This wiki has undergone a migration to Confluence found Here
<meta name="googlebot" content="noindex">

20100516 arb rio wgm agenda

From HL7Wiki
Jump to navigation Jump to search

Architecture Board

Sunday, May 16, 2010 Rio de Janiero Working Group Meeting

Sunday Q1


Name PresentAffiliationE-mail address
Bond,Andy Yes
Curry, Jane Yes Health Information
Grieve, Grahame Yes Kestral
Julian, Tony Yes Mayo
Koisch, John No Guidewire Architecture
Loyd, Patrick No Gordon point Informatics LTD.
Lynch, Cecil Yes ontoreason
Mead, Charlie No National Cancer
Nelson, Dale No
Ocasio, Wendell No Agilex
Parker, Ron Yes CA
Quinn, John No Health Level Seven,
Shakir, Abdul-Malik Yes Shakir
Koehn, Marc No Gordon point Informatics
Hufnagel, Steve Yes U.S. Department of Defense, Military Health System
Laakso, Lynn No Health Level Seven
McGaughey, Skip No
Peres, Greg No CA
Robertson, Scott No Kaiser Permanente</td?
Smith, Karen No HL7 Technical
Thompson, Cliff Yes OntoSolutions
Wrightson, Ann Yes HL7

Call to order

The meeting was called to order at 9:19 am UTC -3 by Chair Ron Parker with Tony Julian as Scribe.

Approval of Agenda for quarter

Agenda approved by affirmation

Approval of agenda for WGM

Agenda approved by affirmation

High-level review of peer review

  • Work item - need to pass through - what is the provenance of the information? RMODP?
  • Someone who is familiar with RMODP - Wendell Ocassio?
  • Discussion centered around the usage of the SAIF documents, as well as how to proceed.
  • The implementation guide will have to reset the original material for reference.
  • One of the challenges is to validate dispostion with ARB.
  • Continued to discuss comments.


The meeting was adjourned at 10:30 am UTC-3.

Sunday Q2

Call to order

The meeting was called to order at 11:00 am UTC-3 by Chair Ron Parker with Tony Julian as Scribe.

Continued to discuss comments.


The meeting was adjourned at 12:30 pm UTC-3.

Sunday Q3


Name PresentAffiliationE-mail address
Bond,Andy Yes
Curry, Jane Yes Health Information
Grieve, Grahame No Kestral
Julian, Tony Yes Mayo
Koisch, John No Guidewire Architecture
Loyd, Patrick No Gordon point Informatics LTD.
Lynch, Cecil Yes ontoreason
Mead, Charlie Yes National Cancer
Nelson, Dale No
Ocasio, Wendell No Agilex
Parker, Ron Yes CA
Quinn, John No Health Level Seven,
Shakir, Abdul-Malik Yes Shakir
Dickinson, Gary Yes CentriHealthCentriHealth
Howard, Andrew Yes
Hufnagel, Steve Yes U.S. Department of Defense, Military Health System
Mulrooney, Galen Yes U.S. Department of Veterans Affairs
Thompson, Cliff Yes OntoSolutions

Call to order

The meeting was called to order at 1:30pm UTC-3 by Chair Ron Parker with Tony Julian as Scribe.

Approval of Agenda for quarter

Agenda approved by affirmation.

Tooling WGM information re OHT information Jane Curry

Discussion centered around the OHT Architecture Value Proposal Survey

Ron Reviewed the Agenda for TSC: There was a prolonged discussion on quality plan. Austin will be crafting the plan.

There was a lot of talk about scalability of ONC work in US. Anticipating a lot of new folks, with new perspectives.

Charlie Mead: There are a number in NCI who dont understand SAIF, HL7 or both.

Ron Parker: HL7 is ratcheting up the visibility. In Q3/Q4 we need to prep for the next WG, what tutorials, do outreach, or prep for it. There are a lot of people who are not comfortable commenting on it. We need strategy. I discussed the challenge on how the US as a realm resolved it's context issues, without US agenda taking over the meeting. Canada and NHS have meetings out of cycle with WGM's. This may changed the look/feel of the International WGM.

Andy Bond: Puts more emphasis on role of SAIF.

Ron Parker: There were issues of TSC Communications plan. They touched lightly on references to ARB and timeline. John Quinn's effectiveness is rated on "Companion Table of CTO and TSC Objectives". We need to talk to John about our dates. Needs to be refreshed. I made some assertions on status, and the roller-coaster of capacity. The only concensus was that it is mission critical. We(HL7) need project manager. I mentioned tooling when asked "what happens when SAIF is done?" - we have to operationalize it. We need a place for artifacts, etc. We have not yet determined the requirements. We have a challenge - SAIF provides the backplane. TSC discussed peoples reluctance to record milestones. We need to develop an enterprise document for HL7.

Charlie Mead: I thought the alpha's would solve that.

Ron Parker: Steve Hufnagle's document found a lot of disparaties.

Charlie Mead: ECCF is not an EA.

