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

Difference between revisions of "20120513 arb face to face"

From HL7Wiki
Jump to navigation Jump to search
 
(21 intermediate revisions by the same user not shown)
Line 40: Line 40:
 
##Call to Order
 
##Call to Order
 
##Aproval of Quarters Agenda
 
##Aproval of Quarters Agenda
 +
##DAM - Kaiser Permanente Perspective - Peter Hendler
 
##Impact of Thomas Erl's presentation
 
##Impact of Thomas Erl's presentation
 
##Substantivity???  
 
##Substantivity???  
Line 382: Line 383:
 
##Subject to review after Austin sends out the latest revision.
 
##Subject to review after Austin sends out the latest revision.
 
#DAM discussion from TSC
 
#DAM discussion from TSC
−
##Charlie discussed the heritage of the current [[Domain_Analysis_Model_ArB | Domain Analysis Model ]]
+
##Charlie discussed the heritage of the current [[Domain_Analysis_Model_ArB | Domain Analysis Model ]].  This includes the addition of reference to conformance.
−
#Report from Architecture Project
+
##AbdulMalik: The state of normative tells people that the content will not change.
−
##
+
##Jane: We need some way to document the use case to people who were not part of the discussion.
−
#SAIF IG involvement.  
+
##Ron: For normative you need a degree of formedness that is well defined.  People want degree of predictability.  What is sufficient in a DAM to meet the burden?
−
#What is next for ArB
+
##Charlie: We can define what we consider to be a correct DAM - a quantifiable objective.  The other side is that people can have a lesser DAM that is useful, to be balloted informatively.  We have to not define the subjective metric.
 +
##AbdulMalik: I see this as a matrix.
 +
##Jane: We need to have the dynamics.
 +
##AbdulMalik: You would think it would have a scope.
 +
##Jane: Scope is formed by the purpose.
 +
##Bo: DAM is a conceptual model.  By nature it is always informative at best - and forcing a publishing model does not add value.
 +
##AbdulMalik:  Informative allows you to pass without taking into account comments.
 +
##BO: It goes to stability.  The publishing mechanism does not fit the model.
 +
##Charlie: I will take another crack at this.
 +
##Ron: Sufficiency of a dam is its ability to support the downstream product.
 +
##Jane: If we say you cannot introduce concepts that were not in your dam.
 +
##AbdulMalik: Shalls/Shoulds Requirements/Best practice: All classes shall have a name,and a definition. All classes should have a unique name - how far do you go?
 +
##Steve: Fully formed SAIF model.
 +
##Charlie: TSC would like us to name the artifacts - not the representation.
 +
##Jane: Helpful to have tools for consistency - but a word doc is ok.
 +
##Ron: Could I negative ballot the downstream work product if the dam is faulty?
 +
##Bo: It requires a waterfall.
 +
##Ron: Is the dam sufficient to scope the product?
 +
##Jane: Purpose is the important part.
 +
##Ron: Maybe the completeness is tested through the ballot process for the downstream product.  We should say this is what a dam is, this is how it is use, and this is the downstream result.
 +
##AbdulMalik: Implies that all of our specification are proceeded by a DAM.
 +
##Patrick: that is where we are going.
 +
##AbdulMalik: There needs a way to express traceability to the DAM.
 +
##Lorraine: We are getting into the IG space.
 +
##Ron: Governance has a role in checking the quality.
 +
##AbdulMalik: If we find a quality problem while developing the downstream product, do we have to reballot the DAM?
 +
##Jane: That is why you go DSTU.
 +
##Lorraine: that is why we have DSTU in the first place.
 +
##AbdulMalik: We started with our model, and looked to the BRIDGE model, then went to create message specification (RMIM-DMIM).  They had a different style in the British Model - no traceability.  We discovered importance of multiplicity in the MIM's that we did not expose in the DAM.  At implementation we had a lot of work to keep the DAM, IG, and final product in sync.
 +
##Ron: Am i overloading the DAM?
 +
##AbdulMalik: I am having problem seeing people go back to the DAM.  It is too much work.
 +
##Ron: The problem is repeatibility.  I have seen it at HL7.  Where interoperability requires coherence and repeatibility.  The likelyhood that you will have to reopen the bag is small.  I am prepared to back away from the position.  HL7 needs to get the DAM right.
 +
##Andy: If HL7 had the discipline of a single trajectory it would work well.  It is incredibly difficult to say ultimately what it means if they are not on a single trajectory. How do you recognize the intersection between trajectories?
 +
##Ron: Same problem I have faced on every team.  The most successful was where the rigor was resolved.
 +
##AbdulMalik: The immunization DAM will spawn many products with no cohesion.
 +
##Jane: People want the DAM to be normative to feel it is sufficiently mature to rely on for the downstream products.  If that is what we want to have, we have no choice.
 +
##AbdulMalik: There are some implications. 
 +
##Jane: We changed the MIF's to map more directly.
 +
##Andy: We have the ability to keep things out of the dam.  I would rather have something with less, rather than nothing with more.
 +
##AbdulMalki: I you dont restrict datatype you help the downstream.
 +
##Andy: The less restrictive the better.
 +
##Patrick: We chose to use conceptual data types.
 +
##Andy: I dont care how you get there as long at you pass these three points to get there.
 +
##Jane:I am on the fitness for purpose kick: The perceptual model defines the shared purpose.
 +
##Steve: You have described the OMG model approach.  You might want flexibility as you go down.
 +
##Ron: Traceability controls what you put in you DAM .  It has to provide a sense of context of the domain.  As you drive down you may find you have missed something.  For those who want to reuse the thing you made, redoing the DAM is good.  Charlie see if you can define at what is sufficient at that level.
 +
##Steve: Do we want to define the artifacts that go in the DAM?
 +
##Charlie: That is part of the SAIF-IG.
 +
##Steve: Do we need an HDF, and the groups use their SAIF-IG at different levels?
 +
##Charlie: That is what Lloyd was talking about - and IG for FHIR, V3, V2.
 +
##Steve: ArB needs to develop a higher level IG. Some groups want to do mind-maps, some want to do sequence diagrams.
 +
##Bo: you need an independent IG.
 +
##Steve: The IG got so deep into it they got  burnt out.  It was at such a detailed level, and was very restrictive.
 +
##Patrick: I see something at the level of the HDF which applies to the IG.
 +
##Steve: ArB should be doing a middle level document.
 +
##Lorraine: I thought the IG was that product.
 +
##Steve: Then everyone would do their own thing.
 +
##Andy: The CD is the model.  You have to choose V3/V2 - how do you describe dynamic process.  There are different ways to cover the matrix.  I am not sure introducing another level will help.
 +
##Ron: Conformity with the CD will be needed.  We may discover that we need to frame the CD differently, after we have tested the IG against the CD.
 +
#Adjournment at 3:00pm Pacific
 +
--[[User:Ajulian|Tony Julian]] 22:06, 13 May 2012 (UTC)
  
