20080605 arb minutes

Architecture Board

June 5, 2008

Attendees:

Charlie Mead, John Koisch(early part, Tony Julian, Yongjian Bao, Jane Curry, Mead Walker, Ron Parker(latter part), Rich Rogers, David Hamill,

Charlie Mead-chair; Jane Curry – scribe

Agenda:

• Attendance (ArB members+ observers)
• Protocol for meeting (who talks when)
• Scope for first meeting as well as trajectory to September
• Physical logistics for all three meetings

Minutes:

CM indicated that HL7 would only underwrite costs for ArB members. On that note, since JC had been an active participant of the ArB since before January, and after discussion with John Quinn and Charlie McCay, Jane Curry will be added as a formal member of the ArB.

• Logistics for June meeting & forward:
  • 10 observers, which is the limit
  • 7 ArB members attending in June
    • Only ArB members can vote –
    • Observers can’t go over previously plowed ground, only discuss on topic
  • CM proposed fixed length QA every so often – say every 2 hours, 15-30 minutes interaction. JK, friendly amendment last 30 minutes of each quarter be the interaction
    • Consensus was 3-hours work, 1 hour interaction; 1hr lunch; repeat – at least for the first 2 days, last day spend the pm in wrap-up mode
  • Dates July 15-17, Aug. 19-21
  • HL7 cost is the primary decision criteria and since Rockville offers meeting space at no cost to HL7 and at least 2 members not requiring funding if meeting there, then the consensus was to stick with Rockville for all the meetings.
    • 7 currently anticipated in July – Rockville - NCI
    • 5 currently anticipated in Aug – Rockville – NCI

• Primary focus for initial architecture work is for Services, because it is the least understood – and we don’t want to spend all our time doing the overarching EA and not achieve a Services context draft by September. But if we can do the whole thing – all the better.
  • MW expressed concern that the perception of not respecting current messaging and document paradigm was a risk that must be managed
  • RP indicated that his experience showed that the 3 paradigms were compatible in that explicit models were the key exchange mechanism in all 3 cases
  • The sticking space is how you define messages – payload is static information expressed in a model in any case, but how you deal with externalized messages versus internalized interfaces is a distinction.
  • Dynamic model comes into play so we really will need to tackle this anyway.
  • HL7 has traditionally stayed out of application design, where Services has something to say. Point was made that V3 may not address application design but did start specifying application behavior by defining receiver responsibilities

• The main objective for next telecon is to firm up the agenda for the June meeting so it can be published since there will be observers.

Action Items

none Tony Julian