20090922 arb Atlanta wgm

Architecture Board

September 22, 2009

Name	Present	With	Affiliation	E-mail address
Bond, Andy	Yes	ArB	NEHTA	[| andy.bond@nehta.gov.au ]
Curry, Jane	Yes	ArB	Health Information Strategies	[| janecurry@healthinfostrategies.com ]
Grieve, Grahame	Yes	ArB	Kestral Computing	[| grahame@kestral.com.au ]
Julian, Tony	Yes	ArB	Mayo Clinic	[| ajulian@mayo.edu ]
Koisch, John	Yes	ArB	NCI	[| jkoisch@guidewirearchitecture.com ]
Loyd, Patrick	Yes	ArB	Gordon point Informatics LTD.	[| patrick.loyd@gpinformatics.com ]
Lynch, Cecil	Yes	ArB	ontoreason LLC	[| clynch@ontoreason.com ]
Mead, Charlie	No	ArB	Booz Allen Hamilton	[| charlie.mead@booz.com ]
Nelson, Dale	Yes	Arb	II4SM	[| dale@zed-logic.com ]
Ocasio, Wendell	No	ArB	Agilex Technologies	[| wendell.ocasio@agilex.com ]
Parker, Ron	Yes	ArB	CA Infoway	[| rparker@eastlink.ca ]
Peres, Greg	Yes	Guest	Infoway	[| gperes@infoway.ca ]
Quinn, John	No	ArB	Health Level Seven, Inc.	[| jquinn@HL7.org ]
Shakir, Abdul-Malik	No	ArB	Shakir Consulting	[| ShakirConsulting@cs.com ]

Call to order

The meeting was called to order at 3:45 P.M. U.S. EDT by chair Ron Parker with Tony Julian as scribe.

Agenda approval

MMS to approve the agend below Tony/Loyd Vote:(9-0-0)

• Call to order
• Approval of agenda
• Announcements
• PASS - Cecil
• BF - John Koisch
• Governance - Jane
• Planning for rest of meeting - All
• Adjournment

Announcements

Per email from Charlie Mead:

• John Koisch to resign as CoChair to avoid “NCI overload”
• Ron Parker to assume CoChair position
• CHI to add architect (info per Ron)
• Andy Bond to join ArB (to address “preponderance of influence” issue)

• Please let everyone else on the ArB know about this. In addition, the TSC has asked that the ArB produce a formal Membership Process and Criteria document detailing the ArB’s position on membership qualifications, nominating procedures, Preponderance of Influence potential, etc. The TSC stated that it controls all membership decisions around ArB but wants ArB input as to our position on these topics.

PASS

Cecil Lynch: Question is they are doing a services profile for user authentication: How do they map it to SAEAF? Want it to be reusable across any industry. They have a DAM, and will create rim-based instantiation, which will get them to Platform Independent. They are still at the conceptual level. There use case is passing identifiers - you have to know the source. II datatypes are not alwayse supported by other industries. First concrete advice we have to give.

Jane Curry: Does that mean any of the rim-derived data structures are not platform-independent because of data-types?

Cecil Lynch: Not cross-industry. Only works if DAM does not include data-types?

John Koisch: Is the question methodological?

Cecil Lynch: Need guidance so they can create generic hand-off to OMG, and fit SAEAF.

Ron Parker: How are we going to provide advice for that?

Cecil Lynch: Jane was correct. We should respond to what a DAM is. I volunteered to write the definition of a DAM. Mead is going to write it, and Cecil will help hammer it out. Make explicit definition, to address issues. Must be clearly defined to be used cross-industry.

Ron Parker: Where is this stuff placed? Must be posted, and be accepted.

Jane Curry: How do we publish it?

Ron Parker: Not in just there world.

Jane Curry: We now have a Gforge that could be used. Publishing has the need for a shared-artifact repository, including relationships to other artifacts. With ability to pull out custom packages.

Ron Parker: in the absence of that, do we have a recommendation?

Jane Curry: People need to register the project, post the scope statement in insight, get gforge access, and post it there.

Patrick Loyd: Is that the way it is to be done.

Jane Curry: no, it is just an idea.

