20090415 arb OOC Minutes
Architecture Board
April 15, 2009
Contents
Call to Order
Meeting was called to order at 9:00am by Chair John Koisch with Tony Julian as scribe.
Attendees
| Name | Present | With | Affiliation | E-main address |
| Curry, Jane | Yes | ArB | Health Information Strategies | janecurry@healthinfostrategies.com |
| Davis, Brian | Yes | Guest | 3rd Millennium | bdavis@3rdmill.com |
| Grieve, Grahame | No | ArB | Kestral Computing | grahame@kestral.com.au |
| Jorgenson, Don | Yes | ? | Inpriva | djorgenson@inpriva.com |
| Julian, Tony | Yes | ArB | Mayo Clinic | ajulian@mayo.edu |
| Koehn, Marc | Yes | Guest | Gordon Point Informatics Ltd | Marc.Koehn@GPInformatics.com |
| Koisch, John | Yes | ArB | 3rd Millennium | koisch_john@bah.com |
| Loyd, Patrick | Yes | ARB | Gordon point Informatics LTD. | patrick.loyd@gpinformatics.com |
| Lynch, Cecil | No | ArB | ontoreason LLC | clynch@ontoreason.com |
| Mead, Charlie | Yes | ArB | Booz Allen Hamilton | charlie.mead@booz.com |
| Mulroney, Galen | Yes | ? | VHA | galen.mulroney@vha.gov |
| Nelson, Dale | Yes | ArB | II4SM | dale@zed-logic.com |
| Ocasio, Wendell | Yes | ArB | Agilex Technologies | wendell.ocasio@agilex.com |
| Parker, Ron | No | ArB | CA Infoway | rparker@eastlink.ca |
| Quinn, John | Yes | ArB | Health Level Seven, Inc. | jquinn@HL7.org |
| Shakir, Abdul-Malik | Yes | ArB | Shakir Consulting | ShakirConsulting@cs.com |
| Walker, Mead | No | ArB | Health Data and Interoperability Inc | dmead@comcast.net |
Agenda
- Q1 Review SAEAF, Governance, Behavioral Framework, RASCII?
- q2 Issues of conformance/compliance v/v frank Oemig's issue, RASCII, then to Roll out
- q3 Enterprise Architecture Roll out, Matrix
- EOD gap analysis,
- Thu RASCII chart
- Q1: SAEAF -
Q1
Charlie Mead: We committed to do a next version of the SAEAF. We were surprised that we got as much acceptance as there was. There was a lot of stuff that needed to be worked out. Governance was there as a vague concept. Tried to have ready for Orlando, but the magnitude involved. We are trying to unbind ourselves from tool specific/architecture specific. EA will bubble up, but it could take 4 billion years. The content was getting significantly large to break into modules. The idea was that the decks would be the outline for the documents. The candidate document is being written. We are on the edge of loosing control of the document, it is at 400 pages. Charlie has asked for formal paid editor, graphics person to manage the graphics, and a way to publish as a book.
John Koisch: This is a similar experience to what the NCI is seeing. We are not trying to subvert or stop grass-roots development of projects.
We are putting heavy restriction on the development. THe TSC needs to make the architecture primary - not a bakseat
thing. Otherwise, people will continue to do what they always did.
Charlie Mead: The objectives are to generate the documentation and education of the SAEAF. The 'F' is the most important letter - it is the top/down creation of a framework in which all of the work of HL7 occurs. It needs to be managed as a corporate asset. We need to get to another level.
Jane Curry: We need to do content management.
Charlie Mead: Yes, it is very hard. THe decks and the document are interrelated.
John Koisch: NCI is trying to keep on WIKI. It is not ideal.
Abdul-Malik Shakir: We need to see the architecture as a living document.
Charlie Mead: It is a framework, and it will grow.
Abdul-Malik Shakir: The chapters are at different levels of maturity. SOme of it is part of the mind-set - some say they are following the HDF, which is not finished, and some things will become lore. The idea of having dedictate editors is a good idea. Editors cannot author, only maintain coherency.
Charlie Mead: The management of text book of medicine. There are contributors, editors, master editor, and graphics editors. They take pages out, and put pages in.
Galen Mulroney: There is broad agreement with Charlie, at least three people have mentioned the affect of the SAEAF with other projects that are going on concurrently. We need a HL7 governance that controls the rollout. e.g. HDF may have to be told that we are changing things, and you need to comply. We may step on toes and hurt feelings.
Charlie Mead: We will step on toes, and hurt feelings.
John Quin: We are spreading widespread fear about a document that has not been finished or approved.
John Koisch: John Quinn, with regard to this, what is the thinking about balloting and formal approval.
John Quin: Balloting must come from the ARB.
Mark Koehn: We work within what we understand. THe roadmap says 'a balloted architecture'. The change-management whether bolt-on or change , it needs to be way better than 70 slide decks. We are trying to find the major milestones, decision points...
John Quin: Balloting is an overloaded term. Peer-review is also overloaded. We need to be careful of the terms used. Balloted to some people means 90% accepted, but it could be DSTU. We need to propose the transition to the TSC. Eliminate the chicken- little aspects.
John Koisch: On agenda for Q1 tomorrow - later today we will discuss conformance/compliance, HDF (are we able to influence), I believe that we need to act with authority
John Quin: We act with the authority we have - it is unusual, not dictatorship, but very high. Mark, we have not done our job to give you the materials you need. What is going to be our response, we have been open, and there are personal issues with those who have not attended the meetings.
Galen Mulroney: We are ignoring organizational change management.
Mark Koehn: We will talk of these things, not 'authority', but a roadmap to the future, being clear and open.
Charlie Mead: We can give you point-by-point examples of the pain-points, of implementing a framework, not an archecture. The hardest thing at NCI is continual reaction of those who feel micro-managed, when in fact we provide a framework.
Abdul-Malik Shakir: In addition to unintentionally hurting people, there are those who will offer resistance. We should anticipate them and work closely with them to minimize specially SOA, MnM, Conformance/Compliance.
Brian Davis: Which is what Charlie and John did at NCI.
