2017-07-18 Full MAG Meeting (Conf call)

Introduction

Timescales

New Observers

PWI

YGA

• See section 6.2

• Reuben, Tomasz and Michael Osborne now Observers

• Brandon Ulrich has been proposed to the nominating committee

• Update from last week's MAG DL Subgroup meeting (YGA)

 Update on SNOMED DL Enhancements

YGA

DM

Update on Classification Service - intention to support output of OWL without necessarily classifying.

https://docs.google.com/document/d/1tqNEA6S4fEF4fgj15OPabYA2E0VTz8epxvRRwczKizQ/edit#heading=h.tbua8b7c0k7k

Discussion: YGA Negation. Singapore is using the full DL profile including negation. International edition will use EL profile. Potential impact on Australia's edition. Potential improvements in classification may bring DL processing times down to what we currently enjoy for EL classifications.

Modularity - Option 2 is the only interpretation. Option 1 will become possible to expressing using OWL directly, but is not an option for RF2 to OWL transformation.

Steps Required

PWI

YGA

Consultation exercise - complete in time for October Conference discussion?

(YGA) End of July - provide paper to AGs for review. Comments back for end August.

September - wider dissemination of simpler format - specific questions?

• Distribution format to represent concrete domains

• Deprecation of Stated Relationship file (suggested for 20190131)

DL Enhancements - Subclass Of, temporary position

For 20180731:

• Definitely in the OWL Refset.

• Will they also appear in the Stated Relationship file? Consensus was No.

• If not a stated relationship, could be expressed in tooling as an IS A with some indicator to show that it should be considered a separate axiom. YGA Keen to see user interface work towards the 'true' endgame of making the separate axiom obvious, rather than some non-obvious workaround (such as an IS A in relationship group 1).

DL Enhancements - Concrete Domains

• Short Term - since the "numbers as concepts" solution is now available for

• Long Term - what will appear in inferred form? Inferred concrete file to augment inferred concept file.

• 3 Options:

  • Refset in the mode of Singapore and Australia

  • Modify existing inferred file

  • Create additional counterpart to inferred concept relationship file: "inferred concrete relationship file" (datatypes relationships file - ML)

  • (Not desirable) Continue to generate concepts as numbers 'on-the-fly'

See 5.3.1.8 Substantial discussion in document comments.

DL Enhancements - Static OWL

GR / HS ?

Neither present on call.

Is the OWL Refset draft available?

AOB

PH Asked for Role Chain example - see MAG DL pages.