – not necessarily forever as separate document – decision later. Focus on publishing AM as separate WD. Timeframes discussed, settled most easily on presenting AM doc to entire WG on 28th. Spend time in Telcon to describe AM and highlight issues still outstanding. AM doc doesn’t have to be finished in this case. Have WG work on this in the following week and then send to W3C on 9th to be published on 16th. Dissolve AMG once WD has been submitted to entire WG. Individual authors of AM doc section would probably still be responsible for their sections. WG presented with 2 telcons per week. One mandatory, like the current WG one, one technical like this call, no attendence.
Stuart – Fair representation of what was discussed.
Marc – Question on timing of proposed second telcon. 4pm GMT works well ?
David – 8am PST is good time for me. Not so good for Asia.
Mark J – Section 4 (Applications and Modules) has been generating lots of discussion. Not possible to generate cohesive proposal in 1 week. Timing tight for sections not authored yet. 4.1 may be inconsistent with other section and no representation of current discussions. Can summarize discussion, but not synthesize proposal.
Stuart – styles quite different between sections
Mark J – Suggestions went back to SOAP, discussions suggests some WG members want to go in a different direction. Issues of mappings used to invoke behaviour on receiver side. Terminology perhaps too abstract.
Henrik – Would like to see this related to SOAP. Ask question “is there anything (requirements etc) that cannot be do in SOAP model”.
Mark J – People have widely differing views of how SOAP actor should be used in XMLP.
Henrik – All specs have different interpretations, but we have to get beyond the “it would be nice” way of thinking – back to SOAP model.
Stuart – Could represent section 4 as is, or could include summary of discussion. WG can continue work on this section.
Mark J – Can produce the discussion part.
David – This approach is fine. Mark J text has already been through one iteration.
Stuart – Module, Handler, Block discussion has unraveled somewhat in recent discussions.
Henrik – discussion has been about what you target with.
Stuart – That was another discussion.
Henrik – proposed term: hosted handlers instead of application.
Henrik – application is as meaningful as object – different for everyone.
Lots more discussion on meaning of application.
Stuart – still at tail end of process discussion, 15 min over already.
David – Sounds like a slippage of about a week over initial timescale.
Mark J – Yes.
David – Another option is to put that part in and designate it as a fluid section.
Mark J – Do you want to maintain AM doc as discussion go forward beyond initial working draft. Will it drive specification?
David – AM has continuing life, but mindful of boundaries on that. Not happy with AM doc driving the specification. Originally planned – take SOAP 1.1 – create abstract model from it. Need syntax (SOAP 1.1), behind that we can abstract out framework. That’s what I originally thought AM would be.
Stuart – don’t think we set out to go outside those boundaries.
Mark J – processing model belongs in AM doc.
David – Agree, concerned that we don’t go out to produce uber-protocol.
David – House in field analogy. AM is blueprint of house. If I want to knock down a wall, AM would tell me whether it was load bearing.
Stuart – Did we agree to slip the schedule by one week.
David – Mark J indicates that one week slippage would enable inclusion of current discussions.
Mark J – Is binding section OK.
Stuart – this is new from Marc Hadley.
David – can I just finish up on dates. Any discussion on future of AMG after publication of AM doc ?
All – discussion of second WG telcon. Particularly around timing.
David – main issue now is what needs to be done to AM doc prior to presenting it to WG.
Stuart – Intro and Glossary OK, Section 2 overview is OK. Any discussion.
Henrik – Application needs to be solved.
Stuart - Service Model, big disclaimer for Henrik’s concern. Has been trying to get proposals for ways to change or replace the section – nothing forthcoming so far.
Henrik – I am partially to blame for this.
Stuart - Invite discussion on what to do with section 3.
Mark J [to henrik] – Is there a sense that section 3 is flawed ?
Henrik – Introduces concepts that are not part of SOAP model. Goes too far in defining request response (3.1) and path model (3.4). SOAP doesn’t define them for a particular reason. There will be a variety of ways that we want to send SOAP messages around – don’t want to be overly proscriptive. Role of the protocol binding. Binding imposes rules on XMLP layer.
Mark J – we don’t have an alternative version to compare.
Henrik – that is the next step.
Start – re: path debate still ongoing on list, section can be adjusted once discussion has played out.
Further discussion on this that I didn’t have time to capture.
Stuart - principle concern is how do we get a document to present to WG. The doc would be liable to change afterwards as issues become clearer.
Mark J – Maybe we could add a new section that describes alternate view of section 3.
Henrik – Lots of people will read the WD – should be cautious with the content.
Staurt – options: leave intact, remove entirely, remove parts or an alternative that doesn’t yet exist.
Henrik – Look at 3.1 and 3.4. Not take them out but move to appendix to clearly mark them as under discussion.
Mark J – doesn’t rest of section rely on these parts.
Henrik – will require some massaging.
Stuart – any other views ?
Several people suggest leaving as is.
Henrik – even though this goes beyond SOAP 1.1 ? Discussion around SOAP 1.1 and the role of bindings. RPC is an abstract binding.
Lots of discussion around the role of bindings that I didn’t capture.
Stuart – still need to decide how to move forward. What is the view of the subgroup.
David – Catch glimpses of what Henrik is describing, but can’t quite get a handle on them. Would like to understand the approach you are suggesting. It’s up to Henrik to describe his approach. Current section is comprehensible, we need an alternative to compare.
Henrik – Happy to produce that. If we have request-response as a first-class primitive how do we support other protocols like multicast.
Stuart – Multicast is difficult with request response.
Henrik – that’s why I picked it.
Action: Stuart, take discussion to email.
Action: Stuart, mapping between AM and SOAP with attachments, will produce something by early next week.
David – Sounds like AM doc by 28th is too soon – 4th April is too late ! This would mean pub by end of April which slips whole schedule by 1 qtr. Can we say we will have AM doc by the weekend after 28th ?
Stuart – believe we can have something by end of month but there may still be some points of discussion.
David – fine.
Stuart – does anyone think we can’t have a draft by the end of the month?
No response – affirmation.
Stuart – any final questions or comments ?
Silence.