Dale Nelson: At the MnM meeting there was a topic of discussion on how to engage with the framework: we feel the need to be involved but where, and at what level?
Galen Mulroney: The SOA RASCII is a first step.
Abdul-Malik Shakir: When you come in at 2:00, after telling your spouse you would be home at 11:00, you need to have your story straight.\
John Quin: We need documents, etc, but we cannot use that to communicate outside this group. The phantom attention is not on the slides the ArB has created, just a snapshot of the snapshot.
John Koisch: The problem we have is "How do you define enterprise architecture?" is that the communication must flow both ways, from the projects to the architecture. The framework is a communication tool to make sure the communcations flow both ways. (Presented the graphic from usadotay i.usatoday.net/tech/graphics/iss_timeline/flash.htm.) This is an excellent graphics that shows the interfaces - an elegant example of the Zacherman drawing.
Charlie Mead: Zacherman defined the differences between engineering and architecture.
John Koisch: Defines the connecting points. This is what we are trying to do - governance happens at the interfaces. We dont have the bandwidth for the components, only the interfaces. The Arb cannot care about the components.
Mark Koehn: You hope the guy who approaches the soyuz interface can walk into the station.
John Koisch: There are definitely places the metaphor breaks down: Every project is affected, e.g. "id looks like this"
Mark Koehn: How we state this is the project.
Dale Nelson: Illustrates one of the big challenges. With all of the interfaces specified, does this allow everyone say 'I conform, but can do what I want in my sandbox'.
John Koisch: A good thing?
Dale Nelson: allows lip service without true conformance.
John Koisch: 10% of the horse leads, 10% is the ass. I dont have the bandwidth to make you a good citizen, but I can make sure you play nicely.
Abdul-Malik Shakir: The cultural change - ArB needs to recognize that the organization has gone through change before: V3 methodology has created fear of change - in the early days MnM were the backroom zealots. We cannot become same. Also the 'not-invented-here' syndrom. We need to reuse things. MnM does not have a handle on change management. Education has a problem.
Tony Julian: We have implemented change management, but 30% are going around it.
John Koisch: People are willing to work with the framework, but just tell me what I need to do. Several alpha projects want to do it now. CTS2 recognizes that they cannot specify without the framework.
Wendell Occasio: THe aproach needs to be an agile approach, put what we have out for excercising - it is complex enough that it is all not mature. Lets get it out, and get it excercised.
John Koisch: we need to do it, or they need to come up with something else. The definitions will be part of the use.
Charlie Mead: That is what the alpha projects are about.
Jane Curry: Everyone wants to be a alpha project.
Jane Curry: Reports
Jane Curry: Open health tools wants us to use the terms conformance/compliance loosely, switching definitions after january, and need to tease apart what governance for whom? The Governance framework went through 5 iterations, and the 6th is being done to define the primative artifacts, instead of the complex artifacts, i went from 75 slides to over 100.
John Koisch: The BF paper was released 5 weeks ago, to be included in the SAEAF. Charlie was to write the other pieces then lay it in. I released the deck this last weekend, I have since made some further changes, that are more implementation focused. As I have tried to make it less dense, I think it is at the right level, for an educational deck for the behavioral framework which was added to the SAEAF, since the existing dynamic model did not support the behavior. THe BF does not need to belong to the top document, but how it gets accomplished, is important. We may pull out an examples deck, which the behavioral deck needs.
Jane Curry: The same topic has different examples depending on the jurisdiction.
John Koisch: the BF feeds into and relies on conformance/compliance. I would encourage everyone to look. I added service classification scheme in detail.
Wendell Occasio: Where the comments I included on tracker in the latest?
Tony Julian: Clear the trackers when the item is included in the document.
John Koisch: Woody and Lloyd will come in tomorrow and talk about moving the BF into the MIF. We need to discuss artifacts and vocabulary that will need to be included.
Abdul-Malik Shakir: Is the BF refined enough to map to the MIF?
John Koisch: Yes
Abdul-Malik Shakir: Will that play into the tooling issues?
John Koisch: Yes
Abdul-Malik Shakir: Will we have a transitioning discussion?
John Koisch: Yes. We need to discuss.
Jane Curry: The tooling is later this year, and the migration strategy from current to new needs to occur when we have a handle on the timing of new tooling. THis calendar year will not change the tools.
Abdul-Malik Shakir: Is this year not about rolling out the SAEAF?
Mark Koehn: We need to have a gap accessment, which I need to produce.
John Koisch: the problem with the BF is that there are tooling solutions and standards, which we tried to back our way into, clearly from the begining. I am using them now, because I need to. THe tooling is an aid.
Wendell Occasio: THe first set of comments for the ECCF there was a dual track. I gave Charlie a spreadsheet.
Charlie Mead: THere were six rounds of the spreadsheet.
John Koisch: A task for charlie is we need to harmonize them.
Tony Julian: Upload the spreadsheet, and close the trackers?
Wendell Occasio: Is there/will there be a practical users guide?
John Koisch: The BF is a practical users guide?
Wendell Occasio: It needs to be across the SAEAF. Tell me what to do, and what bucket?
Mark Koehn: This is a problem of no map. How is it instantiated? We need to itemize the places they are.
Galen Mulroney: When the MIF came up in tooling, we are familar with the issues. Do we care what the MIF looks like, or is it just an interface? - with responsibility to MnM or tooling.
John Koisch: The framework is about capturing requirements.
Galen Mulroney: So, how does the MIF fit in.
John Koisch: The ArB has to say about the state of the MIF as per integration semantics.
Abdul-Malik Shakir: Do we go beyond specifications?
Jane Curry: Arb has responsiblity, but others have the roles.
John Koisch: We are at break: Charlie the conformance discussion is Frank Oemigs stuff, and the BRIDGE conformance/compliance.
John Koisch: We will take a break, and then return to have Charlie discuss the above, while Jane and John 'smith' the RASCII.