 
==Thursday Meeting Information==
 
==Thursday Meeting Information==
Line 471: Line 532:
  
  
−
==Minutes==
+
==Agenda==
 
#Q3
 
#Q3
 
##Call to Order
 
##Call to Order
 
##Aproval of Quarters Agenda
 
##Aproval of Quarters Agenda
 +
##DAM - Kaiser Permanente Perspective - Peter Hendler
 
##Impact of Thomas Erl's presentation
 
##Impact of Thomas Erl's presentation
 +
##Saif pgm - [http://healthlevelseven.projectinsight.net/3c29eb36-dbb6-42af-aa0a-6286a7395932 project 477 CDA R3] - frame an opinion on separating semantics from rendering
 +
##[http://healthlevelseven.projectinsight.net/d237cd97-1afe-4f48-bc49-5f7bd1cb7c16 Risk Assessment and governance for SAIF Architecture] - Assigned by the TSC.
 
##Substantivity???  
 
##Substantivity???  
 
##HL7 Business Architecture
 
##HL7 Business Architecture
 +
###What are the key things to make the decision CDA/V2/FHIR/V3?
 
#Q4
 
#Q4
 
##Maturity of HL7 and the Customer Base(Zoran/Andy)
 
##Maturity of HL7 and the Customer Base(Zoran/Andy)
Line 485: Line 550:
 
##Conference calls will not resume until 2nd week after WGM- May 31, 2012.  Will re-validate during WGM.
 
##Conference calls will not resume until 2nd week after WGM- May 31, 2012.  Will re-validate during WGM.
 
