- From: Thomas Baker <thomas.baker@izb.fraunhofer.de>
- Date: Thu, 1 Apr 2004 19:53:07 +0200
- To: SW Best Practices <public-swbp-wg@w3.org>
Dear all, I joined the group yesterday and will attend the teleconference twenty minutes from now. I posted this Introduction three or four hours ago, or so I thought until I saw I was using the wrong email address, so here it is again... I have been involved since 1996 with the Dublin Core Metadata Initiative and currently serve as Chair of the DCMI Usage Board, which means I am also a member of the DCMI Directorate. I also helped start the DELOS network in Europe and have been involved in various European projects over the years. My background is in the social sciences. I grew interested in the Internet in the 1980s while at Stanford (and on The Well). I tend to approach things like metadata from a linguistic angle and can take some credit for suggesting the notion of Dublin Core as a pidgin. Deep down, I suspect that pidgins or cores, fuzzy and lossy as they are, represent the best we can reasonably expect in terms of "semantic interoperability" on a really broad scale. To me, DCMI is about making the pidgin idea work -- with a core vocabulary; a process for growing the vocabulary; policies for declaring and versioning the vocabulary; an etiquette for playing well with complementary vocabularies; conventions for documenting the vocabulary; a simple grammar; and methods for dumbing down to a Core. In the spirit of "walk before we run" (danbri), then, I am most interested in the plain-vanilla issues around managing and using small vocabularies. Some issues of possible interest to BPD WG: 1) URI policy: DCMI has a formal "namespace policy" [1], and the CORES Project brokered a "first-step" agreement among maintainers of some key standards regarding such policies, raising more general issues [2]. How far could BPD WG go in formulating best practice in this area? 2) Versioning terms and term sets: DCMI has a de-facto method for versioning terms, though it is not yet formally supported by DCMI policy. (It is an event-based model which uses URIs to link changes in Term Versions to Decisions, which in turn are linked to supporting documentation.) Is the model good and might it be generalized? 3) Assertion etiquette and "good neighbor" policies: DCMI is working with Library of Congress on developing an RDF schema in which LC asserts a set of MARC Relator terms to be subPropertyOf dc:contributor. DCMI wants to then endorse these assertions. Might we consider such formalities in BPD WG? 4) Vocabulary documentation (see also Dan [3]): DCMI is looking to the SW community for guidance on what to publish at its namespace URIs (it currently publishes RDF schemas). In terms of work flow, DCMI generates the RDF schemas along with ready-reference Web pages from a common source using XSLT scripts, though surely more sophisticated editing and validation environments are available. 5) Declaring versus reusing (see also Libby [4]): "Should I use an existing term, get DCMI to declare one, or declare my own? Where would I put it? Should I make an Application Profile -- ...whatever that is?" 6) Syntax and interoperability. In the DCMI context, Andy Powell and others have developed an Abstract Model for clarifying the extent to which different syntaxes (e.g., XMTHL vs XML Schema vs RDF) support distinctions between multiple entities or between URI-identified "resources" as opposed to "string values" [5]. Lots of issues there... 7) Scalability and complexity: From DCMI one can see how much effort it takes to maintain, grow, document, and explain a vocabulary of just 90 terms. As DCMI moves beyond the start-up phase, we hope this will turn into a well-oiled routine. However, it does raise the more general issue of how much effort "best practice" will ultimately cost, and how that effort will scale to vocabularies and ontologies much larger and more complex than Dublin Core. Tom P.S. I would be most grateful if someone could send me an archive of this mailing list in the native mbox format. [1] http://dublincore.org/documents/dcmi-namespace/ [2] http://www.dlib.org/dlib/july03/baker/07baker.html [3] http://lists.w3.org/Archives/Public/public-swbp-wg/2004JanMar/0016.html [4] http://lists.w3.org/Archives/Public/public-swbp-wg/2004JanMar/0017.html [5] http://www.ukoln.ac.uk/metadata/dcmi/abstract-model/ -- Dr. Thomas Baker Thomas.Baker@izb.fraunhofer.de Institutszentrum Schloss Birlinghoven mobile +49-160-9664-2129 Fraunhofer-Gesellschaft work +49-30-8109-9027 53754 Sankt Augustin, Germany fax +49-2241-144-2352 Personal email: thbaker79@alumni.amherst.edu
Received on Thursday, 1 April 2004 12:50:52 UTC