Charlie Mead: A few of us were at the first meeting, and made a list of things to do. We did not launch SAEAF until May, but we discussed a business model of the organization, a project was approves, the affected were listed, and we decided that we should do a framework. SO the TSC would handle what happens after the framework was developed. We can talk about Frank Oemigs stuff in 5 minutes. We dont want to publish an enterprise architecture, just a framework. I think we can come up with something reasonable. I think we could re-calibrate.
John Koisch: that is the reason for the RASCII.
Mark Koehn: How did we arrive at the buckets of work in the RASCII?
John Koisch: we did not define the buckets.
Mark Koehn: we need to define the real work/
Charlie Mead: Would that be helpful?
John Koisch: Doe the chart?
Jane Curry: We know we have to do different ones - since this one is SOA.
John Koisch: I was hoping jane and I would re-do the columns, using BOD, ArB, OMG
Mark Koehn: What is the boundary? How do we know which goes to whom?
John Koisch: This is soa work-groups document, we differ on some buckets. We need to re-constitute to deliver to the TSC.
Galen Mulroney: Foundation SD wants to re-do?
Wendell Occasio: What is the RASCII to the RASCII?
Mark Koehn: We need to have a broad project plan, with the 5-7 big buckets, filling in the roles should be homework.
John Koisch: We have dicussed disagreements with SOA, but there was no advancement.
Mark Koehn: We need to refine into others: We need to define process, confirm it, and do it.
John Koisch: TSC needs a RASCII of RASCII's?
Jane Curry: As we introduced, We should do one from our perspective, resolve conflicts with liasons, and take the conflicts to the TSC.
Galen Mulroney: There should be an official starting place.
Don Jorgenson: This was a communication tool from the SOA wg's viewpoint. Gather from all the groups, and harmonize. Community built does not seem to be the direction. You cannot push it.
Mark Koehn: start with charlies mode. What decisions, when, and propose a RASCII with the left column, I would be willing to do.
John Koisch: Break return at 10:55.
Q2
Charlie Mead: Frank Oemig'a document is a document of conformance in V2. Alan Honey did a detailed map to HSSP. We need to do a similar thing to Frank's document. If there is an application in SAEAF we have to determined whether there is a teminology problem - does he use the same terms. We need to do the same level of analysis.
John Quin: Try to remember that German is Frank's languages, and his definitions are concise. He has a crystalline structure, and if we need to move him it will take some work. Patrick Loyd: Is that possible.
Jane Curry: He uses conformance in the same terms. He believes that compliance is not releavant. He has levels of conformance in the sense of he does not use the compliance.
Wendell Occasio: Whyw was it published?
John Quin: A work group defined it.
Galen Mulroney: Either we use the terms that HL7 balloted, or come up with a good reason to change.
Charlie Mead: We used the definitions from RM-ODP - we did not make this up.
Galen Mulroney: I like that argument.
Abdul-Malik Shakir: What makes it correct?
John Koisch: We relied on ISO definitions in RM-ODP.
Abdul-Malik Shakir: We prefer to use the ISO terms.
Charlie Mead: We use the terms in the balloted RM-ODP.
Jane Curry: That is what we said.
Wendell Occasio: This was done by HL7 Conformance/Compliance.
John Quin: The argument that ISO has defined the terms Frank will accept. HL7 works with ISO. If it was an issue where ISO differed, then it would be a more difficult issue. If ISO has a definition, we should discuss it.
Dale Nelson: The V2 artifacts are ANSI standards.
John Koisch: I added an item to the glossary of terms.
Jane Curry: We reference EHR-FM, but not conformance profiles.
John Koisch: We talk about service profiles.
Charlie Mead: Would Wendell and Jane be willing to look at this?
Jane Curry: No bandwidth.
Wendell Occasio: Consistancies with the HL7 published content
Jane Curry: It was published in 2008.
Patrick Loyd: The document from the document you sent was corrupt.
Jane Curry: I can send it out.
Wendell Occasio: RM0ODP talks about conformance and compliance. We want to use HSSP and SOA-ML as much as we can. Cull out anything specific to 2.x - if it does not apply to V3.
Jane Curry: It is a V2 proposal.
Wendell Occasio: Conformance for V3?
Dale Nelson: It has been balloted.
Wendell Occasio: If it is a V2 document, why is it crucial?
Charlie Mead: It has been put forth as "This has already been done, why does not the ArB comply with this?".
Wendell Occasio: Was it analyzed in the context of V3?
Jane Curry: From what I can see on the document, it was designed with the expectation that it would be used for V2, and could be the basis for V3, but there is no explicit V3 definitions.
Charlie Mead: This makes our job easier. We need to map to SAEAF.
Galen Mulroney: That wIoana's question, why did we not start with this document?
Charlie Mead: I will try to help Wendell after a first cut.
Mark Koehn: Do we have a log of issues? So at the end we have a track of issues?
John Quin: We have done an extended peer review, and there should be a log.
Mark Koehn: It would stand us in good stead if we have the log, and the principles used to make the decisions. You can debate and disagree.
Wendell Occasio: It was articulated in the Introduction to the SAEAF.
Galen Mulroney: So we have documented that we are useing RM-ODP. We need to analyze the conflict, we will have the same with different people, to anticipate futures.
John Quin: HL7 has done things on its own, where there is not reasonable material from other standards bodies. There are now a number of othere standards that we should adopt instead of building our own. Modeling we broke every tool we touched.
Charlie Mead: Good Idea - we need to have a log of issues. e.g. why not use soa-ml? AMS did an analysis of why it could not be done. This Frank Oemig's thing is the same.
Mark Koehn: This will be successful if it does not have 5000 rows.
Charlie Mead: Why did you not do this?
Abdul-Malik Shakir: It is a matter of choice.
John Quin: it is acceptable if we have used documented.
Abdul-Malik Shakir: Use CTS2 instead of ISO.
Mark Koehn: Choices were made based on principles we defined.
Charlie Mead: The choices are on a continuum, some we are sticking to, there is a catalog in the first few slides showing why we made the choices we did. They were approved by the Board and TSC.
John Quin: They should be in the heading of each document.
Charlie Mead: Wendell will look at the document, and Patrick will help.