#Adjournment
 
#Adjournment
 +
==Minutes==
 +
Lorraine, Jane, Steve, Ron, Bo, Andy, Charlie, Anil, Patrick Loyd
 +
Peter Hendler, AbdulMalik Shakir, lloyd
 +
#q1
 +
##Call to Order
 +
##Aproval of Quarters Agenda
 +
##DAM - Kaiser Permanente Perspective - Peter Hendler
 +
###B4 the DAM it was clear what HL7 V3 meant - using model, vocabulary, Data types
 +
###many of the DAMS dont have an object model
 +
###It is said "Domain experts cant understand the RIM"
 +
###They can learn the difference  between a Noun and a Verb
 +
###Can you at least separate the nouns from the verbs
 +
###More complicated, but physicians can understand that a Person can be a Patient at Kaiser, the same at intermountain, but a Physician elsewhere.
 +
###Why cant you at least have them thing this box is a "thing", this box is an "action", this box is a "Role"
 +
###They should be able to understand the nouns and verbs
 +
###Cardiology DAM does not differentiate between boxes, and some are vocab objects.
 +
###It confuses things that boxes that are vocab as part of the model.
 +
###Not only are the boxes arbitrary, if you do a search you find no reference to RIM.  They should state in a paragraph that there is an intention to grow up and become something that is HL7 V3 - and may become a document or a message.
 +
###To say Clinicians cant understand UML, they can understand Nouns and Verbs.
 +
###Suggesting a DAM style guide?
 +
###To say that a DAM describing the knowledge of a domain is not enough - tha tit should be relatively and consistently rendered.
 +
###We are anticipation that downstream artifacts have traceability to a DAM.
 +
###Maybe the RIM is super-complicated, but organizing your thoughts into Entities, Actions, and Vocabulary.
 +
###Imposing RIM grammer on a domain may not be logical- hard to differentiate roles to participations.  Purpose of a DAM is to get the thoughts on paper.
 +
###Not necessarily the RIM, but at least Nouns and Verbs.  That they are aware of the differences.  The ones we are reading are wrong.
 +
###No problem with the assertion ArB is to provide a definition of a DAM without specifying the contents.  Also looking at Governance and Risk assessment.  This would be the ideal place to explain goodness, and minimials.  There are dialogs on whether a DAM can be normative.  Ideal if DAMs were consistent.  Have not gotten to them yet.  There are tools that are useful in an iterative use.  DAM must identify the scope of the domain, or purpose.
 +
###If HL7 is going to vote on a DAM with no structure, it is wrong.
 +
###Normative DAMs will require expression as defeind by the ArB.
 +
###IF a DAM has only cardiology information, there is nothing to vote on.  Can only vote on things that apply to HL7.
 +
###What is an HL7 person supposed to vote on?
 +
###An HL7 DAM needs traceability to the subject.  Lab person would expect expression at a sufficient level that I would know what to do.
 +
###What is the nature of the persons who expressed the scope of interest.
 +
###We agreed that we have noting to say about clinical correctness.  Is the form sufficient?
 +
###To come up with an HL7 artifact - is it more trouble to get to V3, or to redo with a mappable.
 +
###Some would be in a DAM implementation guide.
 +
###A style guide would be useful - if I knew the form of expression and arrangement it would help.
 +
###Dont argue that Clinicians cannot do this.
 +
###In evaluating a DAM there are two:
 +
####Correctness
 +
####Completness from a definition standpoint.
 +
###Cant navigate an RMIM both ways.
 +
###We should be able to specify that Class names should be Nouns.
 +
###We are talking about a degree of expression.
 +
###Correctness we have as much exposure in RMIMs as we do DAMs
 +
###Nothing in the RMIM keeps you from expressing my dog as a caregiver.
 +
###HL7 people arent voting on the dogs participation, that the DAM is correct.
 +
###If it is unstructured, why are you voting?
 +
###There is some merit to traceability - you will talk about things in a aggregate way.  We have to provide lattitude to expression of usefullness.
 +
###Because we are talking about clinical domains, if there is reason to express it, we need to clarify the purpose, and declare the expectation that we can only vote on the structure - participation in your group should be broad enough to state that this is clinical information to support downsteam products.
 +
###In terms of clinical correcness, HL7 may never be able to vote on this.
 +
###HL7 can vote on the accurateness from a modeling standpoint, not the clinical decisions.
 +
###In certain disease whether the bone is broken or not is sufficient.  In Rheumatology, not so much.
 +
###Do we have an inventory of all dams?
 +
###Yes - the Board decided that DAMS would be given away for free.
 +
###Not all DAMs were gobledygook.
 +
###If it does not meet minimum quality, we should not give it away.
 +
###I do have the makings of a style guide.  It is not an HL7 style guide.  We have a guide in the HDF - for the RIM.  It would be useful to compare the DAM to the style guide.  Do we modify the style guide, or the model.  Using Classes for concepts might not be allowed.
 +
### we might find patterns of things people are doing that we have not expected.
 +
###Clinicians have expressed the concern over the DAMs in multiple settings this week.
 +
###Patrick is starting a listserv for DAM modeling
 +
###Wiki for DAM modeling.
 +
###Sparx is doing UML training.
 +
###Need to provide a non-clinical DAM expression.
 +
###NCI did a golf-club model, and got slammed because it was not relevant.  Need a clinical example everyone agrees on.
 +
###Might be useful to have a bad example.  Widely understood is a pharmacy order. I have product to be administered.  The DAM should never say I am prescribing peniciliin for a malady, rather a substance for a patient.
 +
###Some people mix up instance diagrams with the Model structure.
 +
###There is a lot of interest in stating nomenclature.  Mind maps and OWL are being used to understand the terms.
 +
###Isnt a DAM the first point of call?  Usually more about what should NOT be there, than what is
 +
#Impact of Thomas Erl's presentation (10 Minutes)
 +
##What were peoples comments?
 +
##Was it the right level for the audience?
 +
##Books are going to be useful - hope they get past seven pages.
 +
##The CD's are better.
 +
##It is clear that their is a depth of thinking and volume of work.
 +
##It was generous of him to provide the product.
 +
##Saif pgm - [http://healthlevelseven.projectinsight.net/3c29eb36-dbb6-42af-aa0a-6286a7395932 project 477 CDA R3] - frame an opinion on separating semantics from rendering#Q3
 +
###You should be expressing the semantics. We should take a page from Xform. 
 +
###The rendering should be separate.  Lloyd pushed back on it - Austin is in favor, Bob not so much.
 +
###The stated concern was that the committee did not understand it.
 +
###In the informative ballot there will be two different representations of an instance.
 +
###This has implications to the architecture.  If you cannot shift stuff around without chopping it up, there is a problem.  There is a risk from the community voting on something that is not good for HL7.  I recommended that the ArB that there should be a preference on the form.  It wont stop the community from keepig to status quo.  It will indicate there is no desire in the community for interoperability.
 +
###We are going to see more opportunity to allow the members to make an informed decision toward one architecture or another.
 +
###Is there a middle ground?  Can we let R3 finish its course, and let R4 take on the issue?
 +
###CDA R3 will die because the community is not enthused with R3 changes.  The ArB if it is going to be useful and relavant to the organization, should know what is going on, and make their comments known to the community where relevant.  Stepping into that role as an ArB is a good thing, and it is in this case appropriate.
 +
###Was well received by the Architure Program.
 +
###"Arb with Teeth" requires that we weigh in on the issue.
 +
###Need to determine process -ArB alone or through the TSC.
 +
###ArB's recommendation needs to be to the program.  I would understand where the powers are.  If you have the authority, do it, otherwise the TSC will have to weigh in.
 +
 +
##[http://healthlevelseven.projectinsight.net/d237cd97-1afe-4f48-bc49-5f7bd1cb7c16 Risk Assessment and governance for SAIF Architecture] - Assigned by the TSC.
 +
###Jane needs to update the PSS before proceeding.
 +
##Substantivity???
 +
##HL7 Business Architecture - discussed in Q4.
 +
##What are the key things to make the decision CDA/V2/FHIR/V3?
 +
###What form of expression would this take?
 +
###SWOT analysis - with decision tree. Informative artifact.
 +
###RIMBAA has posted a white paper on the analysis.
 +
###The audience is both public and internal.  Today I know I will use 2.5.1 for Immunizations, because the government says so.
 +
###Sometimes it is the function of the most mature specification at the point of decision time.
 +
###We can work out where to place it.
 +
 +
==Q4 Agenda==
 +
#Maturity of HL7 and the Customer Base(Zoran/Andy)
 +
#FHIR discussion
 +
#Architecture of templating
 +
#Other business and planning
 +
##Conference calls will not resume until 2nd week after WGM- May 31, 2012.  Will re-validate during WGM.
 +
#Adjournment
 +
 +
==Q4 Minutes==
 +
#Maturity of HL7 and the Customer Base(Zoran/Andy)
 +
##Common problem of driving from the bottom up vs the Top down.  We are defining capability measures for interoperability.
 +
##The organization can follow a path for maturity - which may not be relevant across every customer.
 +
##It is a way of framing the maturity path.
 +
##Andy will share the modeling that has been done.
 +
##Should feed to risk profiles
 +
#FHIR discussion
 +
##What would Arb like to discuss?
 +
##Gerald wants to understand how the ArB would see it in the context of an exchange paradigm around services.  Would like to see from and ArB perspective on how to fit FHIR to the parts that make up FHIR into the HL7 Architecture.  Specifically around the REST-ful aspects.  I think we are in agreement and REST that it is a realization of the RLUS functional model.  I can see some emerging concepts of the OO symantics. 
 +
##How we would do such a thing sounds like a project. 
 +
##Gerald: We had a mapping from RLUS platform independent and functional model to HDATA.  I would suggest that we go back and follow through on mapping along those lines, with SAIF implementation as an RLUS service.
 +
##Grahame: FHIR is not an RLUS model.
 +
##Gerald: There is an HL7 model realized by RLUS with OMG.
 +
##Grahame: I would be happy to look at consistency.  I am not comfortable with the expression.  The implication that FHIR will be an implementation of RLUS is not true.
 +
##Lloyd: FHIR is more than RLUS, and has more functionality.
 +
##Ron: If you could provide a dynamic RLUS/FHIR what would that look like?  Conceptually, how would we proceed? Who would drive?
 +
##Grahame: We will not be making FHIR more complicated to apply to RLUS.
 +
##Lee: I heard that you looked at the restful FHIR that maps to RLUS.
 +
##Gerald: It is.
 +
##Grahame: If you say "FHIR" is the RLUS model.  I think it does not.
 +
##Ron: Of interest to ArB: Risk assessment.  How are FHIR resources used in dynamic interchanges.  FHIR does not express how the resources are used. 
 +
##Grahame: We provide a lite messaging expressing.  SOA can define how to use resources.  Technical interoperability works, but there is a layer of business interoperabiligy that is not define in FHIR.
 +
##Gerald: Which is the heart of what pharmacy is doing around medication statements.
 +
##Grahame: Mapping is useful? Bind to conformance?
 +
##Gerald:I did not say conformance.  You can add functional performance on top of that.  FHIR does not address the administrative funtions.
 +
#Charlie: FHIR will be defined and balloted as a resource collection.  Resources will have profiles.  TSC asked ArB to provide a first cut at how HL7 could govern the resources, and possibly the extensions and profiles.
 +
#Charlie: Thomas Erl's material was the basis for the governance document presented.[http://gforge.hl7.org/gf/download/docmanfileversion/6812/9370/GovernanceOverview--20120517Rp.docx Governance Overview: Gforge Copy]
 +
 +
#Architecture of templating
 +
##Problem Statement:Assertion from Ron:  When we think about the problem I am anticipating, over time we are getting interfaces in place, trigger events are stabile, what will not be stabile, physicians start creating new requirements for the content.  The traditional ballot structure of  binding the context to the design
 +
##For a particular interface we template the content representation, and version them - and know that version changes would still be supportable.
 +
##I know the templating, and the mapping - if there is new clinical content, we would notify those who have implemented that there is new content.
 +
##Theoretically the metadata and trigger events are stable, so I would not have to re-test the dynamics of the interaction.
 +
##We need an architecture for template registries.  How do they get populated? Who populates them?  A Clinician just stated that we provide the transport, but not the final information.  We need tooling for those who will need many flavors, and change there minds often.
 +
##How to achieve stability and predictability - and support changes in content.
 +
##Templates business requirements answers some of the questions.
 +
##Template registry - the template service specification might be a way to start.
 +
##Coalescing the material that is flowing around the organization with the title template, and determine the inventory, and the congruency.
 +
##Goes to scalability, and all the other ilities you are in architecture.
 +
#Business Architecture (Q4)
 +
##Bo:It is hard to determine how HL7 operates.  There are processes, working groups, roles.  I would like to know if there is a way to represent the touch-points and the work-group interactions.
 +
#Charlie:People who succeed have a documented business architecture.  No one would contest the value.  Who would do it?
  