Ron Parker: People want a target architecture. People feel that the artifacts in the SAIF need to be in their plan. There are assertions we make that lead to the belief that they are well set.

Charlie Mead: That is the difference between the specification and the implementation guide. You need intellectual framework for what you are doing. We say that the artifacts developed for messages and services will use the SAIF.

Ron Parker: What is missing - people want to know where to go to get what, and a checklist of process.

Charlie Mead: You have conflated SAIF and EA.

Ron Parker: 7 years in to establishing a SOA architecture for healthcare I can now discuss it with end users. We need to discuss a ECCF implementation plan - with I&C, MnM, InM, security, etc. so that all parties are represented. Charlie McKay said "I need that plan - what are the dependent events, and how do we actualize it. Then we need to move the work to those who care about it." We need a lot more hands on it. Q3/Q4 thursday we need to discuss how to do it. John Quinn needs a plan he can communicate.

Charlie Mead: NCI counts on HL7 to maintain the SAIF guide. If NCI finds problems, they need a version to go back to, for harmonization.

Ron Parker: We need to be talking about an EA for HL7.

Andy Bond: ArB can create SAIF, but not an HL7 EA.

Ron Parker: Implementation of SAIF inside and out needs to have a study of the issues

  • where there are issues
  • where there are no issues
  • publishing

and return to the organization. The organization needs to change accountability and organizationally to accomadate.

Charlie Mead: At NCI we have had to understand that coherently defining the business horizontaly.

Ron Parker: We need to architect for change, and change the nature of management.

Cecil Lynch: That is why the alpha's dont meet it.

Ron Parker: I need to tell Charlie McKay that.

Charlie Mead: quoted a colleague" What happens here is what happened in physics years ago, requires changing from individual to team". We need to understand the business before we define the architecture. SAIF is the framework.

Ron Parker: The comments/documents will help us.

Charlie Mead: Lloyds ECCF comments were "this looks ok, would like to see it work" His comments about the BF "I am not sure it will work". NCI's architect defines BF as a fuzzy cloud.

Ron Parker: The templates of how Governance controls the service contract is important.

Jane Curry: We(canada) have a conformance profile, that defines what message for what business. We need to give a business model that is more user friendly. Some CF are mandatory, some are optional e.g. electronic order cancel is not mandatory.

Cecil Lynch: We need governance and leadership. To realize the contract you need someone who agrees to support it.

Ron Parker: We are still in build-and-deploy mode. Each provence has an EA. We have used V3 messages. As an organization brings in more vendors, they will discover that the contracts are inadequate. You need a established governance model that provides a way to modify the contracts to support the new requirements. Have to have shared governance.

Andy Bond: You need the contracts at an enterprise level.

Ron Parker:Yes

Andy Bond: It also creates a lopsided view.

Ron Parker: Thursday Q3/Q4 we need to throttle, discuss our place in the organization, and help HL7 understand the business architecture. I would like to outline a lightweight framework, then work on business architecture. I propose that we need to resurect jane's database that defines the correlations between items to manage, both policies and procedures. It is a starting point. We need to set reasonable expections on support and shaping of SAIF.

Charlie Mead: At NCI the notion that we were doing SOA started off alarms. People coming to HL7 are not business architects.

Ron Parker: The business of HL7 is to author and deliver specifications.

Jane Curry: We can ask members and customers to send the right people.

Ron Parker: If our product strategy works well a new approach will evolve. If we knock up the framework the expertise will become evident.

Galen Mulrooney: SOA projects that fail - the IT department was teh instegators. The successful ones were driven by business. NCI and others are doing a lot of work - could those people not provide the architecture for Soa/HSSP? It might mean sending diffent buns.

Cecil Lynch: Governance must provide leadership for the horizontal implementation of the information framework.

The service contract provides the binding of the information and behavioral frameworks in the light of a section of governance (the services policies). This is not adequate alone to describe the whole of working semantic interoperability, but is a necessary component. The traceability of the services contract must be followed to the persistence layer as well so that the reproduction of data from a persistence layer to surface data for exchange must be lossless in order to meet the service contract requirements in the context of how the data was captured. This “horizontal” informational viewpoint is the fundamental component to give the whole of working semantic interoperability.

Jane Curry: The Board made a commitment to SAIF governance.

Steve Hufnagle: They have made a promise without understanding what governance is.

Andy Bond: I get worried when we move from the business of HL7 to that of the healtcare community. We have to stop merging the conversations.

Ron Parker: There will be handholding necessary when we move the business to those responsibility.

Charelie Mead: Framing question: DO we think that the alpha project efforts will take us anywhere - if not, stop.

Ron Parker: We need to work with them to use their input to refine the 'model'.

Galen Mulrooney: Use the SAIF alphas as educational?

Cecil Lynch: Accomplished communication and education.

Ron Parker: I am not sure we will be able to cycle things, and get enough heavy lifting done.

Charlie Mead: Operational perspective: NCI is moving forward - started an alpha project at the same time as HL7. Why is it succeeding? At NCI it came top down (use it, or leave!). HL7 will be no different. HL7 has fear of governance.

