20090806 arb telcon minutes
Architecture Board
August 6, 2009
Present:
John Koisch, Jane Curry, Charlie Mead,
Cliff Thompson ontosolutionsllc cliff@ontosolutions.com
The TSC has requested the ArB provide a policy statement RE domain analysis models, their context within SAEAF, whether it “makes sense” for a DAM to be published as a normative standard, and if so, in what contexts.
The TSC has requested that the ArB provide this statement no later then Monday, August 24th.
Charlie Mead: We will draft something in Michigan.
Issues:
John Koisch: what is our involvement with vocab/harmonization. WOody and Lloyd want to spend a lot of time on the Behavioral framework, to get it into the MIF.
Jane Curry: we have timing issues from the current BF to the future. That is where Woody and Lloyd are focused. I dont think we are ready.
Charlie Mead: John Quinn says we want to get off the MIF. OMG is looking at it: Lloyd has provided requirements to OMG. We should let things stabilize with Alpha projects and OMG discussion.
John Koisch: There is work going on with specification. Our discussions about what the ballot will look like. There is no clear way we will publish the EA rollout information in the MIF. When they say it is too complex, it shuts down the production. Woody has valid issues, but when someone comes to the table with "this is how we have always done it".
Jane Curry: The fallback for publishing is to publish in a gif, not necessarily machine readable artifacts. The first thing is human understandability, then computable artifacts. We will be in a transition for a number of years. Human artifacts are based on industry-standard tooling, then we will address platform independent and platform dependent. People will try out different techniques for artifacts - when will we understand it, when will we be able to tool?
John Koisch: It comes down to what you are trying to do. In HL7 , the ArB is trying to get people to be specific, not optimized. I suspect Charlie will support me. The gains on being specific come in agility, not computability. I think that is what John Quinn is looking for. It must be brought to ground. We could be computable with the BF soon. It is a paradigm shift - computable does not mean in the MIF.
Jane Curry: What tools will you consume - traceability is essential, but we must have plans to encorporate the artifacts. DAM, will it be balloted normative, international, us affiliate. This set the nature of the domain, as demonstrated with the BRIDG model when you bring in new players. There is a learning curve. What stability is the DAM to sustain when you increase the scope. We need to understand before we worry about normative.
Charlie Mead: The most immediate thing is the agenda for Michigan. The biggest issue about the EA rollout is what artifacts should be generated. We have a good Idea based on the NCI work what the artifacts look like. WE, arb, do not have an idea what the ECCF looks like for messages or documents. There is a call the first day of the F-F for structured documents. They dont understand what they are supposed to do. We (arb/nci) have operated that the issue is in the PIM/PISM rows. With the exception of the TSC thing, is the notion of how to lead the alpha projects - it is expected that the EA projects will do work in Atlanta, with Arb supervision.
John Koisch: THe notion of using a DAM may be clear to some people, but in cross-functional teams, working with bridge and extensions, working with the DAM is a huge shift for most people. That is true about a lot of this stuff. Most teams are not agile enought when choosing standards, aor interactiong with standards. It comes down to how people look at this. Is the DAM a class model? we have to walk them through it. If HL7 adopts DAM, it will require another shift.
Jane Curry: We dont have the walkd from DAm to DIM.
John Koisch: Everybody keeps talking about project based support, but no one is doing it.
Charlie Mead: Interesting. TSC sees projects running in the dark. Project can do the initial steps, then falls into a black hole. There is one that is out of control, and the TSC is beginning to see it. John, your responses are timely. What project?
Charlie Mead: I could, but we are crashing into the DCM project. It is a nightmare with DCM trying to discredit the EHR project publically.
John Koisch: HL7 project?
Charlie Mead: NO, the DCM project is surfacing things from the dark. If they choose to work in the dark, they can do a lot of work, and still be in the dark.
John Koisch: Human business process runs smack into technical process.
Jane Curry: the degree of collaberation of those who signed up, had a lot of sponsors. Who is speaking on behalf of the project, and who are they vetting their comments with.
Charlie Mead: Asking?
Jane Curry: Who is vetting the stuff? They make statements in isolation, the next point at which declarative statements should be made is in Atlanta.
John Koisch: I have it in my head. Lay it out q/q?
Jane Curry: I am writing the governance paper.
Charlie Mead: I have given up asking.
Jane Curry: I will write stuff around the problematic issues, governance.
John Koisch: I am working around 3 things with the BF: one is a paper that is shorter.
Charlie Mead: John Quinn says HL7 will announce the award of the RFP for the SAEAF book. Soon. Good you
are working on it.
