20100401 arb minutes
Template:Arb minutes Architecture review Board Meeting Agenda
April 1, 2010
Contents
Agenda
- Call to order
- Approval of Agenda
- approval of Minutes
- Review of action items
http://gforge.hl7.org/gf/project/saeaf/tracker/?action=TrackerItemBrowse&tracker_id=590
- Peer Review
- Governance
- Introduction
- Information Model
- Review of Wiki Pages
http://en.wikipedia.org/wiki/HL7_Services_Aware_Interoperability_Framework
- Other business and planning for next meeting
- Adjournment
Attendance
Cliff Regrets Karen Smith, Tony Julian, Ron Parker,Jane Curry, CecilLynch, Leon Cameron,John Koisch
Call to order
The meeting was called to order at 11:00 am U.S. EDT with Ron Parker as chair and Tony Julian as scribe.
Approval of Agenda
MMS to approve the agenda(/) (0-0-0)
approval of Minutes
SAIF peer review
- Governance
Jane Curry: I will be devoting my weekend to getting the GF done.
- Information Framework
Cecil Lynch: At NCI we are into vocabulary elements with 1179, which suffers from early binding. Once you have bound a concept you are done. It makes it useless to bind an abstract model. I am working on showing how to bind the ARCRIM model to HL7 V3 CMETS, and carry all of the defining information about the context, without fixing it to the object - a UNL composition association. When you look at the information framework, the DAM is not an appropriate place from which to derive your models, you have to work down one layer. I am woking stp-wise from a cncept all the way down. The issue I am having from Ann Wrightson, she has the feeling that we have to reach-out to people who are using text-blobs (PCDATA), which we can do if we have a formalism. We have talked about we pretty much loose our readers if we go into first order predicate calculas. We will put out to ArB to decided what to include.
Ron Parker: Curious about Ann's rationale.
Cecil Lynch: CDA with a bunch of text.
Jane Curry: Unless you get a physician to sign off on it,
Cecil Lynch: We spent many hours on the phone going through it, trying to understand what Ann was trying to say in her concept map. I am waiting on the input. Charlie will be applying text to our concept map. I am structuring it in ontologies, the best way to see it. My example will be in OWL mapped to the usual artifacts in our VISIO diagrams. They get lost when you throw them into a HMD spreadsheet.
Jane Curry: The whole thing managing ontologies is about subsetting values, and the point where it occurs. It is important to understand, managing to the point where they are bound. Coherence is a problematic explanation - you have to have coherence in environment.
Ron Parker: Fascinating conversation.
Cecil Lynch: I have to have a white-paper by April 2, which will give us a start, it will go to Charlie to edit, and mold to the IF. I will have a good stab at the IF by April 2 - tomorrow. I have a OWL content for that for a NCI meeting monday.
Ron Parker: I can see lots of points where we should choose to pull in MnM, and othere on a dialog, what we need to do in SAIF is to inform not only the way information is represented in payload, as well as contextualizing the interface, in term of contracts. Will expose other things.
Cecil Lynch: I have a RIMBA presentation at the WGM where I will be going over our V3 back-end data model, that we have implementa at MD Anderson Cancer center. It was eye opening as where our models are vocabulary bound - you can abstract it out, render in a generic api for an enterprise service. If you serivce know how to work with 20100 dat types, your service can work with that - a service on the bus that interfaces with that, and allows you to specialize the services. I am leading the semantic infrastructure at NCI. It will inform this at NCI to form an outcome service - everything is a variable on the outcome - address, therapies, genetics, socio-economics. If you are going to build an outcomes service 1 service - to access your data - you have to have a generic service - working with the datatypes, structure your services to deal with it. How to atomize it at the service interface level.
Ron Parker: Add discussion with Charlie about implementation guides.
