PMC Minutes for 2010-02-03

Attendees

Rob Godfrey: JPMorgan
Laurie Bryson : JPMorgan
Pranta Das : Cisco
Angus Telfer : Inetco
Shaun McAllister : Solace
Colin McNaughton : Progress
Hiram Chirino : Progress
Dave Ingham : Microsoft
Gordon Sim : RedHat
Rafael Schloming : RedHat
Carl Trieloff : RedHat
Chuck Linton : Tervela
Alexis Richardson : Rabbit
Alessandro Petroni : Bank of America
John O'Hara : Bank of America
Andreas Moravec : DBS
Hanno Klein : DBS
Matthew Arrott : TWIST
Keith Owen : Credit Suisse
Roshan D'Mello : Credit Suisse
Paul Blanz : Goldman Sachs

Apologies

Allan Beck: JPMorgan
Pieter Hintjens: iMatix
Mark Blair : Credit Suisse
Robert Greig : Barclays

Agenda

• Participation update
• F2F
• 1-0 Recommendation Process
• AOB

Participation update

Expect to have progress on new joiner this week

F2F

2nd and 3rd weeks of May are available (Wed/Thu/Fri)
AR : can wee do end of May
ACTION : Need to get people's availability for May
MA: last week may be difficult due to memorial day weekend

1-0 Recommendation Process

What is needed to be comfortable on 1-0 spec

CT: We will have additional engineers work through it
believe we can do that in current timeline
no specific issues we are concerned about
Aspects have been brought up like JMS, DTX, etc
We can do these as additional amendments, drawing a line under what we have
JOH: how long would you need to review
CT: End of February seems reasonable currently
Need to define mechanism for when a new PR would be required
JOH: Would like authors to have confidence and a quorom of the technical companies
having confidence.
Queues, subscriptions, JMS - need roadmap on those. Happy with drawing a line
on the current work
Would be happy to vote at end of Februaru
CMcN: Like to see JMS mapping
Would be good to have a prototype
CT: We have been prototyping to make sure key things are there.
CMcN: We want authors to have confidence
MA: Number of us not wanting to make the prototype a precondition of the PMC vote
Each organisations repsonsibility to get confidence on their own
TWIST wants before real implementations take place we would like an opt-in
prototype effort.
JOH: Prototype SIG?
MA: Prototyping a way for non-author orgs to become comfortable with the spec prior
to GA. Organisations would be preparing the internal resources for using/
implementing AMQP. Need discussions on who has resources / intent around this.
JOH: What would be your position at end of February?
MA: We will require conformance tests by before can get interoping
implementations. I think the prototype fills that space.
JOH: think we could do interop with connecathon
MA: TWIST is comfortable with end of Feb for Recommendation vote
Need a clear roadmap for GA for 2010, won't be comfortable with that yet
JOH: Should discuss certification next week
CMcN: Prototype not a requirement
Would like to see JMS mapping started
AR: Would like to see pub/sub, queuing works (not worried about JMS for
recommendation). Have a list of issues would be best investigated by
prototyping. Would like to have a longer period before recommendation (e.g.
beginning of May). We will not vote for recommendation before that.
JOH: Would you be able to get yourself into that position within that time
AR: Don't think this would be a good use of our time
MA: Spectrum of technical work between nothing and the whole AMQP spec... Seems
like you need inspection of key aspects of spec.
AR: Not asking for full implementation. Think technologists need to produce report
on what, if any changes need to be made to the spec.
JOH: Are you able to produce resource to work on this.
MA: Think implementation of prototyping is bigger than one org can handle on its own
CT: Think that we are voting now that we can start implementation (e.g. prototypes)
both inside and outside working group.
JOH: We expect to find errors and additions between recommendation and final
CT: Voting to say we are comfortable to invest
RG: Would like a more concrete description of what missing re queues,
subscriptions, etc
Don't want to embark on JMS mapping until core spec is finalised
DI: We are doing a detailed review and prototype
We are finding some issues which we are reporting
Would like to be involved in communal prototype discussions
This is a latered spec, I would like to get the core spec finalised and then
also get the plan for all the parts layered above the spec finalised
End of February would be aggressive... but not as long as end of May
PB: Interested in JMS Mapping
Slightly concerned over lack of common view on how comfort is obtained
End of February seems aggressive based on conversation
CT: Clearly separation of people who rely on the confidence of others
Others need to be confident themselves
We have RedHat, Microsoft and Inetco all working on becoming comfortable - we
should collect this feedback centrally and feed back to others.
RG: We should not make one persons vote contingent on another - leads to vetos.
MA: We will all need be involved - shouldn't have some PMC members sitting back
MA: We should discuss Prototyping Effort and its implications on Recommendation and
GA Votes. I think it would be dangerous to make Recommendation dependent on
prototyping.
CT: System must stay as majority vote to a veto vote
The prototype can lead to people essentially committing third parties to doing
work.
Fine for people to look at each other to influence their decision.
JOH: We have a documentation which has had a lot of work and time has been invested
If we can't sign off on this we can't invest.
If we vote on this then we lock in something we can invest in.
Recommendation is not voting as the finished item.
AT: See this as a technical voting
GA we need to have meetings on interoperability and this we can work on in May
AR: Think we need an independent review
CT: Inetco is reviewing
AT: Nobody is independent
AR: Need people who aren't Rob or Rafi to review
AT: That is what we are doing, no-one other than myself has been talking to Rob or
Rafi, so we are doing this.
AR: Should the Recommendation be closer to GA or to PR2? Think the delta should be
smaller between recommendation and GA.
MA: Two different viewpoints
1 - Recommendation should be best technical and understanding of the spec
2 - Go through the recommendation vote so community can go from a text analysis
to an implementation phase.
We need to decide which of these we are going to be voting on

• Meeting closed 17:20 *