John Koisch: Tuesday q1-q2 rollout review
Charlie Mead: What does that mean? Will we assign point persons? Structdoc is expecting at least one member, John or Charlie. What is in the review?
John Koisch: EA Rollout Review - inventory of projects, expectations, POCs, inventory of artifacts (existing and gaps)
Charlie Mead: John Quinn is expecting that each of the projects will be represented by the ArB.
John Koisch: Focus oin the differences between interoperability paradigms
John Koisch: Sign up for ArB facilitator
Charlie Mead: Who wants to be a EA project facilitator. Otherwise we will not be able to support. Marc is expecting dedicated resources.
Jane Curry: who will be able?
Charlie Mead: Those who have committee responsiblibities may not.
John Koisch: Sign up for capability(per capability).
John Koisch:Dale would be a good candidate for the engineering viewpoint. A mentor per project.
Jane Curry: brings up one of the things, wednesday would be good, with harmonization joining, difference between composite and atomic artifacts. We have artifacts that stretch across viewpoints. The business rules may be expressed in other viewpoints as they relate to those viewpoints. No one has put anything else on the grid. We need those sorted out.
John Koisch: we have had productive discussions. Dale is going to do a paragraph on the problems of separating out artifacts. We have had problems at NCI: discrete artifacts are for the birds. If you build a BAM, you dont just create processes, the refer to concepts. Same with a DAM. It is pure to think of primitives, but I would posit that in fact the only primitives that we care about are services and service specification, which are atomic holes that can be reused practically.
Jane Curry: I dont disagree about decomposition, but I do think we need recognize when we are dealing with aspects in one view point that belong within another - it is discordant if you dont.
John Koisch: revisit composites : what is intended to be re-used, or used: who is the target audience.
Charlie Mead: what HL7 is trying to do, with the notion of an EA framework, we will have to confront the issue of primitives and composites. CMETS were an attempt that does not work well. We have dodged governance. It includes Janes concepts, but governance like HL7 has not experienced. A soon as the TSC understands primitives and composites, primitives to be derived from the alpha projects, it is time to have that discussion
John Koisch: Wednesday Q1-Q2 Behavioral Framework Update
John Koisch: Wed Q3-Q4 Primitives, GOvernance, and Composites
John Koisch: gets out the possible things. That leaves thursday:
Jane Curry: we have to have before we leave what we are going to accomplish between Michigan and the WGM, and what our objectives are.
John Koisch: Thursday Q1 - TSC request for " ArB provide a policy statement RE domain analysis models, their context within SAEAF, whether it “makes sense” for a DAM to be published as a normative standard, and if so, in what contexts."
John Koisch: Thursday Q1 - TSC request for " ArB provide a policy statement RE domain analysis models, their context within SAEAF, whether it “makes sense” for a DAM to be published as a normative standard, and if so, in what contexts."
Jane Curry: we need to lay out what we are going to try to accomplish - development, what are out objectives, and how do we organize our time.
Charlie Mead: what objectives?
Jane Curry: what are we going to bring to Atlanta, what are our activities need to be to meet the objectives. We have to start to schedule.
Charlie Mead: That is what Marc is expecting. Charlie McCay expects us to leave atlanta with a facilitator for each rollout project, and each rolling.
Jane Curry: then we need to figure it out, and what we should do differently.
Charlie Mead: NCI can provide quantitive examples. NCI is doing deployment, HL7 will be doing PIM's or PISMS. We are most efficient with document projects. John and I spent an afternoon trying to fit a messaging project into the methodology. John, do we have this wired up?
John Koisch: It was a good excercise, with a good model. CDA is a good case, I could give you good guidance on how NCI is thinking about the CDA, but it is not down to computable rules. GOvernance comes down to what do you want to govern? How to direct their efforts so some governance group can look at it and pass.
Charlie Mead: Remember when we started this last summer? NCI said "we will tell you what we are doing. Allan and John Koisch have been doing work with the BF. This is a place to start." we spent june questioning, july understanding. If we want the EA projects to be successful, we need to tell them how it will be governed. Let NCI put the stuff on the table, I expect that Ron Parker has some experience: I know from OHT that Andy was supportive. Ask everyone to read the IF document. We need to come up with a candidate governance structure. It may not be right, but they can try them, and see if it works. EA is not an emergent property.
Jane Curry: People need to understand what artifacts, and level of quality to be compliant. Yes, there is room for manuevering when their are unknowns, but there is some element of objective or measurable criteria. We must have something concrete.
Charlie Mead: if we focused on artifacts and governance, we could askt th TSC question in 30 minutes.
Jane Curry: I agree - balloting is about artifacts and governance.
John Koisch: we have an agenda. John will put on the wiki.