Ron Parker: Need to formalize for Alpha projects.

John Koisch: Not gforge - it is a swamp - we need to consider a formal folder structure.

Jane Curry: not posting anonymously.

Jane Curry: we dont have enough knowledge, and dont know recommendations.

John Koisch: you need tortise. There is the issue of models, and how to publish. EA? Otherwise it is a governance problem.

Ron Parker: Agreed, we have to put a framework for them to publish into. It must be visible.

Jane Curry: we need to practice what we preach.

Dale Nelson: Where is it, and how do we use it? I tried to get the document,and I could not find the latest document.

Jane Curry: We have gforge documentation, but it is not well published, or communicated. If part of the SAEAF alpha project includes that, then we should incorporate the intructions.

Ron Parker: Who will take point

Jane Curry: I will take point. We need a recommended structure.

John Koisch: Another thing to publish, so recommend a folder structure in SVN, forcing people to use the files publication in GFORGE.

Jane Curry: Formalized version control.

John Koisch: Different view for outside users than developers. If we develop formal framework, we will have modeling issues. There are solutions we should look at.

Ron Parker: Any precident from TSC?

John Koisch: We have R3, which includes EHR functional model, PASS will put in anything, OO will put in detailed dynamic models, CTS2 will put threads of rainbow documents.

Jane Curry: we need an action item with a deliverable. I will ask Wilfred Bonney to work with whomever, so we can draw from what he has.

Ron Parker: Identify methodology for view and project use.

Jane Curry: Thursday Q1 collaboration tools from electronic services. - should align.

Ron Parker: We need push model to get it to alpha team. Do we have an education process.

Jane Curry: Short gotomeeting walkthru, including MnM in the process.

Ron Parker: So you (jane) are going to kick this off.

Jane Curry: We have two MnM co-chairs in the room.

Ron Parker: Prefer we take it to MnM quickly.

Jane Curry: Who can take to MnM on friday? We need concurrence from MnM, and communication to alpha projects.

Ron Parker: MnM participants in this room.

Dale Nelson: Publishing governance of ITS.

 Tony will post list of facilitators 
 EA Rollout Facilitators

Jane Curry: ArB asking MnM to identify the artifacts to be included int the Grid. First order of business is to produce documentation: What artifacts do we produce, and where do we put them.

Ron Parker: MnM is not ready for this.

Jane Curry: Today MnM would point to HDF as methodolgy.

John Koisch: Does MnM define folders?

Jane Curry: we hand it off to MnM; we do it and force MnM to use it, or we work together.

Ron Parker: MnM has their way now.

John Koisch: Part of rollout is impacts to publishing.

Ron Parker: My proposal - learning process. Once it is sorted out . .

1. Framework introduces new artifacts.
2. We declare them.
3. we sit with MnM and ask what would fit without any problem.
4. What is white space, where are the gaps, and are they relevant.
5. Post as part of methodology.
6. MnM assert discovery process
7. facilatators push to projects.

Patrick Loyd: we have generated artifacts and interactive. We have tooling needs, the first versions will be manually produced.

Ron Parker: I know MnM will get very interested. For Alpha we will need slack.

Jane Curry: We can say that.

Ron Parker: we need to do CF further.

Behavioral Framework (formerly known as Dynamic Model)

Johns Presentation

Jane Curry: I need to show governance framework thursday.

John Koisch: Further modeling efforts. BF separates accountability - who does what when. Services, HL7 Version 3 messages, documents define this using different foundations.

• Services formalise accountability for accountability around the roles.
• HL7 Version 3 message specs formalice accountability to sender/receiver
• SDA accountability is deferred to the receiver
• business process - can be formalized behind interfaces, cn be evaluated more precisely.

Grahame Grieve: The ADA details minimal accountability.

Patrick Loyd: They are not orthogonal. HL7 Version 3 messages is a transport mechanism as well as the transport. Documents dont describe transport.

Ron Parker: I don't disagree.

John Koisch: formalize accountability in a contract.

Grahame Grieve: Services and messages bind at a different time, CDA defers to the user.

John Koisch: what we are trying to do with BF is to give artifacts that they already understand. We want to make sure as we transform between messages/documents/services we do it right.

