20090510 ArB Kyoto minutes

Architecture Board

Present

John Koisch (informal chair)
Patrick Loyd
Ron Parker
Dale Nelson
Michael Van Der Zel
Grahame Grieve (scribe)
Andy Bond

Sun Q1 Minutes

The focus of the week is SAEAF roll out.

Patrick Loyd has been appointed to the position of unofficial sticky situation facilitator (due to the chairs being the primary authors and movers)

Grab bag of issues that ArB members believe we should be addressing:

• SAEAF roll out
  • Canada has a major new evolution of their architecture. One of the things to be addressed is the evolution of v3 stads suite into SAEAF
• behavioural model --> contracts
• tooling implications of SAEAF
• weigh in on vocab issue with regard to CTS2, and other cross-internal-jurisdictional issues
• HDF future directions
• what are our normative artefacts - governance process
• is RIMBAA in scope? how much?
• ballot quality initiative
• paradigms and the future - is messaging in the future
• the templates mess - just how should we do templates?

Agenda reviewed and updated as found on the wiki (Grahame/Patrick Unanimous)

SAEAF Publishing Report

• SAEAF has been broken up into 7 sections
• HDF is going to be repurposed as a SAEAF implementation guide with other content moved into CPP
  • there is still question about how this will actually happen
• parts of the SAEAF can be regarded as mature documents
• there are two internal projects officially using the SAEAF: CTS2 and PASS + several external projects (including projects at II4SM, NCI, and Infoway)
  • Mark is not formally aware of these projects - better coordination is needed
  • ArB members are supposed to fill out the Service Specification Stack Matrix Saeaf_specification_stack
  • Mark is getting more issues than traction
  • will be making a strong statement on Wed Q3

Sun Q2

Messaging is Dead! Long live messaging!

Short discussion of this somewhat radical loose statement made by the scribe which is scheduled for Tues Q4. We agree that our mission is to describe content models and to bind these to exchange models, but that the question is, who binds these things together, and how it is engineered. Of course in the end we'll have messages, so it's not really about messages, it's about how we align our internal process to produce the most consumable artefacts. There is a wide variety of views in the community on this matter.

CTS2

Extensive discussion about the CTS 2 project, potential wider CTS use by OMG, and whether this will lead to an outcome that leads to something we can use.

This problem is exacerbated by preadoption of CTS 2 like solutions by influential members of the community due to the pressure to have something to implement.

Conclusion: ArB should actively ensure that SAEAF & ECCF is followed, If we do a good job of documenting our requirements in a clear and formal fashion, like the abstract data types, a technical representation may be taken as a PIM, so we need to carefully document and describe what we are doing. We need some why to capture the fact there are architectural decisions inherent in conceptual designs.

Motion: that ArB come and advise Vocab concerning the things that need to be in the conceptual model and those things that shouldn't be, so that the committee is clear about this. In addition, to communicate back to the ArB and TSC what the findings are. Ted/Charlie (unanimous)

Action: Tues Q1. John and Grahame to attend the vocab meeting to being this process.