Ron Parker: That was the conversation at the TSC.

Jane Curry: Do we use the big hammer?

Cecil Lynch: That is what works for NCI.

Ron Parker: This requires investment in time and project management.

Charlie Mead: NCI had governance top down.

Jane Curry: The governance is engaged.

Andy Bond: SAIF is not a development platform, it is a governance(conformance) document.

Ron Parker: I would like to frame the questions for the Alphas - where are you, where are you going, what do you need to get there?

Charlie Mead: One of the troubles at NCI is - For the first 5 years at CABIG we had compatibility guidelines. Only required definition at the PSM level. The reason we are not getting the interoperability is because at the CIM/PIM levels are not represented.

Steve Hufnagle: We had to use EHR-FM and HITSP to populate the cells. The BF was the most specific on the artifacts. After coming up with the table, we understood that if we did not have EHR-FM and HITSP we could not have filled in the table.

Charlie Mead: Without top down, we would have not gotten here.

Steve Hufnagle: We found granularity issues in DOD and HITSP specs. Wireframes assumes the architecture comes from the business.

The meeting was adjourned(with exhaustion) at 3:00pm UTC - 3.

Updated:Tony Julian 14:15, 27 May 2010 (UTC)

Thursday Q3


Name PresentAffiliationE-mail address
Bond,Andy Yes
Curry, Jane Yes Health Information
Grieve, Grahame No Kestral
Julian, Tony Yes Mayo
Koisch, John No Guidewire Architecture
Loyd, Patrick No Gordon point Informatics LTD.
Lynch, Cecil Yes ontoreason
Mead, Charlie Yes National Cancer
Nelson, Dale No
Ocasio, Wendell No Agilex
Parker, Ron Yes CA
Quinn, John No Health Level Seven,
Shakir, Abdul-Malik No Shakir
Dickinson, Gary Yes CentriHealthCentriHealth
Howard, Andrew No
Hufnagel, Steve No U.S. Department of Defense, Military Health System
Koehn, Marc No Gordon point Informatics
Laakso, Lynn No Health Level Seven
McGaughey, Skip No ???
Mulrooney, Galen No U.S. Department of Veterans Affairs
Peres, Greg No CA
Robertson, Scott No Kaiser Permanente</td?
Smith, Karen Yes HL7 Technical
Steinen, Michael Yes HL7 Technical
Thompson, Cliff Nos OntoSolutions
Wrightson, Ann Yes HL7

Call to order

The meeting was called to order at 1:30pm UTC-3 by Chair Ron Parker with Tony Julian as scribe.

Approval of Agenda for quarter

Add to agenda I&C meeting Agenda approved by affirmation.


In conversations with Ron and John Q and Charlie McKay, it was felt that the more committes involved, the better off we will be.  Logical committee for the IG would be the I&C.  Met with I&C - they agreed, I&C will take the responsibility of taking on the IG - we will have to help, but in the end they will have responsibility.

Jane Curry: There is a chapter that describes localizations.

Charlie Mead: With help they can take over IG.

Ron Parker: Question of Quality or process?

Charlie Mead: Quality. Frank said that it the outcome was testable conformance, he was for it.

Cecil Lynch: Spent the rest of the meeting mulling over next steps. NCI will have first cut IG next week.

Ron Parker: Close on peer review, and how do we go forward, and what is the strategy on the IF and IG. We need far better sense of artifacts from the Alphas to add to the guide.

Cecil Lynch: What are you taking about?

Ron Parker: you cannot create an IG without understanding of the process. There is a sense in the community that engagement will occur. IG needs a better understanding of the architecture. We cannot conclude IG until we cycle the business architecture.

Jane Curry: We say the GOM reflected through the group using the HDF as the source of the current artifacts. This is our formal process in the GOM. If that is not true, it gives us a picture we can edit and strengthen.

Cecil Lynch: IF has requirement for HL7 side to have a piece that relates to the HDF processes, and walk back to core principles - mapping these things to the IF for HL7 specific view. IG has to have examples of the business processes.

Jane Curry: When will the spec writers have to deal, US and affiliates, vendors.

Ron Parker: Initial IG must be for HL7 International.

Cecil Lynch: Tie it into the RIMBAA matrix. If they have the static semantic artifacts, it will map. If you see the RIMBAA demonstrations it is in there, but you have to see the demonstrations.

Ann Wrightson: HOw does RIMBA fit?

Cecil Lynch: Yesterday i was doing demo in RIMBAA with OWL. One of the things that came out Berndt loved the model. Model generates pure V3 artifacts of static semantics. People looking at different ways of doing things is educational.

Ann Wrightson: I am sure that an artifact coming from this is in danger of making assumptions that are not viable - or will push people in that direction.

Cecil Lynch: There are WG's who will not use it all, but there are those who are building the total into their infrastructure. It does not describe most of the world. You will have to pick things at a higher level.

Ann Wrightson: YOu must start with two black boxs, and what you SAIF is what it takes for the black boxes to interoperate. The people implementing the contract need to worry about the contents of the box, but the contract should be ouside of that.

