20100121 arb minutes
Architecture and Review Board Meeting Minutes
January 21, 2009
|Curry, Jane||Yes||ArB||Health Information Strategiesfirstname.lastname@example.org|
|Grieve, Grahame||?||ArB||Kestral Computingemail@example.com|
|Julian, Tony||?||ArB||Mayo Clinicfirstname.lastname@example.org|
|Koisch, John||?||ArB||Giidewire Architectureemail@example.com|
|Loyd, Patrick||?||ARB||Gordon point Informatics LTD.||firstname.lastname@example.org|
|Lynch, Cecil||?||ArB||ontoreason LLCemail@example.com|
|Mead, Charlie||?||ArB||National Cancer Institutefirstname.lastname@example.org|
|Ocasio, Wendell||?||ArB||Agilex Technologiesemail@example.com|
|Parker, Ron||?||ArB||CA Infowayfirstname.lastname@example.org|
|Quinn, John||?||ArB||Health Level Seven, Inc.||jquinn@HL7.org|
|Shakir, Abdul-Malik||?||ArB||Shakir Consulting||ShakirConsulting@cs.com|
|Bear, Yogi(template)||?||Guest||US Dept. Interior, Park Serviceemail@example.com|
- Call to order
- Approval of Agenda for Quarter/Day
- SAEAF and RIMBA Defered to a telcon
- Results of meeting w/Electronic services( Ron Parker)
Ron Parker: SAEAF deserves its own page, not public. There will be a public page which will refer people to ArB. It will be pointed to from other wiki pages.
Jane Curry: Instead of pointing to gforge, there is now under the ArB front page. Tony will harmonize.
Ron Parker: Met with Conformance and answered their questions.
Ron Parker: Met with Electronic services: We will provide the the HTML to be presented to the public, Electronic Services will apply stylesheets.
Ron Parker: Timelines: PIC has peer review document, Ron talked with Helen about tweaking.
Jane Curry: DId we articulate that the next iteration is due 3week before next WGM.
Ron Parker: For next call will provide project plan for our cycles. We have more budget for Karen's work - two working groups meeting for here time. We need persistant tools to manage versions/content.
Jane Curry: We committed to find out from NCI because they are selecting tools presently.
- Revisit Governance Scope
Ron Parker: Talked to Charlie about expectations - 10-11 pages - trying to scope down governance to match current release of ECCF and BF.
Jane Curry: My pushback: There is a significant scoping excercise within the governance of the ECCF. Your interoperability community is usually a multiple-community effort. Some are constrained at the ECCF. They need to be in more than one stack.
Ron Parker: I want that in. It is difficult to discuss scoping without the language. Charlie said ECCF and BF gave us a grammer.
Jane Curry: I dont have to do specification?
Ron Parker: No
Jane Curry: In the implementation HL7 will have to address coherence across domain.
Ron Parker: What is needed to support current level of participation? What would the alphas need to know?
Ann Wrightson: I am concerned that people tend to assume that when you do this you are introducing across organization. We need to make concious effort that it be larger than a single interop community.
Jane Curry: Interoperability community needs to assess where focus is required, as well as the spec stack scope. When you start drawing pictures, on what subject, it depends on the share purpose at the subject level. They may make different choices for another subject.
Ann Wrightson We discussed in ITS this morning. Some people can do this, but we need to provide a method for doing it.
Jane Curry: I disagree with your disagreement. At what point are you drawing what from where? It is not top-down driven.
Ann Wrightson I am concerned about the triangle. People only populating the implementable.
Andy Bond: Not only tech stack is computable: It should be a logical transformation. Governance can cover policy and regulation. This will control layer appropriate for community. They can get interoperability at the conceptual level.
Ann Wrightson SOme of the interoperability has not been document.
Andy Bond: It consists of some things that dont work.
Ron Parker: It is fuzzy thinking. The governance Jane needs to express is to governance in terms of ECCF - identifying shared purpose and shared governance. What is formal expression and grammer thereto. How do you govern WG's is a separate conversation.
Jane Curry: Explain "What is needed to support current level of interest and participating.
Ron Parker: Focus on the participatants. Identify formalisms or artifacts (existing or new) required for this expression. People are struggling on what the tangable things look like.
Jane Curry: I got out of tuesday a big change is the idea of authoritative requirements - who is providing, and what is their authority? From headspace? Never Declared? VS requirements stated/external with know ownership.
Ron Parker: Anything else.
Jane Curry: Interconnecting systems have to have relationships to an organization. The problem is when you cross boundaries?
Jane Curry: there are two levels - i will try not to do six.
Ron Parker: Opportunities to extend.
- Information Framework Sketch
Ann Wrightson Sketched drawing (past link here).
Jane Curry: You dont have to provide context within an environment.