 +
#Other business and planning
 +
#Conference calls will not resume until 2nd week after WGM- May 31, 2012.
 +
##Baltimore:
 +
###Sunday Q1, Q2, Q3
 +
###Tuesday Q4
 +
###Thursday Q3,Q4
 +
#Adjournment  The meeting was adjourned at 4:45pm Pacific
 +
--[[User:Ajulian|Tony Julian]] 23:49, 17 May 2012 (UTC)                                                                                       
 
[[Category:Arb Minutes]]
 
[[Category:Arb Minutes]]

Latest revision as of 16:41, 4 May 2014

ARB - Meeting HL7 Spring Working Group Meeting

Sheraton Wall Centre Vancouver, BC, Canada

May 12-18, 2012

Agenda

Sunday

  1. Q1
    1. Call to order
    2. Roll Call
    3. Approval of Quarters Agenda
    4. Approval of WGM agenda
    5. Approval of Minutes
    6. Report from TSC (Charlie/Ron)
    7. Report to the TSC for Sunday Night TSC meeting
    8. Discuss Report to Co-Chair meeting Monday Night
    9. Review joint meetings to ensure coverage
    10. Adjournment
  2. Q2
    1. Call to Order
    2. Approval of Quarters Agenda
    3. Substantivity???
    4. SAIF-CD??? aftershocks
    5. Adjournment
  3. Q3
    1. Call to Order
    2. Approval of Quarters Agenda
    3. Strategic Initiatives
    4. DAM discussion from TSC
    5. Report from Architecture Project
    6. SAIF IG involvement.
    7. What is next for ArB
    8. Item 1
    9. Item2
    10. Adjournment

Thursday

  1. Q3
    1. Call to Order
    2. Aproval of Quarters Agenda
    3. DAM - Kaiser Permanente Perspective - Peter Hendler
    4. Impact of Thomas Erl's presentation
    5. Substantivity???
    6. HL7 Business Architecture
  2. Q4
    1. Maturity of HL7 and the Customer Base(Zoran/Andy)
    2. FHIR discussion
  3. Other business and planning
    1. Conference calls will not resume until 2nd week after WGM- May 31, 2012. Will re-validate during WGM.
  4. Adjournment

Meeting Information

HL7 ArB Work Group Meeting Minutes

Location: Sheraton Wall, Vancouver, BC Canada

Date: 20120513
Time: 9:00am Pacific
Facilitator Parker, Ron Note taker(s) Julian, Tony
Attendee Name Affiliation
X Bond,Andy NEHTA
X Constable, Lorraine Constable Consulting Inc.
X Curry, Jane Health Information Strategies
. Dagnall, Bo HP Enterprise Services
X Grieve, Grahame Health Intersections Pty Ltd
X Hufnagel, Steve U.S. Department of Defense, Military Health System
X Julian, Tony Mayo Clinic
X Loyd, Patrick ICode Solutions
. Lynch, Cecil Accenture
X Mead, Charlie National Cancer Institute
. Milosevic, Zoran NEHTA
. Ocasio, Wendell Agilex Technologies
X Parker, Ron CA Infoway
. Quinn, John Health Level Seven, Inc.
. Guests
X Luthra, Anil Health Alberta
X McKenzie, Lloyd HL7 Canada
X Stechishin,Andy CANA Software & Services Ltd.
. Legend
X Present
. Absent
Quorum Requirements Met: Yes