Cecil Lynch: This is why we have the whole CIM/PIM/PSM stack. Eithere you are specific at the PSM level, or you pick some level of abstraction.

Jane Curry: Does that mean we have to limit the CIM level to the minimum capabilities?

Andy Bond: No, instead of the conceptual concept, it may have a different granualarity of detail

Ron Parker: We are going far down the implementation. This(RIMBAA) is just one input (IG, harmonization) - we will need to discuss the nuances.

Joint with SOA

TSC ALpha Peer review comments Community is rallying to the challenges of ARRA stimulus

Slide 2

  • Complete the work at hand - SAIF book
    • Dispostion peer-review comments
      • ECCF is relatively OK
      • BF needs some work in adjusting language, and a thorough edit for style and graphics.
        • Need to remove the HL7ish parts to the IG -
        • remove the HL7 artifact references.
      • GF need to close on this - push it out.
      • IF: cecil has been comissioned to do draft in a condensed time(two weeks).
        • Challenge will be to de-NCI it.
      • IG need to develop a strategy that engages the organization
        • Two IG's - HL7 International - reference guide - maybe a template. NCI is emerging 3 layers to the IG.

1. Adopters: to the person who wants to use it 2. Adapters:Those who want to take a different approach (technology, vehicle) 3. Builders Core of NCI is cell-by-cell discription of artifacts.

    • Review the consolidate comments
    • Categorize the comments and consolidate changes
    • ECCF and INTRO corrections doable
    • BF needs work, perhaps some more positionaing language in overview
  • Consolidate back into DITA

Jane Curry: We have 3 licenses for DITA authorizing tools. It has content management at the enterprise management level. XMETAL and OXYGEN.

Ann Wrightson: Instead of DITA tools you need content management.

Ann Wrightson: OXYGEN is the better tool.

Jane Curry: Yes, OXYGEN has the more current version of DITA.

Ann Wrightson: There is a freeware content manager for OXYGEN.

  • Collaborate with SOA on CTS2
  • Plan /initiate survey of Interoperability capabilites" enabled by HL7 SPecs.
  • EA to support the business of HL7 International

Jane Curry: TSC owns this - are we going to surface things that belong. HL7 Business drives the architecture. HL7 International is not just the meeting attendees. Use "HL7 Enterprise Architecture". We need this soon, to avoid confusion.

 ACtION: Need excercise to clearly define disambiguate and label the business, internal, external, and enterprise architecture

What is the definition of the architecture.

Ann Wrightson: What is the definition of the architecture. Ask affilitates to describe healthcare architecture.

Ron Parker: Create a pattern or an artifact to help them answer the questions. We will achieve discovery, like we need to do internally. There is an issue that we dont know what we are working with. Asking the affilitates will be invaluable at a product strategy level.


The meeting was adjourned at 3:00pm UTC-3.

Thursday Q4


Name PresentAffiliationE-mail address
Bond,Andy Yes
Curry, Jane Yes Health Information
Grieve, Grahame No Kestral
Julian, Tony Yes Mayo
Koisch, John Yes Guidewire Architecture
Loyd, Patrick Yes Gordon point Informatics LTD.
Lynch, Cecil Yes ontoreason
Mead, Charlie No National Cancer
Nelson, Dale No
Ocasio, Wendell No Agilex
Parker, Ron Yes CA
Quinn, John No Health Level Seven,
Shakir, Abdul-Malik No Shakir
Chai, Victor Yes Singapore,
Dickinson, Gary No
Howard, Andrew Yes
Hufnagel, Steve No U.S. Department of Defense, Military Health System
Koehn, Marc No Gordon point Informatics
Laakso, Lynn No Health Level Seven
McGaughey, Skip No ???
McKenzie, Lloyd No HL7
Mulrooney, Galen No U.S. Department of Veterans Affairs
Peres, Greg No CA
Robertson, Scott No Kaiser Permanente</td?
Smith, Karen No HL7 Technical
Steinen, Michael Yes HL7 Technical
Thompson, Cliff Yes OntoSolutions
Wrightson, Ann Yes HL7

Call to order

The meeting was called to order at 3:30pm UTC-3 by Chair Ron Parker with Tony Julian as Scribe.

Go back to PIC with the template and the utility.

  • EA to support the business of HL7 International Pt. 2
  • Pull key principles from SAIF
  • Draw requirements from
    • Project services,
    • quality Initiative
  • product strategy

Not complete unitl the IG is done, however we will iterate.

The organization is not resourced to do the work, and the volunteers dont have the bandwidth. There is no understanding or the ROle of ARB in this process.

  • Prep for the fall meeting:
    • Key message is where on the continuum of start-to-finish the SAIF is.
    • Need to target SAIF at different audiences.
    • TSC need strategy to address

Arb Members should start to attend thursday round-table at Working Group Meetings.

Skype thread:

[11:46:18 AM] Ron G Parker: Because we have more than one "call in" participant, please use the following number:

[11:46:47 AM] Ron G Parker: 1-866-626-0833 x8320876103#

[11:47:01 AM] Ron G Parker: I will be dialing in via Skype so I can't guarantee call quality.