Ann Wrightson all of these are concepts. It applies literally to the UML model, where there are named concepts - same framework in another direction.
Jane Curry: I got into this understanding the granularity of concepts. Definition of addressable storage/atomic unit.
Ron Parker: When I look at this, we are trying to describe a pattern to express the payload in services.
Ann Wrightson You have to bring in the semantic model.
Ron Parker: There is a whole bunch of educational wrapping.
Cecil Lynch: This gets to what I was trying to talk about -tying these together. V2 is not in the charter, but will eventually need to be. We have a v3 framework, but we need a way to provide v2. THe easy way is to use common logic. Trying to explain common logic to everyone is difficult. There are a few things we have to pull in.
Jane Curry: Syntatic model has parts that need to express.
Ann Wrightson It needs work between Cecil and I.Needs to brought forward into practical model. We have to be rigorous about names, and how you find the names. THis is fundamental.
Cecil Lynch: by syntatic model do you mean grammer?
Ann Wrightson This being a schema, but not fixed to. It is the means whereby the structure is expressed - could be a regular expression.
Cecil Lynch: IN common logic Parsed literal would be a sentence. The syntatic model provides the conjuctions?
Ann WrightsonUses the concept representation convention. What you just talked about is at the concept level. You have to do all of the above before you get there.
Cecil Lynch: Common logic covers all of this.
Ron Parker: You are in the problem solving. Jane brought up that many of the community will not understand common logic. YOu have to go through the exercise.
- Items brought up this meeting
Ron Parker: Tony will transfer JK's email into tracker.
Cecil Lynch: CDA R3 project plan is done. Item 10: ROn will take.
Ron Parker: Discussed agenda for next quarter.
- Call to order
- Approval of Agenda for Quarter
- Gforge for action items
Ron Parker: Several people will be getting notifications from Gforge.
- Schedule for SAEAF document Publishing
Ron Parker: Talked to electronic services, and will be pushing out before next WGM. Is it OK to push it out now? Charlie Mead: ECCF is ready. Jane Curry: I need to look at it again. Charlie Mead: I take responsibility for all of the corrections - will push through publishable version of BF by end of weekend. Jane Curry: Intro is original version was from May, We need to ask the intro has timing stuff in it - update all of the time, or take that stuff out. Charlie Mead: Take out timing. Ron Parker: I will review the intro. Charlie Mead: HL7 will have to build implementation guide from HDF. We have releasable ECCF, couple of days to release the BF, first draft of information framework in three weeks. Jane will have governance by Feb 1, 2010. Jane Curry: Will be at same level of scope as ECCF. Charlie Mead: HL7 and others will have to build implementatin guides. Ron Parker: When do we want to do harmonization? Needs to be halfway between WGM's. Ron Parker: There will be an anchor page. I agreed to publish a public declaration on the HL7 Page. Elevator pitch - if you want more, contact the TSC. Charlie Mead: I had to write saeaf-snapshop. I just sent to TSC. We need to all edit it. Charlie will distribute to ArB list. Goal is two pages of text and the tryptic.
Ron will put on next meeting agend.
- Alpha Facilitation
Ron Parker: Charlie McCay said that the facilitators we not known. I will have rotating agenda Item around updates for saeaf, with facilitators giving 5 minutes once a month.
- Peer Review
Ron Parker: Discussed peer review with PIC, and will tweak for our needs, circulate peer review, and complete for next WGM.
- Other business and planning
- Telcon Schedule
Thursdays at 11:00 Eastern Thursdays at 18:00 Eastern We will have standing agenda Items. Charlie Mead: The more we can make problem solving, not meetings to report what can be read in e-mail.
- Naming of the SAEAF.
Charlie Mead: Hl7 Services Aware Enterprise Architecture Framework - and interoperability framework. Ron Parker: We we talked as SOA contextualized they were happy. You could assume it is an entire framework. It(the name) seems loaded to services. We keep running up against misunderstandings. Andy Boyd: My suggestion is SAIF. Services Aware Interoperability Framework. Needs strong statement between.
- Next WGM meeting schedule (day/quarter)
The following will be submitted:
- Sunday Q1
- Sunday Q2
- Sunday Q3
- Tuesday Q4
- Thursday Q3
- Thursday Q4
- The model expresses what is to be consumed. Where is the entry point? We have a provision for expression of business requirements. Do we expect every WG to express this? We need to think about a superset of the WG, wherein enterpriseA can find their place.
- What are the disparities?
- When there is no expression of requirement in the current model I dont know how to do it.