Jane Curry: It is a good document, well defined.
Charlie Mead: The sooner we can address, the better, and include in the ECCF document. PL: there are people who hear the word service, and they immediately close their minds.
Charlie Mead: From the Intro Deck: Starting at Slide 23ff where we adopt 'The Lens of SAEAF'. The purpose of it was to define what we got from where. In the document it quotes the standard.
Jane Curry: Have we pulled out the principles?
John Koisch: Yes, they are service principles.
Jane Curry: Rational for why RM-ODP works better thatn TOJAM. We got handled a bigger problem - we have to define conformance across the whole enterprise.
Brian Davis: You did a good job, but there are additional things to consider.
Dale Nelson: We need to say it better (slide 23).
Mark Koehn: To me, the answer is right - we should use it. This is part of the educational plan.
Charlie Mead: People have to understand that there is a change in the pathway, but the full path has not been resolved. People look at it and ignore until it is done.
Brian Davis: You can use the work these other people have done.
John Koisch: There are people who have not read RM-ODP, but use TOGAF. We need to authority or trust that the issues have been addressed - there needs to be a clear message. We dont have all the answers, and the distrust.
Jane Curry: CSI drives a lot of this- without it we would not need the others.
John Koisch: If you take away any of the slices, you loose important notions.
Charlie Mead: Your perspective is based on the fact you love information.
Jane Curry: Yes.
John Koisch: We can go into why we chose these - CSI is very important - it changes the game.
Mark Koehn: We have a set of principles, resonable people can debate the choices - this thing will go through it's journey - we have a set of artifacts. We need the discussion of the consequences of existing artifacts. Are we willing to have that dialog?
Charlie Mead: We could walk through the recalibration.
John Koisch: In the beginning HL7 was an oval, but we made it a square to show that it encompases.
Galen Mulroney: CCOW does not exist in the oval.
Abdul-Malik Shakir: Yes it was the first.
Charlie Mead: We have 30 minutes:
- Recalibration
- Charged to do the EAS.
Charlie Mead: An EAS needs to emerge, but without a framework it will not. We need to re-calibrate. There is one paragraph in the ECCF document Paragraph 2.1 and 2.2. Everyone including John to read - This comes from our fundamental at the NCI.
John Quin: I would rather this group do the mappings, than the organization members.
John Koisch: The charge is to build the SAEAF that includes conformance and govarnance. and to publish as EAS to be published as the HL7 Enterprise Architecture Stack(SS).
Mark Koehn: Non-architects read other things into this document. Our consumers build space stations, HITSP, and others cannot do it although they can build networks. EAS need to be defined better.
Wendell Occasio: If you are talking about the constraints of HL7, versus the constraints of an individual consumer. In the ECCF we have gotten there.
Jane Curry: That train led me down to find the primitives in the artifacts.
Mark Koehn: What do you mean in the EAS?
John Koisch: One thing would be a scheduling service specification - (or any other paradigm) at NCI we will have full-on implemented things. The EAS is the sears catalog - you open it up, you order it, and you get that.
Galen Mulroney: Is the scheduling service specification an instance?
Charlie Mead: An EAS to HL7 is a set of constructs that support interoperability that can support others. E.G. Datatypes is a component. It will show up in a number of other components. It is a primitive. Engineering has to build primitives that can be used by others. CMETS were supposed to be primitives, just not worked out that way. Specification stacks are containers for artifacts. ArB needs to specify the primitives.
Galen Mulroney: I thought I heard you say that there could be multiple architecures that would fit in the framework? Will CDA have its own architecture?
Charlie Mead: Look at the MDA from the SS, at the conceptual level all will have the same artifacts, but they will look different. Canada Infoway and NCI could interoperate by looking at the constraints to determine the gaps.
Galen Mulroney: The EAS in question are the consumers, who should describe theirs in relationship to the SAEAF.
John Koisch: The notion is not that everyone is SAEAF compliant: It makes the specification implicit, so that at least they will understand, and be able to map to their stack. The point is that we will never agree to agree, but if we build against common specifications, and the components are explicit enough, we can figure out how to interoperate.
John Koisch: It is not a system specification.
Mark Koehn: We have the framework, which will be how we build specifications. Some words need to be changes to make it mor understandable.
Charlie Mead: There are two architectures: We need to make our stuff in the same framework.
John Quin: We need to get our arms around it. HL7 has spend 98% of its existence figuring out how we do our stuff, only 2% of how other use our stuff.
Jane Curry: If it does not specify how an organization consumes the archetecture. Where did the constraints come from?
Charlie Mead: ANdy bond from Austrailia said "We are talking about the aspects of architecture that effect us.
Mark Koehn: Implementing the SAEAF means implementating the EAS.
Charlie Mead: Part 2 is governing so that it works.
Charlie Mead: Governance bolts them together.
John Quin: The last 10 minutes need to be put into graphics/slides.
Abdul-Malik Shakir: That emergences is not certain or spontaenous. Someone needs to do it.
Charlie Mead: Hence governance.
- MIssion/Goal
Charlie Mead: We have an SAEAF and a EAS timeline we need to bolt together
Mark Koehn: What does that mean? How do we do it?
Charlie Mead: We have done something, and the framework path has been adopted by the BOD and TSC. We are doing it!
- Commitment
- TSC, BOD 16 months ago
- Deliverables
- coordination
- Impact
- Open issues - we need to keep an issues list.
Charlie Mead: we have published a table of HSSP. AMS published a nice analysis of SOA-ML.
Galen Mulroney: WHere is it?
Charlie Mead: SOA-ML is in the BF document.
Charlie Mead: We need to produce the coordination. There are others adking why we did not do it the other way.
Mark Koehn: ECCF specification stack - are we assuming at the end of the game there should be an artifact for each intersection,
where and when is that clarity (why do I need to build what)?
Charlie Mead: We have 3 projects that are trying to do those documents - 1 had produced a candidate list.
Mark Koehn: What committees?
Charlie Mead: All projects will be involved in all aspects.
John Koisch: we need to stop working as committees, and start working as projects.