[11:47:27 AM] John Koisch: Hi Ron ... is there an agenda?

[11:50:33 AM] Jane Curry:

[11:52:43 AM] Cliff Thompson: Hi Karen

[11:55:27 AM] John Koisch: Hi all ... except when Ron is talking, the speaker is incredibly muffled

[11:55:43 AM] Karen Smith: The relay operator is unable to hear the speakers

[11:55:49 AM] Andy Bond: will keep up a running commentary for you

[11:55:57 AM] Karen Smith: ok thanks appreciate it

[11:56:56 AM] Karen Smith: operator can hear Ron but not the others

[11:57:12 AM] Andy Bond: I&C meeting with ARB/Charlie resulted in agreement to move forward

[11:57:35 AM] Andy Bond: need help but not unwilling

[11:58:36 AM] Andy Bond: questions at I&C about reconciling v2 vs v3

[11:59:56 AM] Andy Bond: need to extend I&C expectation to multilayered conformance rather than just through implementation guide

[12:02:50 PM] Andy Bond: they are waiting on nci eccf implementation guide to get a better idea of what needs done

[12:03:28 PM] Karen Smith: will we also have a hl7 eccf implementation guide?

[12:04:03 PM] Andy Bond: yes, nci one will just provide a sample for I&C to get better idea of what will be needed

[12:09:11 PM] Karen Smith: what are they talking about now?

[12:12:17 PM] John Koisch: Andy ... did you get tired?

[12:13:05 PM] Andy Bond: sorry ... managed to break firefox and things went awry

[12:13:30 PM] Grahame Grieve: shouldn't use a mac, see

[12:14:28 PM] Andy Bond: talking about info framework

[12:14:28 PM] Andy Bond: cecil talking about rimbaa and how it relates to saif

[12:14:49 PM] John Koisch: we will be publishing a paper on RIMBAA saif interaction soon

[12:14:57 PM] Andy Bond: nah nah nah ... shouldn't install vpn software on the fly :)

[12:15:44 PM] Andy Bond: ann now shouting :)

[12:15:53 PM] John Koisch: you guys need a podium and everyone takes a turn at the Mic

[12:16:05 PM] Andy Bond: we have one but its ron's laptop

[12:16:22 PM] John Koisch: self-made podium!

[12:18:41 PM] Andy Bond: you guys hearing ron/cecil?

[12:18:51 PM] John Koisch: y

[12:27:43 PM] Karen Smith: Regarding the BF, the EA diagrams have been converted to SVG, so they should be more readable. If the font is still too small, let me know and i'll increase the font size.

[12:28:17 PM] Andy Bond: will pass that on ...

[12:28:17 PM] Karen Smith: I left some of the BF terms with initial caps because I wasn't sure if they were keywords or not. can someone clarify?

[12:28:23 PM] Karen Smith: ok thanks

[12:28:37 PM] Karen Smith: I'll go thru the BF to ensure the style is consistent

[12:28:48 PM] Karen Smith: but it'll be up to you guys to clarify the language, I guess

[12:31:13 PM] Jane Curry: The glossary need was expressed in the peer review

[12:32:13 PM] John Koisch: A glossary has been provided for the BF

[12:32:36 PM] Andy Bond: came through in ECCF too

[12:32:51 PM] John Koisch: y

[12:33:14 PM] Andy Bond: 71 inconsistency problems in ECCF from feedback

[12:33:28 PM] John Koisch: that's consistent with my take

[12:33:40 PM] Andy Bond: 123 instances of same in intro

[12:34:01 PM] Karen Smith: were there also inconsistency issues in the BF?

[12:34:30 PM] Andy Bond: oh that had far more exciting stuff ... inconsistencies were the least of the coments :)

[12:37:35 PM] John Koisch: y .... lloyd had a bunch of really fundamental issues

[12:37:41 PM] John Koisch: not sure how to respond to some of those

[12:38:01 PM] Andy Bond: SAIF as a HL7 neutral doc and HL7isms in implementation guide

[12:38:49 PM] John Koisch: so then HL7 v3 is a platform .... ? I would be happy with that characterization.

[12:39:50 PM] John Koisch: but I suspect others won't be

[12:40:19 PM] Andy Bond: we had an interesting chat with I&C about v3 and v2 as PIM's

[12:40:28 PM] Andy Bond: rather than platforms

[12:41:22 PM] John Koisch: we have spent a lot of time on v3 at the NCI .... pulling it into constituent parts so that we can have RIM-based parameters for Service calls

[12:42:35 PM] Andy Bond: if RIM is a platform then would still need a model for information PIM

[12:44:04 PM] John Koisch: RIM is not a platform ... it is a set of choices about how to express information

[12:44:10 PM] John Koisch: and v3 is not RIM

[12:44:13 PM] John Koisch: at least in our world

[12:45:18 PM] Andy Bond: yes meant in info viewpoint, it would be RIM as a PIM and then creation of a V2IM as a PIM ... I should have been more specific

[12:45:46 PM] Karen Smith: would RIM also apply to the CIM and PSM levels?

