background image
DigiCULT 37
On the other hand, practitioners were not sufficient-
ly resourced or trained to participate in such RTD.
Gilliland-Swetland demanded incentives to foster
`the development of testbeds from which generalisa-
ble data, benchmarks, and outcomes can be generated.
More practitioner training/professional education in
this area needs to be encouraged and provided. Fund-
ing initiatives should highlight this as a priority area
to encourage existing researchers to work in this area.'
She thought that, with appropriate funding for RTD
and practitioner education, some good testbed devel-
opments could be achieved around 2008.
The most detailed suggestions of what could form
an RTD plan in the field of `intelligent heritage'
came from Martin Doerr. He is head of the Cen-
tre for Cultural Informatics of the Institute of Com-
puter Science (ICS), which forms part of the main
national research centre in Greece, the Foundation
for Research and Technology Hellas (FORTH).
In order to give as complete a summary as possible
of Doerr's ideas and suggestions, in what follows we
include larger quotes from his three-page contribution.
Doerr started off with the suggestion that what has
been achieved over recent years in technological and
economic terms in enterprise knowledge access via
data-warehouse technology `could be used as an anal-
ogy for large clusters of cultural and other institu-
tions to provide interdisciplinary research nets'. Over
the next 10-15 years, `we could come to a situation,
where large clusters of semiautonomous, complementary
(and not only homogeneous in content) information
sources can be accessed for research, education and
public interest under virtual global schemata of great
generality, which are basically top-ontologies. Knowl-
edge Organization Systems (thesauri, gazetteers etc.)
can be used like Dynamic Linked Libraries as global
resources that allow for relating information from dif-
ferent resources immediately without costly and ambig-
uous data clearing. A kind of Semantic Grid could be
based on a generic architecture of how to link unique
references to events, persons, places, periods, objects,
event types, person roles, object types across servers.
This could effectively connect archives, libraries and
museums on the base of referential networks and not
only on the often meaningless equality of subject.'
Related to this, Doerr thought that `new query par-
adigms fully supporting access to partially complete
knowledge and to answer inductive/deductive ques-
tions could be defined' and also that `content-based
analysis could bring recognition of factual relations
and events to the same level as recognition of con-
cepts and entities'. For feature recognition in images
he expected little progress with respect to fully auto-
matic subject detection; even in 10-15 years' time this
would remain an area `restricted to limited, highly
controlled contexts'.
Asked what currently represents the main RTD
limitations or gaps for sound progress in `intelligent
heritage', Doerr gave the following list:
`(a) Interdisciplinary collaboration... Even though the EC
propagates it, the evaluation is done by domain spe-
cialists that declare any true interdisciplinary work as
non-innovative.' With respect to what was missing in
particular, Doerr mentioned `systematic investment
in semiautomatic approaches to indexing, metadata
and KOS creation and quality control', and `systemat-
ic investigation of the real user questions and the dis-
course structure in science and humanities'.
The Centre for Cultural
Informatics `pursues a
comprehensive, cross-
disciplinary approach to
supporting the entire
life-cycle of cultural infor-
mation and documentation
procedures for the benefit
of the preservation, study
and promotion of cultural
heritage'. It has several cur-
rent co-operative projects
with cultural institutions
and scientists from the
humanities such as, to name
an outstanding example, the
Special Interest Group for
the CIDOC Conceptual
Reference Model (CRM).
Cf. http://www.ics.forth.
gr/isl/cci.html; CIDOC
CRM, http://cidoc.
DCTHI7_271104.indd 37
06.12.2004 8:37:41 Uhr