- Introduction
Ron Parker: I am done with intro. Conclusions from the material. The intro was 72-78 pages, and had several components. I am proposing a framework - move the implications for HL7 as an appendix, or adjunct document. John Quinn has said that ONC is interested. We need to compartmentalize for our casual reader, intermingled with internal group. I have externalized HL7 implications as a 13 page adjunct document. In the history section, we had dialog about HSSP and the implications. Tis is draft for peer review. Same thing for relationship with other SDO's. There are elements of this piece that should be in the intro, and executive summary. Could be 3 paragraphs in the INTRO. There is contextualization of "Arb thinks" - I have left in history, change in intro to sound like HL7, and SAIF as an expression of content - shared assets of the documents. Now is 44 pages - with heavy emphasis of contextualizing about services aware. This issue around re-contextualizing SAIF as a services-aware, patterened around a framework. We are not trying to be a new TOGAF or RMODP. RMODP is a good starting place, especially the viewpoints. We are working around proven approaches to architecture, - the way products are produced. Someone in progress can incorporate the concepts into the work they are doing. Instead of rolling your own everytime, you should be able to use the SAIF to increase your probability that as a design parameter, you will be able to interoperate. The rest is a bit of editorial - making readable. We will have to review the material again. It would be appropriate to pass through it - we wont have time to concensus build.
Jane Curry: Are you putting the change management process in the intro? People need to know they have an opportunity to kibbutz.
Ron Parker: It is absent - not sure it is in intro - very important to engage with SAIF withou burdening around the process models. Letter, or intro to intro - change management, accept to evolve and adapt. I want the INTRO to be about SAIF, not the process.
Jane Curry: People may not know about our process, or their participation.
Ron Parker: Karen and I will be re-generating to a numbered PDF. I have not put out announcement, until INTRO is done. I dont feel we can put out ECCF and BF withou contextualizing the INTRO.
Ron Parker: Implementation Guides - Charlie Mead has made it clear that we need to strongly differentiate between HL7 implementation guides, and external implementation guides. We will produce more than one IG - at least 1 for HL7.
There is absolutely intent to inform external organizations on how to structure enterprise architecture. There is a template or guide to HL7 use of SAIF, that needs to be understood by the community.
John Koisch: What should HL7 be standardizing?
Jane Curry: ECCF says to refine for your community.
John Koisch: You need to know how it maps to your stuff.
Jane Curry: It becomes difficult to understand how to localizing if you dont have a structure to localize from. If you do have one, then describing it is easier.
John Koisch: Attractiveness of Information Models, localizing. Engineers make decisions - having gone throght the specs, there are assumptions made that may or may not be ok. We spent a lot of time cogitazing on this, the faster we build stuff, the sooner the separation will be apparent. In RIMBAA and HL7 V3 there is a lot of stuff. NCI is having to take into account HL7 work.
Ron Parker: They need to understand the requirements - do they resonate to our purpose? We have things we need to do that will not be in the specs that are contextualized. It would be interesting, we should provide template for IG - I dont know how we will mature without insight into the way it is being implemented. As people discover things that dont work - we have jurisdictions that tried without SOA, and are changing their minds.
Ron Parker: I will push out the release on tuesday - expect to get a bunch in a flurry before the end of the month - so we will have to prep for RIO. We had conversation about information framework - IG discussion will be preview - the SOWHAT of SAIF.
Review of action items
http://gforge.hl7.org/gf/project/saeaf/tracker/?action=TrackerItemBrowse&tracker_id=590
Other business and planning for next meeting
Ron Parker: Calls are struggling - should we make them 1/2 hour longer? My ability to engage is stressed.
Jane Curry: We went to bi-weekly to accomadate Austrailia. We have not increased participation.
Ron Parker: We could go to bi-weekly, and still rotate.
Jane Curry: Mike Kingsley can create perma-links. John Quinn was telling me about the good new/bad news - NCI is implementating, but we will be under a microscope. Communication strategy is as important as the produce. We will put peer review material there - did not want to go so public quickly.
Jane Curry: JQ was comfortable with draft work but we need to be responsive.
- Call to order
- Approval of Agenda
- Approval of Minutes
March 25, 2010 Minutes April 1, 2010 Minutes
- Review of SAIF presence on HL7 Page
http://www.hl7.org/permalink/?SAEAF
http://www.hl7.org/permalink/?SAIF
- Review of action items
http://gforge.hl7.org/gf/project/saeaf/tracker/?action=TrackerItemBrowse&tracker_id=590
- Review of Wiki Pages
- Other business and planning for next meeting
- Adjournment
Adjournment
The meeting was adjourned at 12:01 EDT. Tony Julian 19:39, 1 April 2010 (UTC)
