background image
36 DigiCULT
Kirk Martinez (University of Southampton, UK)
expected for ontologies in the next 10-15 years `a
more common use of relatively few ontologies (so
that interoperability is easier) relating to cultural her-
itage (e.g. CIDOC CRM but others are need-
ed too!)'. This would also be one of the cornerstones
needed for a European Semantic Web based search
engine for cultural heritage in member states of the
European Union and beyond.
Martinez saw major current limitations as `imma-
ture ontologies and relatively few users of it; dif-
ficulties getting a unified approach to using it and
building semantic interfaces to digital collections in
museums, libraries etc.' To kick-start the next level of
RTD in these areas, funding needed to be reserved
for some key (and maybe risky) projects. Good doc-
umentation of emerging standards for use in real
deployments would help to achieve a more unified
approach throughout the heritage sector. Martinez
noted, furthermore, that there is currently `little re-
use of technologies made in one project software
is written and dies/is archived rather than used. The
Open Source community shows us the way to keep
the ball rolling on this area.' He urged that reinvent-
ing the wheel in each project should be stopped, by
sharing maintained Open Source components.
A participant from the archival field noted that
`work on ontological harmonisation around a robust
conceptual reference model is only beginning (i.e.
CIDOC CRM is now starting work on harmonisa-
tion with FRBR, EAD and TEI
73
). When this work
is complete we can proceed to testing this as a possi-
ble strategy for deeper query searches... In the next
five years we should begin to have a good idea of
whether this sort of logic will lead us in the right
direction.' Besides ontology harmonisation and testing
for good data-mining results, visualisation and clus-
tering would also need to receive considerable fund-
ing in order to develop innovative forms in which to
represent the results from deep query searching.
Paul Mulholland from The Open University (UK)
was convinced that applications using future semantic
frameworks could also support new forms of mean-
ingful exploration of resources, beyond current search
& retrieval interfaces. However, he currently saw a
clear lack in `the right semantic interfaces for explo-
ration', e.g. for future ambient access spanning the
use of different devices in different contexts.
Anne Gilliland-Swetland (Department of Infor-
mation Studies, UCLA, USA) wanted over the next
10-15 years to see a lot of new work `to enhance
information retrieval of cultural materials that go
beyond the limitations of classic approaches in par-
ticular, exploiting the richness of archaic, and alterna-
tive language usage, and different notions of context',
and this also should lead to `new ways of developing
virtual exhibits of cultural heritage materials and of
facilitating classroom learning'.
The main RTD gap Gilliland-Swetland saw in
today's dominant form of technology development
where, for example, information retrieval showed
little interest in the `particular challenges offered by
cultural, other language, and historical materials'.
`Knowledge preferenda'
Michael Moon, CEO of GISTICS Inc., provi-
ded DigiCULT with the following model of
future collaborative knowledge creation, sharing
and delivery: `Each individual represents a mul-
ti-dimensional user type; each user type represents
a unique hierarchy of "need to know" catego-
ries of knowledge or "preferenda". In the future,
individuals will "subscribe" to a variety of speci-
fic knowledge-domain agencies; each agency will
model the entire contents of the Web according
to its domain-specific taxonomy (hierarchy of
preferenda); a community of practice will create
and maintain each domain-specific taxonomy and,
therefore, the quality of search-results "outputs".
Thus, each serious user will subscribe to a mana-
ged commons and community of practice; each
subscription will constitute a "standing request"
for relevant inputs of a particular domain. As in
the iPod nation where "you are your playlist", in
the managed knowledge commons "you are your
standing request with a tree of knowledge". One's
subscribed profile for need to know will define
your information use pattern and, therefore, your
role and responsibility as a knowledge worker in a
post-information society.'
RTD for realising this scenario would need to
concentrate on `collaborative tools for auto-clas-
sification of "dark content" (text or image with
robust metatagging) as well as good garbage (for
garbage in, garbage out) in the form of robust-
ly characterized content and assets'. Also necessary
are communities of shared concern or knowled-
ge who develop `a global/federal "content-data
model" by which to define, organize, and type the
info preferenda'. In future information exchange
environments for nomadic users, peer-led learning
circles would interact via mobile multimedia mes-
saging or mobile broadband.
73
Functional
Requirements for
Bibliographic Records
(FRBR), http://www.
ifla.org/VII/s13/frbr/frbr.
htm; Encoded Archival
Description (EAD), http://
www.lcweb.loc.gov/ead/;
Text Encoding Initiative
(TEI), http://www.
tei-c.org
DCTHI7_271104.indd 36
06.12.2004 8:37:39 Uhr