20080207 arb minutes

Participants

Abdul Malik Shakir Cecil Lynch Jane Curry John Koisch Grahame Grieve Yongjian Bao Charlie Mead Tony Julian

Agenda

• Roll-call
• Minutes from last meeting
• Tooling responsibilities
• Outstanding ArB issues
  • Localization (‘V2 field length’)
  • GE ballot issue
  • Statement on Backwards Compatibility
• Re-org project
• Status of Mission Statement
• Ballot QA project
• Scope: Current vs Future Architecture

Minutes

• JC concurred with Charlie's assessment of the TSC's view of the ArB's role in tooling. The summary is (from TSC) 'requirements will be generated by the Tooling Committee, as will documentation of the current fucntionality/process.' ArB only needs to weigh in at some point once the requirements have been generated to make sure that no significant architecture issues have been raised. CL - ArB role is making sure that the tooling meets the real use cases of the pieces and parts that the tool might touch. JC is here to make sure that the requirements are understood. JK suggested breaking down tooling discussion based on rm-odp viewpoints. JC noted that the discussion of how the standards and specifications fit together will directly influence the tooling. GDG is concerned that the tooling is led by the MIF, and that the MIF is neither proper documentation of the HL7 meta-model, nor is the scope of the MIF properly resolved in terms of what the MIF doesn't do. This is a real issue at this point of time because Lloyd is working on Vocab MIF at this point and doing things like SNOMED subsetting and representation in the MIF. CL - there are more robust industry tools that should be integrated into HL7 tooling. CM noted TSC does not want ArB to worry about tooling as its number 1 priority, at least not until the Tooling Group (aka Jane and company) document the combination of requirements and current processes/algorithms etc. JohnQ is concerned that OHT needs requirements and current specifications and HL7 can't provide them...it's a much more tactical issue than it is an architecture issue (at this point in time). JC is concerned about scope as well - it will hopefully be clarified in the next few weeks as the discussions go on about where HL7 stops and other groups pick up. There are hints in other collaborative orgs to establish a common industry standard.
• CM : the issues are posted as 'need to be resolved by the ArB' on the TSC/tracker site. HQ called me today and said that we need to weigh in on 3 issues (plus finish our mission statement).
  • Candidate response for the v2 issue: "The issue of length normativity (?) is one of which technical bindings should be implemented at the local level. It is not the purpose of the ArB, MnM, or of HL7 to mandate or standardize technical bindings, recognizing that they are a matter of local implementation and governance. To that end, the standard should stand as written, which says that length is normative in the standard, but may be overruled in local implementations between trading partners." GG - believes in a more measured response. We should look into why lengths are important at all. YB - Length of field is part of normative definition. GG - why do we bother defining lengths at all? Changing actually increases the price of implementing HL7. CL: "We changed the text to informative." CM: i agree that the notion of length is an instance of localization...you localize and you pay a known price in non-interoperability. why does it need to be more complicated than that. the normative standard is to maximize interoperability. when you deviate, you get less, when you comply, you get what's promised. TJ - "Receiver must be able to receive up to the maximum field length" - this is in section 2.5.3 of the standard. AMS - two separate questions: #1 - Someone stating compliance. If vendors have not handled max length in the standard, then are they compliant? When you differ from the standard are you still in compliance? #2 - Should HL7 be dealing with length issues at all? JK suggested offering a statement like "ArB recommends that INM crafts a layered standard of conformance for HL7 that acknowledges localization, and ArB suggests that TSC gives INM the authority to look into why lengths are being made normative or informative at all." We will look at this.
  • GE Ballot Issue: CM: GE voted negative on a ballot because the ballot contained material that had not been through harmonization. They will not withdraw their negative and have asked ArB to weigh in. CM's opinion: tell GE that ArB is now sponsoring a ballot QA project that would prevent this kind of thing from happening in the future and ask the responsible committee to a) check all of their material a second time and b) submit it for a new ballot in May. AMS - If this is an item that should have been harmonized, then you cannot find it non-persuasive. If it does not need to be harmonized, then it can find .... CL: Note to TSC: ArB will find out whether the item needed harmonizing at all and then make a ruling. CM: there is no need to go back to the TSC with a question about the harmonization-relevance of the content. the committee agrees that it was supposed to have been harmonized and wasn't...they just don't accept that as reason for a substantive negative. GG: Suggests adding that the ArB review the difference between Balloting and Harmonization as competing governing processes. If everything gets punted to harmonization then you are balloting content that is not semantically correct. GG: There is a fundamental issue here that makes balloting difficult - balloting v harmonization v ArB. AMS: Motion to say: "This is a reasonable substantive negative, and the committee should proceed based on that assumption." JK seconded. CL called the Question. Motion passes 5 - 1 - 1. Tony Julian