[12:46:41 PM] John Koisch: there is some grey area in the way that you use the specificaiton levels. We choose to expr3ess our PI Specification Information Models as R-MIMs

[12:47:09 PM] John Koisch: and, in fact, we have two levels of R-MIMs ... both expressed in the PI Specificaiton level.

[12:47:11 PM] Andy Bond: what formalism do you use at the CIM for IM?

[12:47:39 PM] John Koisch: a domain model (uml class diagram) that errs on the side of expressivity

[12:47:43 PM] Patrick E Loyd: DAM?

[12:47:56 PM] John Koisch: well, a DAM includes behavior

[12:48:02 PM] John Koisch: so ... not specifically

[12:48:51 PM] John Koisch: I think we are calling it a DIM ...

[12:49:09 PM] Cecil Lynch: You are right Andy. I think that whatever we call these things (Concepts, datatypes, objects, reusable metaobjects, domain metaobjects etc, we will still have to have all of the things we use in HL7 to package data and information. The service side of this is more challenging but Thomas Erl is a good source to throw names at it that everyone will agree have nothing to do with HL7.

[12:49:33 PM] Andy Bond: is there a letter we don't use in front of IM? :)

[12:49:43 PM] John Koisch: here is what we find ... some concepts at the CIM level of models express as a collection of classes in an RMIM and some single cloned classes in an RMIM may express multiple concepts

[12:49:59 PM] Karen Smith: how about ZIM?

[12:50:02 PM] Karen Smith: JIM? haha

[12:50:11 PM] Patrick E Loyd: LOL

[12:50:14 PM] John Koisch: heh

[12:50:23 PM] John Koisch: we will create a ZIM asap

[12:50:47 PM] Karen Smith: sure

[12:56:11 PM] Karen Smith: what content management software were they discussing? The operator didn't catch the name

[12:56:17 PM] Andy Bond: oxygen

[12:56:31 PM] Andy Bond: ann w recommended it over the other one that I didn't catch

[12:56:51 PM] Karen Smith: yes, I'm using a trial version of oxygen and I like it a lot. Reasonable price and keeps up with the latest DITA standards.

[12:57:07 PM] Andy Bond: also talked about needing a content manager too

[12:57:19 PM] Karen Smith: my trial license runs out in 10 days :(

[12:57:36 PM] Karen Smith: yes, a content manager also will help manage the large volume of dita files, graphics, and aid in searches.

[12:57:47 PM] Karen Smith: oxygen can be integrate dwith a content manager or database

[12:59:45 PM] Jane Curry: It sounds like Oxygen is winning the race - I also know that IHTSDO liked it. We would need to find, preferably open source freeware, a content manager

[12:59:59 PM] John Koisch: telcon cut out ... is it connectivity on your end Ron?

[1:00:04 PM] Karen Smith: I plan to start evaluating XMetal tomorrow for comparison purposes. Oxgen already supports many of the new DITA 1.2 elements although the official standard hasn't been released yet.

[1:00:34 PM] Karen Smith: yes, I am leaning towards oxygen. Their tech support is great.

[1:00:51 PM] Andy Bond: someone must have something negative :)

[1:01:11 PM] Andy Bond: ron just said we're off for break and back in 30 mins

[1:01:12 PM] Karen Smith: well, oxygen doesn't have the most intuitive interface, so it takes a little longer to learn.

[1:01:44 PM] John Koisch: heh

[1:01:45 PM] Karen Smith: let me know when you're back online. I'm ready for a break! Thanks all.

[1:01:46 PM] Jane Curry: I also think the integrated content manager would be nice

[1:02:14 PM] Jane Curry: Karen, when you come back - we could just do chat and let you release your telephone person

[1:02:33 PM] Andy Bond: we're looking for one at nehta too so will look at it

[1:04:53 PM] Karen Smith: let me know which content manager you recpommend?

[1:05:06 PM] Karen Smith: Jane, I'm off the video relay for the break.

[1:05:44 PM] Andy Bond: ann w was talking about a good free one ... will ask which one she was talking about

[1:06:16 PM] Karen Smith: that would be great, thanks

[1:12:46 PM] Karen Smith: Oxygen works with the documentum content manager (but that is $$$$)

[1:13:18 PM] Jane Curry: How much?

[1:13:57 PM] Karen Smith: I don't know, will need to investigate. Many tech writing groups use documentum.

[1:14:22 PM] Jane Curry: We need something that can work with more than one project

[1:14:28 PM] Karen Smith: Oxygen works with several types of databases - IBM DB2, JDBC-ODBC, Microsoft sQL server, Oracle, for example.

[1:14:35 PM] Andy Bond: we have new technology for a mic

[1:15:31 PM] Karen Smith: I've heard of Alfresco, an open source content manager, but I'm not familiar with it. I will have to check with Oxygen tech support for a good open source content manager that would be compatible and not too terribly hard to configure.