John Koisch: There are elements in HL7 that can build a platform-independant model.
John Koisch: ECCFapril09-v5 slide 14 assumes another layer of primitives - the reference things. Mentors could be assigned to each cell. It has to be project-oriented, and behavioral.
Mark Koehn: Are we ready for it? You have to fake the rational portion e.g. on this spec you can ignore box 1 and 2.
John Koisch: I sent an eclipse process framework that uses this as a workflow. If you buy into rup, and the boundaries. MK How are the primitives will emerge in a concensus based mechanism?
Charlie Mead: They will emerge by architecture.
Jane Curry: They will emerge from those familiar.
Don Jorgenson: It looks like the ArB pushing up will support.
Charlie Mead: NCI will prime the pump, a short term team of a few people in place to build architectural primitives - a few core level services. You know they are primitive, used across projects.
Jane Curry: We used to say data-types was a Lower level protocol implementation. This is a turn-around.
John Koisch: THe architecture emerges from architectures, not implementors.
Don Jorgenson: Those solving healtcare problems, should be outside of their worries.
Charlie Mead: Were there any missing points
Q3
Wendell Occasio: Presented a analysis of Frank Oemig's proposal for Conformance/Compliance.
- Two separate efforts to define conformance
- Both documents attempt to provide definitions and a framework for "conformance".
(See slide deck Concept Comparison)
Abdul-Malik Shakir: Specification is a better term than standard.
John Koisch: We call a balloted specification a "Standard". This is an opportunity to draw that out.
Abdul-Malik Shakir: Standards and Profiles could be specilizations of specification.
- concept 3
Abdul-Malik Shakir: V2: compliance is comparison to a standard, conformance is comparision to a profile.
Mark Koehn: If rule a is original and rule b is the derived - is there a rule?
Wendell Occasio: An RMIM needs to be compliant with a DMIM.
Wendell Occasio: In the SAEAF we use the term compliance for constraint, but it may also be applied to the binding. e.g. binding to SNOMED
Charlie Mead: ECCF slides 38-40. define SAEAF compliance v/v RM-ODP.
Abdul-Malik Shakir: Given a PQ, which may be integer or real, they are not consistent.
Wendell Occasio: You can follow all of the rules of compliance and conformance and still be incompatable.
Jane Curry: There is potential loss between patterns. The selection of constraints my yield incompatable results.
Wendell Occasio: You have a sender/receiver sending PQ. IF the sender constrains as integer, and the receiver as real, they are compliant. If the sender constrains as real, and the reciever as integer, they are not compliant.
Jane Curry: Not certification of the implementation, but compatibility as well.
Abdul-Malik Shakir: Compatability also comes into play for backward compatability.
Mark Koehn: a big leap that constraint is a compliant behavior?
Jane Curry: Conformance at the instance level.
Wendell Occasio: conformace is to rules.
Jane Curry: Tooling uses validation at the artifact level as well as the tooling level.
Q4
Mark Koehn: We have work that builds the framework. We have alpha projects. We know we need to develop EAS, components, primitives, tooling, and certification of the EAS to the SAEAF. We also have an idea of the endgame. All of the committees will review and participate as well as build primitives. We have not evolved to define the seeet of activities. We need a roadmap.
SAEAF
John Koisch: Explained the SAEAF stack matrix. Each cell has a list of artifacts that releate.
Mark Koehn: When is it "Done"?
- Elaborate the SAEAF matrix.
- artifact types
- cross-lane compatability/reconcillation/harmonization
- Best Practices for development
- Quality Criteria
- some way to be formally accepted by the organization or mandated by the TSC/BOD?
Charlie Mead:
- candidate framework,
- BOD acceptance,
- Implementation guide/machinery,
- alpha projects,
- review SAEAF in view of projectst.
John Koisch: Formal balloting?
Charlie Mead: Formal acceptance.
Jane Curry: TSC approval?
Mark Koehn: Whatever we do, people will throw rocks at it.
Charlie Mead: notion vs content. the SAEAF defines the concepts, but the implementation guide is important.
Mark Koehn:
- Candidate SS (SEAAF BOD/TSC accepted)
- Iterations of SS/SAEAF as influenced by alpha projects
- Peer review - we dont have a mechanism for a formal peer review.
John Koisch: lets work backwards: In Vancouver we talked to Charlie McKay about sunset projects based on the SAEAF. Do we need soft approval, then usage?
Mark Koehn: soft approval is a non-ballot approval route.
Wendell Occasio: How about a broad statement that no work can be done that does not fit within the framework.
Jane Curry: Governance and Operations Manual (GOM) already has this in it.
Wendell Occasio: The end result is a framework, with an update to the GOM to include the criteria to require conformance/compliance to the SAEAF.
Charlie Mead: We stumbled on the fact that there is an architecture to HL7. The purpose is to generate pieces that go into other peoples architecture. At the last reality check, the business archetecture changes were not the ARB's problem.
Wendell Occasio: Not the ArB's problem, but certainly belongs on the activity diagram.
Charlie Mead: We make recommendations, the TSC will do it.
Mark Koehn: The conditions of the SAEAF will trigger a GOM update. Is there anothere activity between the states of the SAEAF taking into account the Alpha work.
Jane Curry: we need to have a form of peer review, then apply the changes, and document the changes.
Wendell Occasio: GOM should specify the governance requirement for the SAEAF.
John Quin: DO the current processes in the GOM meet our needs? If not, we suggest a new process, and provide to the TSC.
Wendell Occasio: Process for broader concensus for which we dont have a process. We commited to explain the journey - the major decision points.
Charlie Mead: I dont think the TSC will allow any major direction changes to happen.
Wendell Occasio: We need to document the journey - if the TSC approves it, then it is done!
Charlie Mead: How do we get projects to work.
Charlie Mead: rollout was to get alpha projects to verify the pardigm. I can speak for the services paradigm from an external project by September 2009. Ron Parker has a messaging project he can do.
Wendell Occasio: Is there a requirement for a project that cust across more than one pardigm?
John Koisch: Were we limiting ourselves to 3?