Minutes Q1

  • Report from TSC:
    • Charlie: Unresolved points from TSC: Austin and John Quinn both committed to notion that SAIF architecture, Fast Healthcare Interoperability Resources (FHIR), and strategic initiatives connecte via SAIF. There is no SAIF-IG for HL7. The leading resource is Lloyd McKenzie, who is working on FHIR.
    • ArB proposed to take over SAIF-IG from MnM. There was angst, concluded when Austin declared that TSC owns the SAIF-IG.
    • LLoyd: there should be a SAIF-IG for V3, V2, FHIR.
    • Charlie: Will start with a single SAIF-IG, with branch points. There will be a discussion Sunday night of the notion of FHIR governance vs FHIR management vs FHIR architecture.
    • Ron and Charlie discussed the TSC meeting. Helen Stevens brought up the vertical/vs horizontal in the SAIF-IG.
    • Ron: The metrics are pointing out the lack of consideration of HL7 priorities in the current projects.
  • Report to the TSC for Sunday Night TSC meeting
    • Grahame: FHIR SAIF-IG - I want the FHIR SAIF-IG to be the methodology guide, and I intend to write it. FHIR is only a partial methodology - not enventing anything new, just a different way to deploy existing methodology. The FHIR imformation model is borrowed from somewhere else.
    • FHIR proposes REST out of the box - which limits the dynamic model. FHIR puts together a flexible lightweight messaging model that matches those of version 2. If you cannot meet your needs by using REST or messaging, define a service yourself.
    • Jane: Documents?
    • Grahame: Yes, out of the box, but not necessarily CDA-R3.
    • Lloyd: you can have the charistics of a document that you can exchange using the same dynamic model as other content.
    • Jane: Same as V2 passing CDA?
    • Grahame: Actually better than MDM.
    • Lloyd:Documents in V2MDM are imbeded, FHIR documents are exposed.
    • Governance:
      • Grahame: FHIR is posted on HL7 servers, so is now HL7 IP. Until the first full normative version it will be freely available from the HL7 website at [| hl7.org/fhir].
      • Grahame: I propose that the TSC owns FHIR, and the FHIR governance board(FHIRGB) reports to the TSC. FHIRGB -Consists of persons who provide time commitements or money.
      • Ron: FHIR creates a defacto footprint, which wil allow other value propositions for the business of HL7.
      • Grahame: Derivative works are possible.
    • Ron: We need to discuss actioning. Charlie asserted that you cannot talk about governance unless you think of it from risk management.
    • Charlie: Do you have any sense as to how someone can claim conformance to the dynamic portion?
    • Steve: You can set the allowable states
    • Jane: Detail the roles - which are human, and which are automated at the conceptual level. If you dont, you have people guessing. Triggers in a more global sense, e.g. time conditions, something that triggers something to occur, with logical precedence (you cant cancel what you did not order).
    • Steve: You can define that a medication order has different behaviour from a lab order.
    • Ron: Patterns are emerging - elevating and layering. Properly dicomposing the services can be very powerful.
    • Jane: You can have a generic pattern that then forks with the complexity.
    • Ron: Told clinicians that a referal is an order fullfilment pattern - to which they objected.
    • Grahame: The clinicians argue about the requirements of the referal. It is appropriate to say that any specification that claims conformance to the DAM must explicitly express the conformace to the DAM.
    • Andy: This is emerging- in the end whatever you write down is a constrained pattern.
    • Ron: One of the measures of success will be external groups asserting their position on HL7.
  • Report to co-chairs Meeting Monday Night
    • Ron: Morning slides will discuss our topics
    • Ron: ArB is doing stuff - why do I care?
      • Jane: Strategic Initiatives(SI) are marching orders from Board to TSC. We need to discuss alignment with the SI.
      • Ron: Semantic web and the impact to HL7 - IBM is adding RDF to rational.
      • Jane: there may be alignment opportunities
      • Lloyd: I educated Eric on HL7, his questions need to be discussed with Grahame.
  • adjournment at 10:30am Pacific

--Tony Julian 17:34, 13 May 2012 (UTC)


Meeting Information - Q2

HL7 ArB Work Group Meeting Minutes

Location: Sheraton Wall, Vancouver, BC Canada

Date: 20120513
Time: 11:00am Pacific
Facilitator Parker, Ron Note taker(s) Julian, Tony
Attendee Name Affiliation
. Bond,Andy NEHTA
. Constable, Lorraine Constable Consulting Inc.
X Curry, Jane Health Information Strategies
. Dagnall, Bo HP Enterprise Services
. Grieve, Grahame Health Intersections Pty Ltd
. Hufnagel, Steve U.S. Department of Defense, Military Health System
X Julian, Tony Mayo Clinic
. Loyd, Patrick ICode Solutions
. Lynch, Cecil Accenture
. Mead, Charlie National Cancer Institute
. Milosevic, Zoran NEHTA
. Ocasio, Wendell Agilex Technologies
X Parker, Ron CA Infoway
. Quinn, John Health Level Seven, Inc.
. Guests
X Luthra, Anil Health Alberta
. Shakir, Abdul Malik Shakir Consulting
X Stechishin,Andy CANA Software & Services Ltd.
. Legend
X Present
. Absent
Quorum Requirements Met: Yes

Minutes Q2

  1. Call to Order
  2. Approval of Quarters Agenda
  3. Substantivity???
  4. SAIF-CD??? aftershocks
  5. Adjournment


Meeting Information - Q3

HL7 ArB Work Group Meeting Minutes

Location: Sheraton Wall, Vancouver, BC Canada

Date: 20120513
Time: 11:45pm Pacific
Facilitator Parker, Ron Note taker(s) Julian, Tony
Attendee Name Affiliation
X Bond,Andy NEHTA
X Constable, Lorraine Constable Consulting Inc.
X Curry, Jane Health Information Strategies
X Dagnall, Bo HP Enterprise Services
. Grieve, Grahame Health Intersections Pty Ltd
X Hufnagel, Steve U.S. Department of Defense, Military Health System
X Julian, Tony Mayo Clinic
X Loyd, Patrick ICode Solutions
. Lynch, Cecil Accenture
X Mead, Charlie National Cancer Institute
. Milosevic, Zoran NEHTA
. Ocasio, Wendell Agilex Technologies
X Parker, Ron CA Infoway
. Quinn, John Health Level Seven, Inc.
. Guests
X Luthra, Anil Health Alberta
X Shakir, Abdul Malik Shakir Consulting
. Stechishin,Andy CANA Software & Services Ltd.
. Legend
X Present
. Absent
Quorum Requirements Met: Yes