[1:19:52 PM] Karen Smith: has the mtg started back up yet?

[1:20:04 PM] Jane Curry: Not quite

[1:36:53 PM] Jane Curry: Ron's back - starting

[1:38:31 PM] Cliff Thompson: are we using the same phone #

[1:38:43 PM] Ron G Parker: Back online on the conference call

[1:39:50 PM] Ron G Parker: Call in number is:

[1:40:11 PM] Ron G Parker: 1-866-626-0833 x8320876103#

[1:40:38 PM] Cliff Thompson: thanks

[1:43:27 PM] Andy Bond: you all need to laugh at ron's jokes :)

[1:46:16 PM] Cecil Lynch: I usually just laugh at Ron.

[1:46:25 PM] Cecil Lynch: :D

[1:53:07 PM] Andy Bond: ron is walking the mic around when he remembers

[1:54:12 PM] Cecil Lynch: In contarst to laughing at Ron's jokes., please don't laugh at Ron's memory. :)

[2:04:29 PM] John Koisch: cecil does not think funny thoughts

[2:06:27 PM] John Koisch: all stages

[2:08:23 PM] John Koisch: my hand is up

[2:08:29 PM] Andy Bond: ok dokee

[2:09:20 PM] Andy Bond: you're next

[2:09:58 PM] Andy Bond: almost next

[2:10:03 PM] John Koisch: yeah, I guess

[2:10:14 PM] John Koisch: I have some history that is pertinent here

[2:10:30 PM] Andy Bond: very good ...

[2:14:22 PM] Andy Bond: graham in flight home

[2:14:37 PM] John Koisch: as usual, he is up in the air

[2:14:59 PM] John Koisch: this is governance issue

[2:15:01 PM] Andy Bond: but online ... maybe I spoke prematurely

[2:15:29 PM] John Koisch: interestingly, I show you (andy) off line right now

[2:15:45 PM] John Koisch: ooops ... back on line

[2:15:55 PM] Andy Bond: yes there is ... no point blaming someone if the governance isn't being used

[2:16:06 PM] Andy Bond: stealth mode

[2:16:10 PM] John Koisch: heh

[2:16:16 PM] John Koisch: stealth governance?

[2:16:51 PM] Andy Bond: stealth mode visibility of me ... governance has to be anything but stealth

[2:17:05 PM] John Koisch: i like the idea of stealth governance

[2:17:18 PM] Andy Bond: only if we get ninja uniforms

[2:17:25 PM] John Koisch: oh YES

[2:17:35 PM] John Koisch: ninja architects

[2:17:36 PM] Karen Smith: don't want a big brother! 1984

[2:18:08 PM] Karen Smith: what would ninja architects do -- fighjt for their design ideas *=(Ha)

[2:18:47 PM] Andy Bond: they would take out evil designers

[2:26:12 PM] Karen Smith: Is this EA for HL7 separate from SAIF?

[2:26:57 PM] Andy Bond: yes

[2:27:54 PM] Anthony(Tony) Julian: Karen: SAIF tells you how to build an EA.

[2:30:01 PM] Karen Smith: ok

[2:31:48 PM] Andy Bond: charlie dropped a doc on the ArB list ... this bit on twin architectures is useful ... 4.1 The “Two-Architecture” Problem

PhUSE 2009

Every organization – regardless of its business practices – has a “business architecture,” i.e. the set of people, places, systems, roles, relationships, responsibilities, accountabilities, processes, deliverables, and patterns-of-practice that collectively define the organization both internally – i.e. to its employees, managers, etc. – as well as externally – i.e. to its customers, suppliers, regulators, etc. Whether an organization’s business architecture is explicit or implicit is immaterial in the sense that an implicit business architecture is every bit as real as one defined explicitly. (The latter representation is, however, much easier to expand, modify, evolve, govern, or otherwise interact with than an architecture that is simply carried in people’s minds as “the way we do things here.”) HL7 and CDISC both have their own business architectures.

In addition, organizations like HL7 and CDISC that develop technical specifications that enable connectivity between IT systems both within a given organization as well as between organizations – i.e. whose business architectures are (hopefully) optimized to enable the efficient, effective, efficacious, and economically-viable production of “specifications” that enable interoperability – are faced with a “second architecture” with which they must be deeply – and ultimately explicitly – aware, i.e. the technical/enterprise architectures of the various enterprise/technical architectures that utilize the specifications to enable interoperability.

The necessity of being aware of both architectures, i.e. to have a business architecture that effectively addresses the needs of the enterprise/technical architectures supported by the products produced by the business architecture, is often referred to as the “Two-Architecture Problem” specifically because experience has shown that lack of awareness of this problem leads to business-architecture-based processes that do not effectively address the needs of the customers’ enterprise/technology architectures. Both HL7 and CDISC have experienced the difficulties that can arise when specifications are developed without sufficient awareness of the Two-Architecture Problem, and both organizations are now actively involved in reorganization efforts aimed at correcting these difficulties.