Charlie Mead: TSC wanted at least one for each paradigm.
Alpha Projects
The discussion below talks to the following points:
- Outside projects:
- Need contract/commitment - obligations of participants
- Knowledgeable of the SAEAF
- Use Stack
- Use HL7 Core Stuff
- Uses/contributes to intelectual property of domain commitees
- Willing to give back.
- Artifacts
- Engagement/feedback on the SAEAF/Processes - lessons learned process/tooling
- Engage through governance processes.
- Need contract/commitment - obligations of participants
- Recruitment
- Inside Project Engagement
- Ask SD's to call for projects
- Ask committees who we know are creating new standards
- CDA-R3
- EHR-R2
- HDF/core principles
- Privacy and security services
- CTSR2
- OO
- Responsibilities
- Example artefacts
- Feedback on the SAEAF
- Ballot would be good, but there may be a tooling problem.
- inside project
Russ Hamm says CTS2 should be ready
Don Jorgenson: PASS could cross the paradigm, with basic access control.
John Koisch: what is the outcome of the alpha projects?
Jane Curry: artifacts and verification of the process.
John Koisch: Do we want an ANSI balloted artifact.
Mark Koehn: Do we have to ballot to drive the changes to the SAEAF, or just a full-featured development.
John Quin: We have funding to build tools, etc. ONCHIT and CCHIT are in favor of the project. We have a plan for tools for both publishing and end users.
Jane Curry: BF has impact based on the current methodology and creation of the MIF.
Mark Koehn: process will pop out tooling requirements. Who is documenting the requirements?
Brian Davis: CABIG is documenting the gap analysis.
Mark Koehn: whose job is it?
Don Jorgenson: Alpha leadership should make recommendations to tooling.
Jane Curry: also feedback loop from tooling to alpha.
Jane Curry: there are placeholders in the tooling plan.
Mark Koehn: is there a resourcing issue?
Jane Curry: It is being addressed.
Galen Mulroney: there have been a series of meeting with MnM, InM and tooling to solve a problem. Our decisions are being made also. Alpha produce tooling requirements, but we are producing tooling requirements. We need to head off the tooling decisions being made.
Jane Curry: the tooling considerations have been announced, and have a 2 point plan: those things we need to do no matter what, as well as taking advantage of current requirements.
Galen Mulroney: do we need to make a interim review of the current direction of tooling?
Mark Koehn: How do we align, I suggest that it needs action.
Galen Mulroney: Those making decisions who are not formal members of the ArB are not aware. We have to be carefull about surprising people. CDA R3 will be another problem, so we should cut it off in the bud.
Mark Koehn: Continuing on the alpha projects: What are the major activities of the alpha projects:
Charlie Mead: I dont totally agree: Finding projects outside of HL7 is easy - they are willing. Even if they are successful, there will be those within HL7 who will not accept it. We dont know messaging and documents. I would alpha the alpha project commitments.
Abdul-Malik Shakir: How far should it go?
Charlie Mead: at least one level.
Abdul-Malik Shakir: I have an analysis project I can do by September.
Dale Nelson: Uses/contributes to intelectual property of domain commitees
Charlie Mead: Willing to give back.
Mark Koehn: do we expect the project to bring the artifacts forward through a notional feedback?
Charlie Mead: If we dont engage, we will bifurcate.
Abdul-Malik Shakir: Are we going to require the correct?
Charlie Mead: They must create top-down.
Abdul-Malik Shakir: Some are useing UML instead of HL7 notation.
Jane Curry: we are defining the differences.
Abdul-Malik Shakir: some will do datatypes R1 instead of R2 since R2 is not in the RIM
Don Jorgenson: Is this an implemention
Charlie Mead: We are looking for knowledge about the stack/ implementation.
Wendell Occasio: Do they need to cover every cell of the SS?
Charlie Mead: I think there are more projects - we need to take a look at all of the projects> PL: FInding an implementation project that is DAM building.
John Koisch: is the external project meet the criteria
Mark Koehn: How do we recruit alpha projects?
Charlie Mead: External recruitment process comes from those around the table who know of projects.
Mark Koehn: the challenge is internal?
Dale Nelson: what is the value of doing a projects? We write papers about them?
Abdul-Malik Shakir: We tried this in the implementation committees, and they never shared their artifacts.
Mark Koehn: Is the concensus that outside ones are easy to specify? Specially with the changes to existing artifacts.
Dale Nelson: Outside projects are easy to find from open source.
Charlie Mead: We have talked about internal and external projects. From a change management process, can we invite internal projects, and if they dont show up, then we will still learn something.
John Quin: anyone developing HITSP requirements, or CDAR3. If they dont choose to participate, then we let them in later.
Charlie Mead: Then we tell them the rules we made that they will have to handle.
John Koisch: Outside are set up to manage the projects. HL7 is not set up for that e.g. the meeting space is apportioned by committee, not by project.
Mark Koehn: If we found one or two, we could get a room for them.
John Koisch: There are barriers for entry. There are ways to game it out internally to make it more attactive. OO is a ripe candidate, but it would have to be a project, headed to a standard, not just a test project.
Charlie Mead: there are significant differences from external projects that are due to SCOPE. NCI is building a SOA, not just one service. We can learn the mapping of the SAEAF, BF, and SS from an external project.
John Koisch: attend conference calls, attend WGM's to discuss content.
Don Jorgenson: Outside sponsored.
Charlie Mead: the difference is that the groups that showed up were those creating an implementation, not trying to build huge concensus.
Wendell Occasio: Galen was mentioning internal decisions that may have impact on our stuff, that could be a project.
Jane Curry: There are t Wendell Occasio: EHR- R2 and EHR- R3, which have opportunities to align with other stuff, if we educated them to the value of adding the value of the SAEAF to their project.
Mark Koehn: How do we make them happen? Why would the projects not be directed to change their scope to be alpha projects.
John Quin: I cant imagine why they would not want to take the opportunity.
Charlie Mead: Can we encourage them?
John Quin: Yes.
Jane Curry: Bob Dolin is interested, but he does not know what it means.