Charlie Mead: You are confusing instances of things generated from saeaf with saeaf content. What you really are saying is does the saeaf have the grammer. Jane Curry: SAEAF on the top right should be the sum total of the things build from the saeaf catalog. Charlie Mead: This is the SAEAF derived HL7 . Ron Parker: Next thing to do, as an architecture board, how are we providing guidance on how to use the HL7 assets. How are we providing a way for a 'newbie' to consume the content. Ron Parker: Requirements for a WG are understood and generated by a WG, not expressed as architecture understand by other parties. Andy Boyd: There should be a layer of SAEAF's. Then there is an instance (WI) on the diagram. We dont play SAEAF as the enterprise architect. Saeaf is not a container for instances - just a reference framework. Ron Parker: We need to make it real. We need an entry point to understand the spec, and the mapping to the architecture. Charlie, we need to talk about this soon. If I was a large enterprise wanting to bet the farm on this, I would want to know persistance and traceability. I would not invest in this if I could not map my architecture to the HL7 WI. Charlie Mead: I have collegues who think I am leading NCI over the cliff. NCI believes that HL7 stuff that is NOT SAEAF is safe - solid/stable. SAEAF brings BF, and testable conformance framework. NCI is signing up and trusting HL7. NCI wil do this irrespective of HL7 - use framework to build their own implementation guide. NCI is depending on the grammer, not implementation guide. Ron Parker: Not sustainable model - it can't just be NCI. This idea formalizing requirements can be convuluted. Lloyd McKenzie: How far to building artifacts? Ron Parker: Requirements space - nada. Lloyd McKenzie: HL7 has only story boards for requirements. Ron Parker: Blueprint 2015 is building a business architecture. Jane Curry: From the artifact list, what had to be validated was a scope statement. Ron Parker: EHR-FM fits there. Someone has to filter, do checklist, and validate what exists. This allows someone to play with the enterprise. Vendors have an entry point into the requirements spec. Ron Parker: I propose work activity. Lloyd McKenzie: Until tools are in place, we have nothing. Charlie Mead: NCI is investing in tools that it will give away. Expect HL7 to have input. Lloyd McKenzie: How can you do it without the artifacts? Charlie Mead: Two projects have developed some artifacts. CAEHR team running under SAEAF, is suppose to flesh out artifacts, which will drive the tooling. Lloyd McKenzie: Would be quite useful to feedback to MnM, vocab. From the Arb. Charlie Mead: Not from the ArB, from the projects. They have complex interrelationships that must be managed with tools. Andy Boyd: There is an expectation that we have to fill in all the blanks, or we fail. In some cases we have to extend what we have done. Lloyd McKenzie:My expectation we will have requirements. From the dynamic model perspective it will change - static model will stay the same until we realize expansion. Cecil Lynch: Is there expectation that someone will say At this pism level you will use this cmet? Lloyd McKenzie: Making services aware, you remove assumptions on how things work, or organized, or structured. All messages make assumptions about content. In services constrained schemas instead must be partitioned. Need artifacts at the universal level to accomplish. The real services aware part means looking again at the universal level. Cecil Lynch: That is why this is dynamic - we constrain in a tool - in such a way it makes sense. Lloyd McKenzie: You have a dynamic environment, developing the static model has to reconfigure itself. Ron Parker: When organizations engage, i need chunks - i have embraced the model. Goes to fundamental model of providing soa design. Lloyd McKenzie:1. Getting the modellers into that state, 2. Normally, you say here is my enterprise, so I can do soa. HL7 does not have a single enterprise - we have hundreds. Ron Parker: Different classes of consumers. Solutions at low level are problematic. Lloyd McKenzie: I understand how to do it in Canadian context, but in the diverse framework environment, it gets more challenging. Jane Curry: There are implications for system boundaries. Documents have no system boundaries. Messaging have system boundaries, but no implied behavior. Ron Parker: What level of WI do we need to express. Jane Curry: People have to adopt cost to finish the work. Cecil Lynch: We are providing a model - eventually it needs to be concrete. ECCF requires backward chaining. You have to be able to compare what you are doing to the model. Then you can suplement what is missing. Ron Parker: Methodology may evolve. Lloyd McKenzie: Core approach. Historically we have defined structures at the lowest level. Most cannot be implemented at the univerasal leve. You cannot implement any spec without constraint. We can derive schemas with XS:ANY, or constrain. Cecil Lynch: Different level of compliance. Ron Parker: Next working group meeting, who(what audience) needs to discuss this. Lloyd McKenzie: I expect a couple of phases. Changing the currents requirements gathering. Ron Parker: Beyond what we have expressed with SAEAF book, there is another relation to requirements gathering. Lloyd McKenzie: We will not start tooling until we have designed the whole stack below:
SAIF Implementation Guide
- Requirements Acquisition CIM level (with tooling)
- Artefact Development
- Atrefact Balloting
- Artefact Publishing
- Registered use of Product
6:00pm Eastern On jJanuary 28, 2010 Late,
Adjourned Tony Julian