Minutes Q3

  1. Call to Order
  2. Approval of Quarters Agenda
  3. Substantivity
    1. Jane:consider substantivity as a risk mitigation strategy to validate the measure/definition.
    2. Ron: The V2 definition does not apply to V2, and the definitions do not work between artifacts.
    3. Charlie: You want to cast various flavors of substantivity of risk?
    4. Jane: Risk they mitigate. Then we can have semi-objective criteria.
    5. Charlie: We have not had success with TSC because they do not have the time. FHIR is going on record that once a resource is ballot it will not be changed - it may be replaced.
    6. Jane: With V3, since XML is primary representation, if you change anything you have violated its structure.
    7. Charlie: FHIR is changing from design by constraint to design by extension.
    8. Ron: Motivator is around usage - 80/20 rule.
    9. Jane: CDA has not changed for five(5) years.
    10. Ron: This will hit us like a ton of bricks on Thursday.
  4. Strategic Initiatives
    1. Subject to review after Austin sends out the latest revision.
  5. DAM discussion from TSC
    1. Charlie discussed the heritage of the current Domain Analysis Model . This includes the addition of reference to conformance.
    2. AbdulMalik: The state of normative tells people that the content will not change.
    3. Jane: We need some way to document the use case to people who were not part of the discussion.
    4. Ron: For normative you need a degree of formedness that is well defined. People want degree of predictability. What is sufficient in a DAM to meet the burden?
    5. Charlie: We can define what we consider to be a correct DAM - a quantifiable objective. The other side is that people can have a lesser DAM that is useful, to be balloted informatively. We have to not define the subjective metric.
    6. AbdulMalik: I see this as a matrix.
    7. Jane: We need to have the dynamics.
    8. AbdulMalik: You would think it would have a scope.
    9. Jane: Scope is formed by the purpose.
    10. Bo: DAM is a conceptual model. By nature it is always informative at best - and forcing a publishing model does not add value.
    11. AbdulMalik: Informative allows you to pass without taking into account comments.
    12. BO: It goes to stability. The publishing mechanism does not fit the model.
    13. Charlie: I will take another crack at this.
    14. Ron: Sufficiency of a dam is its ability to support the downstream product.
    15. Jane: If we say you cannot introduce concepts that were not in your dam.
    16. AbdulMalik: Shalls/Shoulds Requirements/Best practice: All classes shall have a name,and a definition. All classes should have a unique name - how far do you go?
    17. Steve: Fully formed SAIF model.
    18. Charlie: TSC would like us to name the artifacts - not the representation.
    19. Jane: Helpful to have tools for consistency - but a word doc is ok.
    20. Ron: Could I negative ballot the downstream work product if the dam is faulty?
    21. Bo: It requires a waterfall.
    22. Ron: Is the dam sufficient to scope the product?
    23. Jane: Purpose is the important part.
    24. Ron: Maybe the completeness is tested through the ballot process for the downstream product. We should say this is what a dam is, this is how it is use, and this is the downstream result.
    25. AbdulMalik: Implies that all of our specification are proceeded by a DAM.
    26. Patrick: that is where we are going.
    27. AbdulMalik: There needs a way to express traceability to the DAM.
    28. Lorraine: We are getting into the IG space.
    29. Ron: Governance has a role in checking the quality.
    30. AbdulMalik: If we find a quality problem while developing the downstream product, do we have to reballot the DAM?
    31. Jane: That is why you go DSTU.
    32. Lorraine: that is why we have DSTU in the first place.
    33. AbdulMalik: We started with our model, and looked to the BRIDGE model, then went to create message specification (RMIM-DMIM). They had a different style in the British Model - no traceability. We discovered importance of multiplicity in the MIM's that we did not expose in the DAM. At implementation we had a lot of work to keep the DAM, IG, and final product in sync.
    34. Ron: Am i overloading the DAM?
    35. AbdulMalik: I am having problem seeing people go back to the DAM. It is too much work.
    36. Ron: The problem is repeatibility. I have seen it at HL7. Where interoperability requires coherence and repeatibility. The likelyhood that you will have to reopen the bag is small. I am prepared to back away from the position. HL7 needs to get the DAM right.
    37. Andy: If HL7 had the discipline of a single trajectory it would work well. It is incredibly difficult to say ultimately what it means if they are not on a single trajectory. How do you recognize the intersection between trajectories?
    38. Ron: Same problem I have faced on every team. The most successful was where the rigor was resolved.
    39. AbdulMalik: The immunization DAM will spawn many products with no cohesion.
    40. Jane: People want the DAM to be normative to feel it is sufficiently mature to rely on for the downstream products. If that is what we want to have, we have no choice.
    41. AbdulMalik: There are some implications.
    42. Jane: We changed the MIF's to map more directly.
    43. Andy: We have the ability to keep things out of the dam. I would rather have something with less, rather than nothing with more.
    44. AbdulMalki: I you dont restrict datatype you help the downstream.
    45. Andy: The less restrictive the better.
    46. Patrick: We chose to use conceptual data types.
    47. Andy: I dont care how you get there as long at you pass these three points to get there.
    48. Jane:I am on the fitness for purpose kick: The perceptual model defines the shared purpose.
    49. Steve: You have described the OMG model approach. You might want flexibility as you go down.
    50. Ron: Traceability controls what you put in you DAM . It has to provide a sense of context of the domain. As you drive down you may find you have missed something. For those who want to reuse the thing you made, redoing the DAM is good. Charlie see if you can define at what is sufficient at that level.
    51. Steve: Do we want to define the artifacts that go in the DAM?
    52. Charlie: That is part of the SAIF-IG.
    53. Steve: Do we need an HDF, and the groups use their SAIF-IG at different levels?
    54. Charlie: That is what Lloyd was talking about - and IG for FHIR, V3, V2.
    55. Steve: ArB needs to develop a higher level IG. Some groups want to do mind-maps, some want to do sequence diagrams.
    56. Bo: you need an independent IG.
    57. Steve: The IG got so deep into it they got burnt out. It was at such a detailed level, and was very restrictive.
    58. Patrick: I see something at the level of the HDF which applies to the IG.
    59. Steve: ArB should be doing a middle level document.
    60. Lorraine: I thought the IG was that product.
    61. Steve: Then everyone would do their own thing.
    62. Andy: The CD is the model. You have to choose V3/V2 - how do you describe dynamic process. There are different ways to cover the matrix. I am not sure introducing another level will help.
    63. Ron: Conformity with the CD will be needed. We may discover that we need to frame the CD differently, after we have tested the IG against the CD.
  6. Adjournment at 3:00pm Pacific

--Tony Julian 22:06, 13 May 2012 (UTC)

Thursday Meeting Information

HL7 ArB Work Group Meeting Minutes

Location: Sheraton Wall, Vancouver, BC Canada

Date: 20120517
Time: 1:45pm Pacific
Facilitator Charlie Mead/ Parker, Ron Note taker(s) Julian, Tony/????
Attendee Name Affiliation
. Bond,Andy NEHTA
. Constable, Lorraine Constable Consulting Inc.
. Curry, Jane Health Information Strategies
. Dagnall, Bo HP Enterprise Services
. Grieve, Grahame Health Intersections Pty Ltd
. Hufnagel, Steve U.S. Department of Defense, Military Health System
. Julian, Tony Mayo Clinic
. Loyd, Patrick ICode Solutions
. Lynch, Cecil Accenture
. Mead, Charlie National Cancer Institute
. Milosevic, Zoran NEHTA
. Ocasio, Wendell Agilex Technologies
. Parker, Ron CA Infoway
. Quinn, John Health Level Seven, Inc.
. Guests
.
. Legend
. Present
. Absent
Quorum Requirements Met: Yes


