2016-04-14 SEC Call Meeting notes

RM Release 1.0.4 currently 5 issues:

• • need final agreement on the digital signing one - I propose to do no more changes now, and to raise a new 1.1.0 CR
    • Heath Frankel (Unlicensed) - gave server-signed IHE XDS zip (CDA) file example
    • Heath Frankel (Unlicensed): should look at PKI-flavour of signing
    • Sebastian Iancu: we do svr-side flavour of signing
    • AGREED: push main changes to Release 1.1.0; limit current changes to 1.0.4
  • 2 CRs have been finalised
  • 2 more need to be accepted - relating to the rendering of Hash types in the RM.

Heath to elaborate on RM complexity issues

• Heath Frankel (Unlicensed) - Ocean will try single Entry Composition; problems remain in persistent Compositions; ?errors in clinical models? Need feedback to CKM community?
• Heath Frankel (Unlicensed) - radical new way of doing Problem List by reference; current problem is we don't have 'blame' for Entries in Prob List. How do I review a consultation and see changes done just from that consultation.
• Ian McNicoll - we should use a 'business level service' idea with API to deal with some of the tricky logic.

We can just tag what is there now and call that Release 1, as a baseline; this involves a) some technical aspects and b) doing the appropriate things on Jira to leave the correct 'paper trail'

• Ian McNicoll - progress:
  • is using Java version as baseline; 
  • tag existing ver of 'terminology' repo as '0.x.x' to reflect that it has never been used; then commit a copy of current state of Java files & add a tag 1.x.x (= currently used terminology); then make changes and tag that as 1.1.x or 2.x.x as appropriate.
  • stick with 2 file split with adjustments;

I have been held up on two issues:

• there is a need for a redefined node to replace its parent node without the parent still being available. I spent a couple of weeks looking at this and determined that we probably need a new keyword or marker that means 'replace' for a C_OBJECT node, which prevents the parent node from still being available.
• there is an ongoing discussion about the 'rules' part of the archetype, as per the recent wiki page. I have decided it is better to just complete the specification in the original form - re-using the new 'rules' BASE spec we agreed in Stockholm (you have seen the draft). I don't think this is the final answer, but at least it is based on separated rules; if no-one uses it - good, it gives us the time to design a new approach for representing cross-dependency and get it added via proper PRs/CRs

I don't think this needs much discussion now - probably just SEC agreement in principle that we go ahead with the rules bit, declare Release-2.0.x (I think it will be 6)

Heath Frankel (Unlicensed) - what is the impact of ADL2/AOM2; what is the impl experience?

Diego Bosca - provided ISO update; Current version of AOM is accepted as a first draft for the renewal of ISO13606-2. Some countries have provided feedback that can probably be translated into CRs.

Thomas Beale

Rong Chen

• replace existing UML with UML model of BASE rules (by inclusion)  extra when/then rules as per GDL. This should be done in MagicDraw so that we take advantage of the same extraction approach as with the other specs.
• There are some formatting and heading numbering anomalies that I think need to be addressed.
• currently the syntax is inlined, but with Asciidoctor, we can actually do an include of the real syntax (where is it?) - I can show how to do this.
• I am happy to work on this with Rong, or do most of it, or help someone else do most of it (i.e. just help with MagicDraw and Asciidoctor specifics) - up to the SEC to decide.