Abstract
1. Operational Issues : Pedagogical
Indexation & Pedagogical Retrieval
is
all
1.1 ARIADNE: A Multicultural,
Multilingual Application Context
TM

applications have been developed and are
currently under test in European
Institutions and Organizations. While the
critical mass of resources may take several
years to be accumulated in the KPS, it is
easy to foresee that large scale reuse will
depend on teaching level arbitration,
linguistic equivalency and pedagogic and
semantic similarity seeking mechanisms. The
teaching level arbitration issue has been
partially addressed in Europe by the ECTS 4
Project, but not the basic problem of
semantic and pedagogic interoperability
which is crucial in the context of actual
teaching material interchange.
1.2 Pedagogical Indexation
From the design stage of the KPS, the need
for an adequate set of descriptors for the
training resources under consideration
became obvious. But could an existing
library-originated indexation systems, be
used to describe at least the semantics of
these electronic teaching documents? After
close inspection of the features of the two
most promising thesauri: the Dewey
classification and the Universal Decimal
Classification (UDC), the idea was
abandoned as (i) semantics appeared to
depend on the fact that the resources were
to be used for teaching; (ii) the vocabulary
was far from being sufficiently detailed; (iii)
the organization of knowledge domains
seemed unusual and (iv) the corresponding
thesauri of terms did not exist in
@compatiblei versions in more than one
language. This prompted us to design the
ARIADNE Pedagogic Header, later
recognized as a metadata set'*', to be used
by indexers through the so-called Pedagogic
*** We refer to the system implemented in the current
(Sept. 1998) version of the ARIADNE tools, not to our
"+Educational Metadata Recomrnendation~r a \[see:
http: / /ar iadne.uni l .ch/metadata.htm\] nor to the
IEEE LTSC Metadata working document, based on
ARIADNE-IMS work . \[see:
http: / /manta.readadp.com\]
Header Generator tool.
1.3 ARIADNE Educational Metadata
The current set of descriptors in the
ARIADNE educational metadata is
organized in four categories, two of which
need little elaborationt: he so-called general
and technical categories.
The general category regroups information
about the origin, availability and metadata
authoring of the resource. Most of the
current descriptors of the Dublin Core can
easily be mapped into it (with the notable
exception of the subject). Interoperability
will stem from the standardization of a
conventional set of descriptors.
The technical category, absent from the
Dublin Core and other classifications, is
obviously needed when dealing with
electronic resources that may run on a
variety of platforms and operating systems,
may or may not include multimedia
components, may or may not exist in a
compressed format and may have to travel
long distances on networks of unspecified
bandwidth. Thence the various descriptors
that belong here. Interoperability is here at
its usual, computer-related interpretation.
The two other categories address the core of
our problem#: the so-called semantic and
pedagogical characteristics. Leaving aside for
the moment the issues related to multi-
linguality, let us consider the descriPtors
that have been chosen and their
implications on interoperabilities of various
kinds.
2. Semantic
Mechanisms
Interoperability
The semantic characterization of a
"tlearning object-t a relates to the knowledge
(or the know-how) field in the context of
which the teaching or training action is to
take place (rather than takes place)#: we
should mentally adopt here the attitude of
the indexer who guesses at the kind of
educational use that could be made of the
SIGMOD Record, Vol. 28, No. 1, March 1999 21