Agenda

  1. Q3
    1. Call to Order
    2. Aproval of Quarters Agenda
    3. DAM - Kaiser Permanente Perspective - Peter Hendler
    4. Impact of Thomas Erl's presentation
    5. Saif pgm - project 477 CDA R3 - frame an opinion on separating semantics from rendering
    6. Risk Assessment and governance for SAIF Architecture - Assigned by the TSC.
    7. Substantivity???
    8. HL7 Business Architecture
      1. What are the key things to make the decision CDA/V2/FHIR/V3?
  2. Q4
    1. Maturity of HL7 and the Customer Base(Zoran/Andy)
    2. FHIR discussion
    3. Architecture of templating
  3. Other business and planning
    1. Conference calls will not resume until 2nd week after WGM- May 31, 2012. Will re-validate during WGM.
  4. Adjournment

Minutes

Lorraine, Jane, Steve, Ron, Bo, Andy, Charlie, Anil, Patrick Loyd Peter Hendler, AbdulMalik Shakir, lloyd

  1. q1
    1. Call to Order
    2. Aproval of Quarters Agenda
    3. DAM - Kaiser Permanente Perspective - Peter Hendler
      1. B4 the DAM it was clear what HL7 V3 meant - using model, vocabulary, Data types
      2. many of the DAMS dont have an object model
      3. It is said "Domain experts cant understand the RIM"
      4. They can learn the difference between a Noun and a Verb
      5. Can you at least separate the nouns from the verbs
      6. More complicated, but physicians can understand that a Person can be a Patient at Kaiser, the same at intermountain, but a Physician elsewhere.
      7. Why cant you at least have them thing this box is a "thing", this box is an "action", this box is a "Role"
      8. They should be able to understand the nouns and verbs
      9. Cardiology DAM does not differentiate between boxes, and some are vocab objects.
      10. It confuses things that boxes that are vocab as part of the model.
      11. Not only are the boxes arbitrary, if you do a search you find no reference to RIM. They should state in a paragraph that there is an intention to grow up and become something that is HL7 V3 - and may become a document or a message.
      12. To say Clinicians cant understand UML, they can understand Nouns and Verbs.
      13. Suggesting a DAM style guide?
      14. To say that a DAM describing the knowledge of a domain is not enough - tha tit should be relatively and consistently rendered.
      15. We are anticipation that downstream artifacts have traceability to a DAM.
      16. Maybe the RIM is super-complicated, but organizing your thoughts into Entities, Actions, and Vocabulary.
      17. Imposing RIM grammer on a domain may not be logical- hard to differentiate roles to participations. Purpose of a DAM is to get the thoughts on paper.
      18. Not necessarily the RIM, but at least Nouns and Verbs. That they are aware of the differences. The ones we are reading are wrong.
      19. No problem with the assertion ArB is to provide a definition of a DAM without specifying the contents. Also looking at Governance and Risk assessment. This would be the ideal place to explain goodness, and minimials. There are dialogs on whether a DAM can be normative. Ideal if DAMs were consistent. Have not gotten to them yet. There are tools that are useful in an iterative use. DAM must identify the scope of the domain, or purpose.
      20. If HL7 is going to vote on a DAM with no structure, it is wrong.
      21. Normative DAMs will require expression as defeind by the ArB.
      22. IF a DAM has only cardiology information, there is nothing to vote on. Can only vote on things that apply to HL7.
      23. What is an HL7 person supposed to vote on?
      24. An HL7 DAM needs traceability to the subject. Lab person would expect expression at a sufficient level that I would know what to do.
      25. What is the nature of the persons who expressed the scope of interest.
      26. We agreed that we have noting to say about clinical correctness. Is the form sufficient?
      27. To come up with an HL7 artifact - is it more trouble to get to V3, or to redo with a mappable.
      28. Some would be in a DAM implementation guide.
      29. A style guide would be useful - if I knew the form of expression and arrangement it would help.
      30. Dont argue that Clinicians cannot do this.
      31. In evaluating a DAM there are two:
        1. Correctness
        2. Completness from a definition standpoint.
      32. Cant navigate an RMIM both ways.
      33. We should be able to specify that Class names should be Nouns.
      34. We are talking about a degree of expression.
      35. Correctness we have as much exposure in RMIMs as we do DAMs
      36. Nothing in the RMIM keeps you from expressing my dog as a caregiver.
      37. HL7 people arent voting on the dogs participation, that the DAM is correct.
      38. If it is unstructured, why are you voting?
      39. There is some merit to traceability - you will talk about things in a aggregate way. We have to provide lattitude to expression of usefullness.
      40. Because we are talking about clinical domains, if there is reason to express it, we need to clarify the purpose, and declare the expectation that we can only vote on the structure - participation in your group should be broad enough to state that this is clinical information to support downsteam products.
      41. In terms of clinical correcness, HL7 may never be able to vote on this.
      42. HL7 can vote on the accurateness from a modeling standpoint, not the clinical decisions.
      43. In certain disease whether the bone is broken or not is sufficient. In Rheumatology, not so much.
      44. Do we have an inventory of all dams?
      45. Yes - the Board decided that DAMS would be given away for free.
      46. Not all DAMs were gobledygook.
      47. If it does not meet minimum quality, we should not give it away.
      48. I do have the makings of a style guide. It is not an HL7 style guide. We have a guide in the HDF - for the RIM. It would be useful to compare the DAM to the style guide. Do we modify the style guide, or the model. Using Classes for concepts might not be allowed.
      49. we might find patterns of things people are doing that we have not expected.
      50. Clinicians have expressed the concern over the DAMs in multiple settings this week.
      51. Patrick is starting a listserv for DAM modeling
      52. Wiki for DAM modeling.
      53. Sparx is doing UML training.
      54. Need to provide a non-clinical DAM expression.
      55. NCI did a golf-club model, and got slammed because it was not relevant. Need a clinical example everyone agrees on.
      56. Might be useful to have a bad example. Widely understood is a pharmacy order. I have product to be administered. The DAM should never say I am prescribing peniciliin for a malady, rather a substance for a patient.
      57. Some people mix up instance diagrams with the Model structure.
      58. There is a lot of interest in stating nomenclature. Mind maps and OWL are being used to understand the terms.
      59. Isnt a DAM the first point of call? Usually more about what should NOT be there, than what is
  2. Impact of Thomas Erl's presentation (10 Minutes)
    1. What were peoples comments?
    2. Was it the right level for the audience?
    3. Books are going to be useful - hope they get past seven pages.
    4. The CD's are better.
    5. It is clear that their is a depth of thinking and volume of work.
    6. It was generous of him to provide the product.
    7. Saif pgm - project 477 CDA R3 - frame an opinion on separating semantics from rendering#Q3
      1. You should be expressing the semantics. We should take a page from Xform.
      2. The rendering should be separate. Lloyd pushed back on it - Austin is in favor, Bob not so much.
      3. The stated concern was that the committee did not understand it.
      4. In the informative ballot there will be two different representations of an instance.
      5. This has implications to the architecture. If you cannot shift stuff around without chopping it up, there is a problem. There is a risk from the community voting on something that is not good for HL7. I recommended that the ArB that there should be a preference on the form. It wont stop the community from keepig to status quo. It will indicate there is no desire in the community for interoperability.
      6. We are going to see more opportunity to allow the members to make an informed decision toward one architecture or another.
      7. Is there a middle ground? Can we let R3 finish its course, and let R4 take on the issue?
      8. CDA R3 will die because the community is not enthused with R3 changes. The ArB if it is going to be useful and relavant to the organization, should know what is going on, and make their comments known to the community where relevant. Stepping into that role as an ArB is a good thing, and it is in this case appropriate.
      9. Was well received by the Architure Program.
      10. "Arb with Teeth" requires that we weigh in on the issue.
      11. Need to determine process -ArB alone or through the TSC.
      12. ArB's recommendation needs to be to the program. I would understand where the powers are. If you have the authority, do it, otherwise the TSC will have to weigh in.
    1. Risk Assessment and governance for SAIF Architecture - Assigned by the TSC.
      1. Jane needs to update the PSS before proceeding.
    2. Substantivity???
    3. HL7 Business Architecture - discussed in Q4.
    4. What are the key things to make the decision CDA/V2/FHIR/V3?
      1. What form of expression would this take?
      2. SWOT analysis - with decision tree. Informative artifact.
      3. RIMBAA has posted a white paper on the analysis.
      4. The audience is both public and internal. Today I know I will use 2.5.1 for Immunizations, because the government says so.
      5. Sometimes it is the function of the most mature specification at the point of decision time.
      6. We can work out where to place it.

