Title:        Report from public comment period on DCMI Abstract Model
Creator:      Pete Johnston
Identifier:   http://dublincore.org/documents/2007/04/02/abstract-model/comments-received.html
Date:         

Period: 2 April - 30 April 2007

This document summarises the principal comments received during
the public comment period on the DCMI Abstract Model (Version
) Proposed Recommendation [1], a revised version of
the previous Proposed Recommendation (Version ) [2].

[1] http://dublincore.org/documents/2007/04/02/abstract-model/
[2] http://dublincore.org/documents/2005/02/05/abstract-model/

Proposals for moving forward are presented in the sections below.

OWL-DL, datatype properties and object properties

    OWL-DL will require properties to be declared as either
    datatype properties & object properties. If both literal
    and non-literal values are required, then may need two
    parallel properties.

    This is being taken into account by the Usage Board in
    their consideration of the ranges of DCMI properties.

    Proposals: No change required.

    References:
    -- Comment: Ivan Herman,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0704&L=dc-architecture&P=1308
    -- Reply: Pete Johnston,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=424

Specification of which value string is intended for display

    If statement has a non-literal value surrogate with
    multiple value strings, how does the application decide which
    string to display?

    DCAM does not provide this but can make use of the description
    of a value to differentiate "categories" of label e.g. using
    SKOS preferred/alternative label properties.

    Proposals: No change required.

    References:
    -- Comment: Douglas Campbell,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=56
    -- Reply: Pete Johnston,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=700 

Definition of literal

    Provide fuller definition of literal rather than just
    referencing RDF Concepts.

    Proposals: Expand description/definition of literal in
    DCAM, based on RDF C&AS, XML Schema.

    References:
    -- Comment: Douglas Campbell,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=56
    -- Reply: Pete Johnston,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=700

Requirement to use URI to reference VES & SES

    The DCAM requires that properties, VES & SES be referred
    to in statements within DC metadata description sets
    using URIs.

    Proposals: No change required.

    References:
    -- Comment: Douglas Campbell,  
       first aid stations .  how to find bed bugs 

        This Task Group is for collaborative work on Resource Description and Access (RDA), building on agreements made at a meeting held at the British Library April 30/May 1, 2007. Participants in the meeting came from DCMI and other Semantic Web groups, and the RDA development effort. The Task Group is led by Diane Hillmann of Cornell University and Gordon Dunsire of Strathclyde University.
        -- Comment: Douglas Campbell,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=56
    -- Reply: Pete Johnston,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=700

Language and Typed Value Strings/XML Literals

    Typed literals cannot carry language tags.

    Proposals: No change required (though maybe need to emphasise that 
    typed value strings includes XML Literals?).

    References:
    -- Comment: Douglas Campbell,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=56
    -- Reply: Pete Johnston,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=700

dcam:memberOf equivalent in RDF

    RDF has no "built-in" equivalent of VES, so no built-in equivalent of dcam:memberOf.

    Proposals: No change required.

    References:
    -- Comment: Douglas Campbell,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=56
    -- Reply: Pete Johnston,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=700

Definition of "term"

    Definition of "term" should be part of bullet list in section 2.3.

    Proposals: Add bullet defining term near start of list.

    References:
    -- Comment: Douglas Campbell,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=56
    -- Reply: Pete Johnston,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=700

Set of resources (VES) v Set of literals (SES)

    Not always clear when to use VES v when to use
    SES. A modelling choice. Depends on requirements of
    application. Usage Board choices for DCMI terms based on
    content of existing specifications.

    Proposals: No change required.

    References:
    -- Comment: Douglas Campbell,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=56
    -- Reply: Pete Johnston,  
       http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=dc-architecture&P=700

Description of "subproperty"

    Bullet on subproperty in 2.3 should use the phrasing
    "...and the property is part of a property/value pair"
    (rather than "...and a described resource is related to
    a value by that property")

    Proposals: Change phrasing of sixth bullet in section 2.3.

    References: Offlist comment by Pete Johnston

Error in 2.4

    Erroneous reference to "syntax encoding string"

    Proposals: Change "syntax encoding string" to "syntax encoding scheme".

    References: Offlist comment by Pete Johnston

Abstract Syntax Required

    Alistair Miles re-emphasised the points made in the
    previous comment period: the need to delineate an abstract
    syntax, to describe how to merge DC description sets,
    and to describe inference rules. He also suggested the
    need for use cases to specify the requirements that the
    DCAM is intended to meet.

    Proposals: In short term, move forward with current
    version; in longer term, gather requirements using current
    DCAM as input, and consider revision/restructuring along
    lines suggested.

    References: Discussion during telecon, 

Definition of "vocabulary"

    DCAM uses "vocabulary" to refer to set of "terms"; RDF
    Semantics uses "vocabulary" to refer to set of names
    (URIs/literals).

    Proposals: Current DCAM use is consistent across DCAM
    and Namespace Policy document; usage in W3C docs does
    vary. Retain current use in DCAM.

    References: Discussion during telecon,