[sc34wg3] Dublin Core Abstract Model and TMDM mapping

[Dipl.-Wirtsch.-Inf. Lutz Maicher]

I have written a paper about the mapping of the Dublin Core Abstract
Model (DCAM) and the TMDM. The paper "Mapping between the Dublin Core
Abstract Model DCAM and the TMDM (working draft)"  is available at [1].
What is the purpose of this paper?

The current approaches for standardizing the usage of the DC
vocabularies in Topic Maps are really needed and valuable. But what is
in standardization at the moment? Something like a _modeling
methodology_. It will be defined, how metadata in Topic Maps will be
represented by using the DC vocabularies. Having and propagating this
modeling methodology we can assure, that topic maps created at different
occasions become *mergeable*. With this methodology we assure the
*interoperability within the TMDM*.

This insight directly leads to a requirement for the standardization of
the methodology. In general, the DC vocabularies do only define terms
which will be used as keys of properties (or as types, taking the TMDM
glasses). As consequence, applying a common methodology will assure,
that the typing topics will be merged. This will have great benefits for
applying standardized TMQL queries to such metadata records, but much
more power will be achieved when also the values of the properties
become merged. This will happen, when the standardized usage guidelines
will also define some common rules for the values. One example is the
usage of the DC-typed for the dc:type property. Others for dates,
mime-types, etc. are easy to imagine.

But coming back to the purpose of the paper above. With the current
standardization we only assure the interoperability within the TMDM
(which is quite valuable). But by definition, the DC vocabulary is
specified free of any syntax. The Dublin Core Abstract Model (DCAM)
defines the metamodel for the vocabularies, and in general metadata
should be internally represented as DCAM instances, so called
descriptions. These descriptions will be serialised in different
representation formats, like HTML, RDF, XML or even Topic Maps by using
so called "encoding guidelines".

(In Topic Maps the TMDM applies a similar approach. Each topic map is
represented as a TMDM instance and can be serialised to all Topic Maps
notations where the serialization is standardized.)

Having the DCAM as intermediate, we can assure a interoperability of DC
metadata between different representation methods, like RDF, HTML, and
even Topic Maps. Having a TMDM-->DCAM mapping, we can represent any DC
metadata in a topic map as a DCAM instance (and transform it further to
RDF, etc.), and having the DCAM-->TMDM mapping we can publish all DC
metadata (i.e. gained from RDF, etc.) as topic map.

The paper I introduce with this email defines both, a TMDM-->DCAM and a
DCAM-->TMDM mapping. These mappings will assure the desired
interoperability. On top of these mappings, the "usage guidelines for DC
vocabularies in Topic Maps" should be defined, as it is already done in
the standardization process. Applying these usage guidelines will assure
that the created topic maps will be mergeable with other topic maps
(where the guidelines are applied while authoring) and the created topic
maps will always be mappable to a DCAM description set.

Greetings from Leipzig,
Lutz Maicher

[1]
http://www.informatik.uni-leipzig.de/~maicher/publications/DCAMandTMDM_latest.pdf

-- 
______________________________________________

Dipl.-Wirtsch.-Inf. Lutz Maicher
Abteilung Automatische Sprachverarbeitung
Institut für Informatik | Universität Leipzig
Johannisgasse 26 | 04103 Leipzig

phone: 0049 - 341 - 97 - 32303
fax:   0049 - 341 - 97 - 32299
mail: maicher at informatik.uni-leipzig.de
http://www.informatik.uni-leipzig.de/~maicher/
http://www.informatik.uni-leipzig.de/~tmra/

   ======================================
   TMRA - International Conferences on
   Topic Maps Research and Applications

   TMRA 2007 "Scaling Topic Maps"
   http://www.tmra.de
   ======================================