Don Jorgenson: proposed PASS and CTSR2
Mark Koehn: Looking at what was involved, alpha project could write a charter, the the ArB, etc to support?
Don Jorgenson: Yes, it would.
Mark Koehn: We are working on CTSR2, PASS could be done - if we had these four would that excercise the model? If we spent energy on these four, do you feel that it would kick the tires?
Charlie Mead: There is a political divide that over CTSR2/ OMG
John Koisch: OMG wants to run it through the ontology group.
Charlie Mead: There is a project to build it without going through OMG.
Mark Koehn: I would excersize the goevenance.
Don Jorgenson: SOA, CDCCC
Mark Koehn: Is there a scope statement for PASS within the HL7?
Don Jorgenson: Even though HSSP is involved, we have dispensation. We are ready to ballot in fall. With broad support.
Jane Curry: Infoway and OHT are behind it.
John Koisch: You may disagree - but now there are two flavors of projects - HDP and core projects.
Dale Nelson: ICSR is starting a new project.
John Koisch: what happened with OO? PL: We had use cases, but with no implementations - although they are coming forth now. We are starting ANSI balloted projects. All implementations profile the UV rmim to make a profiled RMIM. The hurdle for messaging is that no-one has implemented the UV specification - a localized version cannot be implemented.
Wendell Occasio: The UV models are not supposed to be profiled.
Galen Mulroney: Something Charlie said threw me: You discounted a project because it would not be implemented.
Charlie Mead: I discounted the internal projects that would not be implemented.
Galen Mulroney: What is the scope for alpha projects?
Mark Koehn: IT would be cool to have a running project.
Charlie Mead: I said there was a project that wants to go to implementation, but cannot go all the way within HL7, so would not be an internal project. PL: OO would be one:
Mark Koehn: We should look for outside projects, and then ask the steering divisions to provide projects.
Mark Koehn: Then we should have a scope document.
Charlie Mead: We should review those for the selection criteria.
Mark Koehn: We are testing with CTS2, waiting on John and Russ. We are going to recruit, document, then run. Then we want our committees to engage through the matrix.
Charlie Mead: I thought you were going to say the projects should report to the ArB.
Mark Koehn: We (the EA) dont know how we work throught this.
Charlie Mead: Like a cat?
Mark Koehn: what?
John Koisch: Meow a lot.
Mark Koehn: WHat is the mini-pmo around the project? The ArB.
Charlie Mead: The arb has to be aware, but not manage the project.
Mark Koehn: Review points, issues
Jane Curry: Rascii C, consultant role, mentoring.
Mark Koehn: How do we sketch that behavior? How do we claify that?
Charlie Mead: There are two kinds of governance - architecture and project. The bodies are in communication, but they are different sets of people.
Mark Koehn: The committess will do the work, and then materials will pop-out. Is there an implied gating process, or exceptional process only. Is there govenance with ArB engagement. We need an interim architectural rule.
John Koisch: They should submit to the governance within the SAEAF?
Jane Curry: it has to follow vocab and rim, quality criteria for the content. The process would be helpful for the less-familiar artifacts around the BF.
Mark Koehn: I understand, but who is accountable for evaluating quality.
Jane Curry: I will produce the quality criteria. At the minimum, the teams should bring exceptions to us.
Mark Koehn: At the end of the project I will give you the places I did not follow the criteria? Or do we need to define the process?
John Koisch: The ArB now has to pretend to be a governance body - with the caveat that we understand there may be incomplete/misunderstood artifacts?
Wendell Occasio: Do we have responsibility for compliance?
Jane Curry: WHen we were reviewing the V2 spec, we found that we could not be inspectors, but rather served as an arbitrating board, as well as delivering criteria for inspection. Are we a gating process, which then would have us bypassed. And serve as an arbiture board.
Wendell Occasio: What is the expected outcome?
Jane Curry: A feedback between concrete data, and you can require demonstrating of self-checking.
Mark Koehn: We do not know who is accountable - ballot producers, ArB, TSC, BOD. We assume that the 'Gates' are under governance. We let the projects run, and see what issues percolate, then modify the GOM.
Jane Curry: There is another relationship - from the learning process to tooling. Part of the quality requirement is testing by tooling, instead of manually -find and trace.
Wendell Occasio: Specially for the internal, is it ok for us to not have any veto or say in there output?
Don Jorgenson: Could we have complete requirements?
Wendell Occasio: So what if we do it as lessons learn at the end, and miss the opportunity to modify the frameworks?
Mark Koehn: NCI produces stuff, they give us lessons learns, then we give them a certification?
John Koisch: There should be some level of veto power.
Jane Curry: Leaning for us is what needs to be in the criteria, or dimensions, so we can learn. Test the criteria, look at the artifacts to see if they are best practices/ harvestable artifact. It is so early in the process, othere than the completeness things, I would not be comfortabe to specify the criteria.
Mark Koehn: We need to assess at the end how close we got. We dont know who is responsible, and someone needs to call 'UNCLE' and come the the ArB. What does the excption process look like? My assumption is that the implementation project doing the workflows issues, what do we need to learn, how do we engage the projects, touch base, and when do they kick-out things. DO we have a kick-off process?
Jane Curry: Face to face?
Mark Koehn: I dont know -confirm the project.
Abdul-Malik Shakir: Each project does their normal stuff, then brings it here too see if it qualifies. We need a project to review the projects.
John Koisch: Would the outcome be what does architecure govenance looks like?
Charlie Mead: External has to relate to the ArB.
Dale Nelson: Suppose it is sort-of-compliant.
Charlie Mead: We would like to publish, let people use, and comment back on the defiencies.
Mark Koehn: That what makes an alpha project
Jane Curry: And see what we learn from it, and see what they do with our comments.
Charlie Mead: NCI may find things that wont work.
Wendell Occasio: Dynamic is different internally.
Charlie Mead: II4sm will be on commercial software developement cycle.
Mark Koehn: assessment and learning process. We should turn this into a roadmap in visio.
Galen Mulroney: when you take this home, and unwrap it, will you know what it means?