Q4 Agenda

  1. Maturity of HL7 and the Customer Base(Zoran/Andy)
  2. FHIR discussion
  3. Architecture of templating
  4. Other business and planning
    1. Conference calls will not resume until 2nd week after WGM- May 31, 2012. Will re-validate during WGM.
  5. Adjournment

Q4 Minutes

  1. Maturity of HL7 and the Customer Base(Zoran/Andy)
    1. Common problem of driving from the bottom up vs the Top down. We are defining capability measures for interoperability.
    2. The organization can follow a path for maturity - which may not be relevant across every customer.
    3. It is a way of framing the maturity path.
    4. Andy will share the modeling that has been done.
    5. Should feed to risk profiles
  2. FHIR discussion
    1. What would Arb like to discuss?
    2. Gerald wants to understand how the ArB would see it in the context of an exchange paradigm around services. Would like to see from and ArB perspective on how to fit FHIR to the parts that make up FHIR into the HL7 Architecture. Specifically around the REST-ful aspects. I think we are in agreement and REST that it is a realization of the RLUS functional model. I can see some emerging concepts of the OO symantics.
    3. How we would do such a thing sounds like a project.
    4. Gerald: We had a mapping from RLUS platform independent and functional model to HDATA. I would suggest that we go back and follow through on mapping along those lines, with SAIF implementation as an RLUS service.
    5. Grahame: FHIR is not an RLUS model.
    6. Gerald: There is an HL7 model realized by RLUS with OMG.
    7. Grahame: I would be happy to look at consistency. I am not comfortable with the expression. The implication that FHIR will be an implementation of RLUS is not true.
    8. Lloyd: FHIR is more than RLUS, and has more functionality.
    9. Ron: If you could provide a dynamic RLUS/FHIR what would that look like? Conceptually, how would we proceed? Who would drive?
    10. Grahame: We will not be making FHIR more complicated to apply to RLUS.
    11. Lee: I heard that you looked at the restful FHIR that maps to RLUS.
    12. Gerald: It is.
    13. Grahame: If you say "FHIR" is the RLUS model. I think it does not.
    14. Ron: Of interest to ArB: Risk assessment. How are FHIR resources used in dynamic interchanges. FHIR does not express how the resources are used.
    15. Grahame: We provide a lite messaging expressing. SOA can define how to use resources. Technical interoperability works, but there is a layer of business interoperabiligy that is not define in FHIR.
    16. Gerald: Which is the heart of what pharmacy is doing around medication statements.
    17. Grahame: Mapping is useful? Bind to conformance?
    18. Gerald:I did not say conformance. You can add functional performance on top of that. FHIR does not address the administrative funtions.
  3. Charlie: FHIR will be defined and balloted as a resource collection. Resources will have profiles. TSC asked ArB to provide a first cut at how HL7 could govern the resources, and possibly the extensions and profiles.
  4. Charlie: Thomas Erl's material was the basis for the governance document presented.Governance Overview: Gforge Copy
  1. Architecture of templating
    1. Problem Statement:Assertion from Ron: When we think about the problem I am anticipating, over time we are getting interfaces in place, trigger events are stabile, what will not be stabile, physicians start creating new requirements for the content. The traditional ballot structure of binding the context to the design
    2. For a particular interface we template the content representation, and version them - and know that version changes would still be supportable.
    3. I know the templating, and the mapping - if there is new clinical content, we would notify those who have implemented that there is new content.
    4. Theoretically the metadata and trigger events are stable, so I would not have to re-test the dynamics of the interaction.
    5. We need an architecture for template registries. How do they get populated? Who populates them? A Clinician just stated that we provide the transport, but not the final information. We need tooling for those who will need many flavors, and change there minds often.
    6. How to achieve stability and predictability - and support changes in content.
    7. Templates business requirements answers some of the questions.
    8. Template registry - the template service specification might be a way to start.
    9. Coalescing the material that is flowing around the organization with the title template, and determine the inventory, and the congruency.
    10. Goes to scalability, and all the other ilities you are in architecture.
  2. Business Architecture (Q4)
    1. Bo:It is hard to determine how HL7 operates. There are processes, working groups, roles. I would like to know if there is a way to represent the touch-points and the work-group interactions.
  3. Charlie:People who succeed have a documented business architecture. No one would contest the value. Who would do it?
  1. Other business and planning
  2. Conference calls will not resume until 2nd week after WGM- May 31, 2012.
    1. Baltimore:
      1. Sunday Q1, Q2, Q3
      2. Tuesday Q4
      3. Thursday Q3,Q4
  3. Adjournment The meeting was adjourned at 4:45pm Pacific

--Tony Julian 23:49, 17 May 2012 (UTC)

Retrieved from "https://wiki.hl7.org/w/index.php?title=20120513_arb_face_to_face&oldid=85972"