The DCMI Abstract Model, which attained the status of DCMI
Recommendation in March 2005, has been revised in light of
discussion and feedback from the DCMI Architecture Working
Group, the DCMI Usage Board, and the broader community.
This revised version of the Abstract Model [1] has been
posted for a four-week public comment period. The major
differences between this revised version and the 2005 version
[2] are summarized below. A revised DCMI Namespace Policy
[3] proposing a new DCMI namespace for Abstract Model entities
has been posted for comment at the same time.
Interested members of the public are invited to post comments
on these Proposed Recommendations to the DC-ARCHITECTURE
mailing list [4], including "[DCAM Public Comment]" in the
subject line. Public Comment will be open from 5 February
through 5 March 2007.
[1] http://dublincore.org/documents/2007/02/05/abstract-model/
[2] http://dublincore.org/documents/2005/03/07/abstract-model/
[3] http://dublincore.org/documents/2007/02/05/dcmi-namespace/
[4] http://www.jiscmail.ac.uk/lists/dc-architecture.html
Changes in the DCMI Abstract Model, 2005 to 2007
-- Added a table explicitly mapping Abstract Model entities to
properties and classes of the Resource Description Framework
(RDF) and RDF Schema (RDFS). The DCMI Abstract Model defines
a particular Description Model on the basis of RDFS semantics.
-- Added a separate Vocabulary Model specifying the types
of "terms" used in Dublin Core metadata descriptions and
including support for formal domains and ranges for properties.
The declaration of a vocabulary of classes and their use
as domains and ranges for DCMI properties is the focus of a
separate Public Comment period.
-- Changed the definition of 'vocabulary encoding scheme' --
defined in 2005 as "a class that indicates that the value of a
property is taken from a controlled vocabulary" -- to mean "an
enumerated set of resources" of which the value is a member.
(A value can be described as an instance of a class by other
means, such as by a separate statement to this effect).
-- Updated the definition of 'rich representation', adding
the notion of 'media type'.
Other editorial changes
-- Tightened the definition of 'syntax encoding scheme',
explicitly mapping the concept to the RDF Schema class
'Datatype'.
-- Tightened terminology and wordings to clarify meaning
(e.g., by consistently using phrases instead of sentences for
definitions; by referring to 'described resource' instead of
just 'resource'; by using the phrase "separate 'description'
about the 'value'" instead of a modeling entity for 'separate
description').
-- Shortened the document by removing sections describing
related issues such as 'dumb-down' (formerly Section 5),
'structured values' (formerly Appendix A), and specific
encoding guidelines (formerly Appendixes B, C, and D).
Much of this material will be provided in revised form in
more user-oriented documentation.
-- Added a table mapping current Abstract Model terminology
to the terminology in legacy DCMI "grammatical principles"
documentation (now Appendix A).
-- Permitted a value string to be associated with either
a language tag or syntax encoding scheme, or neither, but
not both.
-- Added a note to the effect that classes can be declared
explicitly or inferred from the domains and ranges of
properties. Dropped the guideline that in DCMI metadata
descriptions, the class of the resource being described should
be indicated by the value of the Dublin Core Type property.
-- Simplified the Description Model, removing 'marked-up
text' and 'structured value string' as separate entities and
rearranging the diagram to improve readability.
-- Added placeholder URIs identifying DCMI Abstract Model
entities in a new DCMI namespace (as described in the
revised DCMI Namespace Policy, also posted for Public
Comment).
-- Replaced QNames throughout the document with full URIs.
--
Dr. Thomas Baker <[log in to unmask]>
Director, Specifications and Documentation
Dublin Core Metadata Initiative
|