Mark Koehn: Yes. The next step is to breakdown the responsibilities. TO perculate into the RASCII.
Galen Mulroney: we have the EA project, and have not discussed their resposililities.
Mark Koehn: We need another conversation.
Charlie Mead: Is the EA committee a directed aspect of the TSC?
Mark Koehn: we need to figure out what it will take. In Orlando we sketched it out, now we are at phase 1. We dont have enough resources to do all of this. SO we dont have a committee in the normal sense. Whe need to define the amount of work, and how we get it done. Mapping out the assessment, bringing forward, and itemizing them is the Role of the EA rollout committee.
Charlie Mead: I am concerned that this will become its own waterfall. For the two projects I have direct contact, are going to start, hopefully with a threadbare notion of the governance.
Mark Koehn: That is good in theory. The forces necessary are not there.
Charlie Mead: Maybe we should scare the TSC, and tell them there are two projects that will start sending in spec. stacks in two weeks.
Jane Curry: I wont be able to look at it until June.
Charlie Mead: Should we let them take the framework as is?
Wendell Occasio: What features do they need to excersize?
Charlie Mead: SAEAF and BF.
Jane Curry: I am harvesting primitives from the other viewpoints.
Charlie Mead: Please publish before Kyoto.
Jane Curry: I plan to.
Mark Koehn: There is nothing to stop us from kick-starting the process, but we need to document this.
Don Jorgenson: If we are going to shift our PASS-AUDIT etc. to the balloting, how can you process the the project in the timeframe?
Mark Koehn: We would have to spend 1/2 day or more on a project to see what is valid.
Don Jorgenson: THe project facilitator will not know what to do?
Mark Koehn: YOu could use the SS as a management framework? What does it mean to the ballot process? THe one who gets there will be able to assess what things are missing.
Charlie Mead: There are tooling requirements.
Jane Curry: It can be published in PDF.
Charlie Mead: The artifacts are informational.
Jane Curry: yes The tricky part will be if it is RIM based and dynamic model?
Mark Koehn: we are caught up in the tooling - are the process and business rules to be excersized. Ballote what portions are informative, and normative - and starts to drive the EA.
Charlie Mead: We are expanding the set of semantics, running some through the tooling path, others not. THen we will revise the tooling.
Mark Koehn: If all we can offer is DSTU, what is the risk?
Charlie Mead: The board has agreed that the SAEAF is a better way to specify the artifacts.
Mark Koehn: It revalidated the strem, and got explicit actions. We did not find action for steering division calls - who is responsible?
Charlie Mead: Steering division reps. to the TSC.
Mark Koehn: We never came back to the HDF issue.
Dale Nelson: That is tomorrow.
Mark Koehn: Yet another output to be discussed.
Charlie Mead: WHen Ioana show up we will offer to re-write the HDF as a SAEAF implementation guide.
Galen Mulroney: I think Ioana would be a good resource, but we got off on the wrong foot - the HDF is not what it was intended to be, but original methodology got ripped out, now is the time to put it back in.
Mark Koehn: All of the core instruments , as part of the end game, the HDF will be absorbed, . . .
Jane Curry: we have a process alignment issues, under different timeframes, nor do any references the others.
Galen Mulroney: HDF predates the SAEAF, maybe it should be rolled into the SAEAF.
Jane Curry: It is not an ArB responsibility, the rollout should do that part.
Mark Koehn: Key actions are assigned. For PASS and external projects, can we kick-off whatever needs to be done to draft a charter. If someone has a framework, and want so alpha the alpha engagement, Charlie, then we could pattern for the other projects. Additionaly we know the initial scope stuff does not have the things we need.
Jane Curry: SOme of the recent project scope statements are already aligning with the SAEAF.
Mark Koehn: For Kyoto, which we need to align, do you have time to discuss the ArB agenda details?
Brian Davis: COmplements on the project management, do we have a final time?
Mark Koehn: I was afraid to ask that. Hoping the governance is done early 2010. Charlies project will pop-out issues by september 2009.
Jane Curry: we are resource limited, unless we get external funding. I know tooling will not work on a volunteer basis until after 2010.
Mark Koehn: we need the requirements for tools done by early 2010. If this was resourced, designing the changes would be simple, except for the acceptance. One could say by summer next year it would be fleshed out.
Charlie Mead: By september there will be 'interesting stuff', but there will be more.
Mark Koehn: Enough output to cutoff Version 1. Then we can work on Version 2. To make it done by date y, we will need to ask the alpha projects to bring their feedback by x.
John Koisch: There are expectations with the ONCHITT effort, maybe we can only fund one project.
Jane Curry: The list of necessary item to support the framework to flesh out an get developed.
Mark Koehn: If we know when we get them, the can we say by May 2010 Guidance, GOM etc is done, tooling is on the way. If that is what we want to get to, lets take off, and determine what the resources are.
Brian Davis: The top of the timeline: Is it your view that the first thing that has to be done is the templates be created and fleshed out?
Charlie Mead: Lots of it is done.
Jane Curry: For five-or-six we have fleshed out.
Charlie Mead: WE have candidate artifacts.
Brian Davis: You have thought this out.
John Koisch: It is too explicit for what you are asking. Very rarely do I see someone saying "this is wrong", but rather format, terminology, etc.
Jane Curry: The September WGM line, assumes to achieve a draft .9.
Mark Koehn: If this is may, and this is final, you could run alphas through december, so you have people ready.
Charlie Mead: THe biggest problem, for a project to go on the timeline, are service projects -CTS2, II4SM.
John Koisch: OO is one> PL: IF you can stop before ballot OO will be there - where we can play.
Charlie Mead: That is great. and you can execute on that timeline. Galen thinks he can get a CDA project into VA.
John Koisch: YOu are always to negative charlie.
Charlie Mead: Learning from John. DId you get what you needed.
Mark Koehn: Yes, we need to anchor some if this in artifacts, we dont have an anchor point, do we need to write flowcharts for everything we need to do. I will see if my group consents. I think so.
John Koisch: I modified the agenda for tomorrow. Adjourned at 5:18pm.
