SC 22/WG17 Minutes
Date of meeting: June 1st 2020
Time of meeting: 17:00 UTC -- 17:45 UTC

Participants:

    - Markus Triska (Austria, substitute for indisposed convenor)
    - Mark Thom (Canada)
    - Stefan Kral (Austria)

1. Greeting and introduction of participants, administrativa

    Convenor Ulrich Neumerkel is indisposed.
    The group wishes him a speedy recovery.

1.1 Minutes of our 2019 meeting (N280), available via: iso.org/jtc1sc22wg17=
 -> Navigation Menu, N-Documents List -> N280

    No requests for changes.

2 Progress in Prolog systems, discussion of implementation practices and on=
going alignment, current areas of interest in the application and implement=
ation of Prolog

    Technical aspects of the discussion include a new, very compact
    representation of strings as lists of characters, common interfaces
    for attributed variables, and the performance of conforming
    extensions and implementation techniques for constraint programming.

    A clean interface for delimited continuations is being discussed
    as the basis for tabling extensions (SLG resolution) in new systems.

    The group expresses hope that newly arising Prolog implementations
    can benefit from each other due to their strong conformance to the
    standard. One example that is mentioned: New Prolog libraries that
    have recently been shared between Scryer Prolog and Tau Prolog.

3. DCGs N279

    Postponed since the editor of the DCG draft is indisposed.

4. Corrigenda, Discussion of N273

    Drafts for further corrigenda are briefly discussed.

5. New projects and liaisons.

    In the public sector, there is increased interest in symbolic AI
    approaches. How can we automate the application of existing laws?
    NZ government already tries to formulate laws as logic programs.
    COVID-19 has increased the urgency to easily formalize laws
    (e.g., eligibility requirements for grants) and to automatically
    reason about them. Automated decisions must be explainable.
    
    Different Prolog systems are beginning to support Unicode programs
    in incompatible and misaligned ways. It is resolved to work on
    this topic at the next meeting of WG17.
    
    The RuleML community expressed interest in the pure subset of
    Prolog as an interchange format for rule representation. A clear
    delineation of the pure subset of Prolog is asked for.
    Even though Annex A mentions "pure" Prolog, the actual standard
    does not define it.