2018-10-15 & 17 - Face to Face MAG Meeting (Vancouver)

Start GTM Recording

Meeting Introduction and Agenda Review

@Peter Williams

5

DL Enhancements Feedback and Progress

@Yongsheng Gao

@Peter Williams

10

• Stated Relationships file becomes invalid in January 2019 although current timeline suggests available for 2 releases. Suggestion for making file available "On Request" ?

• The Stated Relationships will remain the source of truth for the Jan 2019 release. Complete OWL Axiom reference set will be made available on request.

• In the July 2019 release all stated relationships from the previous release will be published as inactive and the complete OWL axiom refset will be the source of truth.

• As part of logic profile enhancement, plan for the solution development (including concrete domains), documentation and roadmap for implementation

Archived Documents:

Description Logic Enhancements - Briefing Paper

Roadmap for the DL Enhancements.pptx

@Yongsheng Gao to ensure July release clearly states that loading then OWL Ontology Refset into the classifier is obligatory for correct classification.

DL Suggested that we have both Stated and OWL rows live in the January release. General agreement.

Will distribute both OWL refsets in single file.

Logic Profile Specification

@Dion McMurtrie

10

SNOMED CT Logic Profile Specification - Status: Published

Dion - happy with list of datatypes against common classifier capabilities?

OWL Refset Specification

@Yongsheng Gao

25

SNOMED CT OWL Guide

Update for OWL Guide.pptx - Slides for the meeting

Update: "Defined" generally agreed to mean - "there exists a sufficient definition". We will inactivate the Synonym "Necessary and sufficient definition" since we will now have subsumption without necessary conditions.

OWL Refset Specification

@Yongsheng Gao

30

Discussion continues

Working with Historical Associations

@Peter Williams

30

Working with Historical Associations

What is needed here is some clear documentation about how Historical Associations are supposed to work - when to create, when to replace, whether various types should be treated differently.

Historical associations (unlike concepts) continue to grow - an increasing snowball. Why not reuse UUIDs rather than allow that snowball to accumulate?

See also TIG 5.2.1.4 Association Reference Set

Terminology management evolution

@Yongsheng Gao

20

Items arising from Saturday meetings:

• Guidance for modularisation as a way to share content.

• Future work: Review specification for change set exchange ie content + metadata

• Terminology Server Content Exchange Interface

QA for that module would also need to be shared.

Refset Metadata

@Peter Williams

10

This topic comes from SNOMED on FHIR

Also TRAG item: Reference set metadata

In fact, this metadata may be more widely useful when applied to a module.

Use cases: Public Library of Simple Refsets, ?

Two separate situations:

• Human information like what is the refset to be used for, who owns it, etc.

• Missing technical information eg what is the target code system of a map, keys vs data, machine searchable fields.
  
  

Suggestion that we could use the FHIR standard which already defines metadata. DM suggests a JSON format which would be extensible unlike a column based solution.

There is a more general desirable case to be able to annotate various components.

Start GTM Recording

Review of Agenda

@Peter Williams

5

Definition status

@Yongsheng Gao

15

Classification Service, SNOMED OWL Toolkit and generation of NNF

@Kai Kewley

@Yongsheng Gao

30

Classification Service with documentation available in GitHub

• Makes use of the Snomed Owl Toolkit which is capable of converting SNOMED CT Snapshot files into an OWL ontology file (functional syntax) for use in 3rd party tools such as Protégé.

• Review Necessary Normal Form for Inferred relationships (draft)

Concrete Domains Roadmap

@Yongsheng Gao

20

RoadMap for concrete domains: Concrete Domain

• External and technical requirements: Concrete Domains

• Impact to end users

Specification for OWL file

@Former user (Deleted)

30

• Requirements and use cases

• Journal paper on SNOMED OWL

Proximal Primitive Parent modelling vs Immediate Parent

@Dion McMurtrie

@michael lawley

30

"Modeling Content Using More Than Proximal Primitives (201739)"

AOB

@Peter Williams

5

Concrete Domains discussion continue

GR Asked for format of GCI rows in OWL refset to be clearly documented.

Facilitating Namespace owner lookups

@Peter Williams

Facilitating Namespace Lookups

Use cases:

• Extension maintainer wishes to raise modelling issue with originator of concept