Dear members of the dc-international list,
As many of you have read on meta2, DC6 was a momentous workshop for
Dublin Core. For those of you who do not subscribe to meta2, I have
attached Stu's excellent summary below. Some news:
1. My report on the activities of the Working Group on DC in
Multiple Languages was well-received and -- exceptionally for DC6 --
2. At AIT, we have finished a prototype distributed registry with versions
of Dublin Core in many languages formatted in RDF. Before putting this
online over the coming weeks, we will be in touch with many of you about
URIs for these files on your Web sites.
3. The Dublin Core Initiative is adopting more formal procedures for
reporting and archiving the activities of its working groups. The Web
page I maintain at the URL below has "half-migrated" to OCLC, but we still
need to work out a procedure for handling the work-flow through email.
I realize that some documents are out of sync and will give this my full
attention after December 15.
4. Most pressingly (for me), I am writing an article on Dublin Core
in Multiple Languages for the December 15 issue of D-Lib Magazine and
have been in touch with many of you this past week requesting related
information and URLs. I will send you all some mail soon asking for
your advice and comments on aspects of the draft.
The Dublin Core community clearly sees our work as important, and a
number of us are strongly committed to moving forward on the agenda we
have described in our position paper. I look forward over the coming
weeks to reopening discussion of the position paper in light of the
developments that that began in Washington.
From: "Weibel,Stu" <[log in to unmask]>
To: "Meta2 (E-mail)" <[log in to unmask]>
Subject: Summary of the Dublin Core workplan for the coming year
Date: Thu, 12 Nov 1998 09:19:53 -0500
The following outline summarizes the program of work the DC Community has
ahead of it for the coming year. A more detailed treatment will be reflected
in the DC-6 Workshop Report.
I. DC Process
Establishing a formal process for maintenance and evolution of the Dublin
Core is a central requirement for continued progress. The Dublin Core
Metadata Initiative needs a coherent process that makes it clear how we do
our work, how change is initiated and ratified (or not), and a clear record
of that change. The beginnings of this process have been in place for
about a half year, in the form of two advisory committees that provide
guidance on technical issues and policy issues confronting the Dublin Core.
One of the first assignments of these committees has been to prepare a
formal process document, and I am pleased to report that an initial draft of
that document has been produced and will be refined and presented to the
community over the coming weeks.
There are many models of goverance to choose from. The subcommittee working
on this process will identify features from various models that most closely
match the culture of open consensus development that has become a hallmark
of the Dublin Core Metadata Initiative. The process that emerges will
reflect the multi-disciplinary and international characteristics of the
Dublin Core which are at the heart of our progress thus far.
II. Polishing the Finnish finish
RFC 2413 publically defines the semantics of DC that have been largely
stable since December of 1996. We can think of this as DC 1.0. Deployment
of DC 1.0 semantics has made clear that minor changes in the definitions
are necessary to clarify the intended semantics and to reduce ambiguity
concerning their application. A review of the existing definitions will be
undertaken and small changes made as judged necessary. The fundamental
underlying semantics will not be changed, nor will the number of elements.
One of the operational criteria for these changes is that existing
applications should require no changes as a result of these clarifications
of the definitions. Rather, these changes should support greater
consistency across applications, and should make the semantics clearer to
new implementors. It is anticipated that these changes will be incorporated
in a revised document that can be thought of as DC 1.1.
III. Ratifying DC Qualifier Semantics
Few if any current DC applications have been implemented without recourse to
some form of qualifiers. We may not all agree how much qualification is
required, but pretty much everyone agrees that some is necessary. The Data
Model Working Group met in the two days following DC-6 and completed the
majority of its agenda, providing a sound foundation for expression of
Dublin Core qualifiers.