 All should read the education documents posted at:
 http://gforge.hl7.org/gf/project/saeaf/scmsvn/?action=browse&path=%2Ftrunk%2Fdocs%2Fpresentations%2FalternateFormat%2F

John Koisch: Contracts maek specific rules agout content, platforms, and localizations that may or may not be part of a given statndard. e.g. contracts may contain quality of service agreements.

John Koisch: HL7 Version 3 O&O separated responsibility into three parties. (Projected POLB_IN111100).

Jane Curry: tooling is waiting for the BF.

John Koisch: The dynamic model has never been formally approved.

John Koisch: We are not only changing models, messages, documents - the needs of our customers have changed. We need drive by interoperability. You cannot rely on business model.

Grahame Grieve: It is still valid.

John Koisch: Now we have a world where the customers are enterprises, having infrastructure, need structures to apply to their business model. We have presumed for months that our concepts are valid. When you separate accountability conceptually, and it is still valid a lot of the time. But there are a lot of organizations that have formalized infrastructure and processes.

Jane Curry: you cannot pull in design conceptions in the model level.

Ron Parker: From a national project this is a big deal.

Grahame Grieve: We are talking about a maturity model for the industry.

John Koisch: I am not sure that if you apply to existing, it will not be mature.

John Koisch: HL7 Version 3 has captured the semantics of the vertical model of interactions better than other industries. We re-organized it based on feedback. What were were trying to understand it HL7 Roles. (required for RM-ODP - which shoves them onto roles). We can build solutions out of conceptual, platform independent, or platform independent. The solutin package contains the contract template, with a rich behavioral model. We can ballot role: HL7 Version 3 ballots receiving roles + event driven packages. We can ballot behaviors. Now we must ballot dependencies.

Jane Curry: we already said we had to do it.

John Koisch: Dont know how to get around it. half of drive-by interoperability needs services.

Jane Curry: business now understand that the rules are now important.

Ron Parker: Where have you said that?

John Koisch: I said it public to combined educational session. Don Mon was there, ed, Austin was there. This is a formalization of all the change we have discussed.

Ron Parker: What do we have?

John Koisch: Q3 we have a project with OO to provide dynamic model we need.

Grahame Grieve: Express appreciation for John's work.

John Koisch: Wednesday Q3: Hans J. Buitendi

John Koisch: "The agenda is focusing on Composite Order to understand how to apply the Behavioral Model work in progress, as well as get better insight how to ensure that content that is to be conveyed through various mechanisms (using Composite Order and CDA R3 as an example) is to be modeled/documented. The following will be used to test it: a Lab Order and Medication Order with a Lab Test for context as part of a Composite Order, while that same lab test is also used as context in a CDA R3 document that also may document the existence of the medication order that was communicated through the Composite Order.

In today's world and understanding, the way to apply the SAEAF behavorial model to the payload is not clear, and ensuring the core payload is consistent regardless of exchange mechanism is not clear. Hopefully Wednesday Q3 and Q4 will provide substantial progress in these two areas to enable OO to complete Composite Order (and subsequently Clinical Event Reporting), while it can help other work groups as well as the ARB and MnM promote consistency in approach across work groups."

Cecil Lynch: CDA R3 will be consistent between models.

Patrick Loyd: We invited different groups. We will have to do most of it. CDAR3 will have to push consistency. OO will be behavioral framework.

Cecil Lynch: Dont worry about CDA - it is already an alpha model.

John Koisch: The first step is to establish the accountability. We are not doing anything different.

Governance

ran out of time - will take up later.

Agenda for rest of week

• Finalizing agenda for week

• • Thursday Bake-off
    • what does CDA-R3 look like?
    • What are the best way to go?
  • Thursday Q4 Governance
  • Conceptualize HL7 Projects (per JK).
  • Products portfolio contextualized in the SAEAF (per JK)
  • Timeline for the BF formalized acceptance by the ArB(per JK)
  • Other business, and planning for next meeting
• Adjournment

Adjournment

The meeting was adjourned at 5:15PM

Tony Julian, Scribe[Tony Julian]