[2:33:48 PM] John Koisch: sure ... this is the basis for Lloyd's problems with the BF

[2:33:59 PM] Andy Bond: mmmmm

[2:34:00 PM] John Koisch: the BF is all about saying "how we do things"

[2:34:13 PM] Andy Bond: and all depends on who we is :)

[2:34:36 PM] John Koisch: well, HL7 ain't no "we"

[2:34:39 PM] John Koisch: it's an it

[2:34:57 PM] John Koisch: but that is to say that hl7 <> Lloyd

[2:35:21 PM] Andy Bond: so lloyd is not an it? :)

[2:35:24 PM] John Koisch: (or any person in particular)

[2:35:32 PM] John Koisch: nah ... Lloyd is not an It

[2:35:40 PM] John Koisch: but he IS IT

[2:35:48 PM] Andy Bond: now I get it

[2:35:52 PM] John Koisch: yeh

[2:39:28 PM] John Koisch: raising my hand

[2:39:49 PM] Jane Curry: And Lloyd is now in the room

[2:39:55 PM] John Koisch: sure

[2:39:58 PM] John Koisch: good

[2:40:46 PM] John Koisch: (wave)

[2:41:02 PM] Andy Bond: (flex)

[2:44:24 PM] Andy Bond: we don't need to complete the EA but at least agree what we need in one

[2:45:03 PM] John Koisch: well, that was the approach that we took with the saif, and we are now several years down the road on that

[2:45:53 PM] Andy Bond: that is an interesting discussion ... the internal architecture view and the external interoperability view

[2:46:34 PM] Andy Bond: applying saif in HL7 can't be done in isolation of understanding the processes it sits within

[2:46:41 PM] John Koisch: yeah ... we have always played the shell game of "the most rigorous / restrictive / difficult system requirements can be expressed as interoperability requirements"

[2:46:55 PM] Andy Bond: 'less is more

[2:47:28 PM] John Koisch: my #1 lesson from implementing saif for two years ... communities HATE frameworks. they want process

[2:47:47 PM] Karen Smith: and procedures (tasks)

[2:48:10 PM] Andy Bond: they're a left and right hand

[2:49:53 PM] Cecil Lynch: The problem is if the community does not adopt a framework, it is dificult to execute a consistent set of procedures. The bigger the enterprise, the worse the problem becomes.

[2:50:11 PM] John Koisch: yup ... it is the communication problem

[2:50:20 PM] Cecil Lynch: We are seeing that at NCI right now.

[2:50:25 PM] Karen Smith: sounds like the framework is similar to a style guide

[2:50:36 PM] John Koisch: as soon as you abandon the framework, you don't have the means to resolve issues or goals

[2:53:16 PM] John Koisch: i agree with this

[2:53:27 PM] Andy Bond: yeah it needs to be contextualised

[2:53:52 PM] John Koisch: odp ... we need it ... so that we can have the right discussions ... but that is not a reqt for people at large

[2:54:39 PM] Cecil Lynch: Isnt this what the stairway to heaven is saying?

[2:54:46 PM] John Koisch: people need to build "model x" that we know serves as "a community model" eg

[2:54:53 PM] John Koisch: in part

[2:54:56 PM] Andy Bond: but we need to get across the value ... don't have to call it odp ... the viewpoints don't make any sense in isolation

[2:55:03 PM] John Koisch: sth is about specificity ....

[2:55:28 PM] John Koisch: this is why steve hufnagel liked the BF ... it lays out the MDA aspects of how you compose specificaitons

[2:55:39 PM] Cecil Lynch: Well, I think that is what Ann is getting at, if I understand her correctly.

[2:55:46 PM] John Koisch: and it has ODP corrolaries and allegories

[2:55:47 PM] Andy Bond: cross viewpoint obligations don't come through

[2:56:00 PM] John Koisch: yeah ... the BF is all about cross viepoint

[2:56:06 PM] John Koisch: computation is compositional

[2:56:29 PM] Andy Bond: yeah a lot of people are equating BF and computational viewpoint

[2:57:05 PM] John Koisch: they are close ... but the BF is really more about how to compose across the viewpoints taking the computational vp as the apex of compoition

[2:57:08 PM] John Koisch: composition

[2:58:52 PM] Karen Smith: It sounds like significant portions of the BF will need to be rewritten based on the peer review comments. John and Ron -- I would like to provide you with the latest Word source of the BF to use as a base for rewriting stuff. Does that sound like a good plan?

[2:59:28 PM] Andy Bond: that sounds like a goo dplan

[2:59:55 PM] Karen Smith: great, I finished converting the BF to dita a couple weeks ago. I generated a Word output file which I can send to John for revising.

[3:00:20 PM] John Koisch: safe travels

[3:00:25 PM] Anthony(Tony) Julian: Meeting adjourned - thanks to all for attending.

[3:00:36 PM] Jane Curry: Bye guys

[3:01:01 PM] Andy Bond: later gator

[3:01:09 PM] Cliff Thompson: bye - have a safe trip back


The meeting was adjourned at 5:00pm UTC-3 Tony Julian 14:46, 27 May 2010 (UTC)