Clarification of qualifier semantics is critical to promoting
interoperability. Between now and the next Dublin Core Workshop
(Tentatively scheduled for late October, 1999, in Germany), working groups
will be tasked to review qualifier semantics identified following the
Helsinki Workshop (DC-5) and reconcile these semantics with the Data Model
so that they may be expressed as consistently as possible in any given
A parallel effort will be undertaken by these working groups to gather from
implementors additional examples of qualifiers judged to be useful and to
consolidate such semantics in proposals that will be evaluated and ratified
according to formal processes that will by then be in place. Implementors,
this is your advance warning that such requests-for-input will be made, as
soon as the appropriate working groups are formed and populated.
IV. DC 2.0
The controversial Agent Proposal, released to the community only days before
the DC-6 meeting, sparked considerable debate on the list and at DC-6, and
raised related questions (for example, the wide recognition that Source is
just a variety of Relation).
These issues emerged roughly simultaneously in the data modelling process
and, importantly, from the efforts of independent groups of implementors and
educators trying to teach resource description using the Dublin Core. In
the early days of the DC initiative, when we were thinking of DC primarily
as a set of semantic labels, these structural mismatches were not
particularly troublesome. As we have grown more sophisticated in our
thinking, and as more communities have examined the Dublin Core for their
own use, the importance of a sound underlying structure has become evident.
Reconciling the needs of museums and libraries, for example, has raised many
difficulties that are best solved by sound data modelling.
Recent dialog with what I will loosely describe as the rights-management
community has amplified the importance of such models. This community is
represented in the INDECS project, an EU-funded project to reconcile the
functional metadata requirements of several well established content
communities (including the DOI Initiative). Those of you who read Godfrey
Rust's criticisms of the Dublin Core in last July's DLib magazine will be
pleased to know that Godfrey and his colleagues made valuable contributions
to DC-6, and further, that our preliminary discussions leave both groups
(INDECS/DOI and DC) strongly optimistic that the underlying data models of
the two efforts can be reconciled. This reconciliation will benefit both
communities. I have for some time now been using the notion of the Internet
Commons as a metaphor for the importance of the DC community's
cross-disciplinary focus. This development is the strongest example thus
far of the importance of such cooperation.
A working group has been formed to carry forward this effort jointly. The
expected result of such efforts will be a proposal for restructuring the
Dublin Core and might be thought of as Dublin Core 2.0. It will be largely
backward compatible with DC 1.x. It will probabaly have fewer top level
elements. It will depend on formal qualifiers for refining the semantics of
the top level elements. It will have an underlying data model that will be
accomodate the changes in the electronic environment that cause tension in a
resource description model based in traditional bibliography.
DC 2.0 will provide an evolutionary path for Dublin Core. The great
majority of work on qualifier semantics will find a natural home there. It
will not isolate DC 1.x applications, anymore than HTML 4.0 has isolated
HTML 1.0. It will provide a development path that gives the Dublin Core
room to grow with our own increasing sophistication about metadata.
I personally came away from DC-6 with a good deal of optimism about our path
forward. Each of previous workshops was marked by controversy that sparked
important steps forward. DC-6 was no exception. The major schism in this
meeting lay between those who want stability for what we have done thus far
and those who are reaching for the next level. It is essential that BOTH of
these objectives be met. Fortunately, the great majority of the supporting
tasks for the two objectives lie on a common path. I look forward to
working with all of you to accomplish them.
Nach Frankfurt im Oktober, 1999!
Dr. Thomas Baker [log in to unmask]
[log in to unmask]
AIT - Asian Institute of Technology, Thailand
ERCIM - European Research Consortium for Informatics and Mathematics
DCML - Working Group on Dublin Core in Multiple Languages
Personal : c/o FES, GPO Box 2781, Bangkok 10501, Thailand
Work : c/o AIT-CSIM, P.O.Box 4 Klong Luang, Pathum Thani 12120, Thailand
Home (11-12 hrs ahead of USA) : +66-2-300-3434
Fax ("for Tom Baker") : +66-2-246-7030, voice: +66-2-246-7013
Office at AIT : +66-2-524-5705, message: +66-2-524-5700