background image
DigiCULT 39
tion; categorical knowledge in science, biology, ethnology, geol-
ogy, paleontology, sociology etc.'
Doerr added an important note in which he
addressed how the predominant notion of a Seman-
tic Web needs to be adapted for `intelligent herit-
age' sources and knowledge providers: `Museums
and scholarly researchers are curators of knowledge,
and not administrators such as libraries. Whereas the
Semantic Web paradigm assumes that interaction
with the source providers is not possible/not scala-
ble, in the cultural scenario the source providers have
a strong interest to improve their sources and work
on specific scientific targets. Therefore they can be
mobilized to produce federated systems with very
high data quality standards. The goal must be to cre-
ate larger and larger bodies of integrated knowledge
without losing their validity and truth warranty. Har-
vesting scenarios are only auxiliary tools as first find-
ing aids.'
Further ideas and suggestions on this issue are to be
found in the chapter on Natural & Enjoyable Interac-
tion (p. 45).
he following table gives a condensed overview
of what the experts thought could be achieved
in this RTD area over the next 10-15 years. After a
short summary of what they considered to be the
current limitations or barriers, the experts' suggestions
are grouped into the phases 2005-2009, 2010-2014,
2015 and beyond. The timeframes and, where given,
years indicate when a certain methodological and/
or technological gap could be closed or some oth-
er RTD breakthrough be achieved. These assessments
are of course dependent on the condition that appro-
priate funding levels, RTD collaborations and other
requirements are met.
2004: Current limitations/barriers
Some of the major present limitations are very well
summarised by Kirk Martinez as `immature ontol-
ogies and relatively few users of it; difficulties get-
ting a unified approach to using it and building
semantic interfaces to digital collections in muse-
ums, libraries etc.'
In particular, there is a need
to further develop domain, middle- and top-
level ontologies;
to employ new conceptual frameworks (e.g.
to evaluate how these work with descriptive
standards such as FRBR, EAD and TEI;
to provide detailed documentation of emerging
`good practices' in real deployments;
to develop semantic interfaces that provide
strong evidence of the cost/benefit ratio of
intelligent heritage in terms of improved access
to, and use of, knowledge resources.
Overall: A phase of further development in theories and
methods, and first convincing results
A stronger focus on harmonisation of (domain)
Better harvesting, clustering and visualisation of
descriptive information (metadata)
Advancement in the engineering of top-level
ontologies with wide applicability
Further development and testing of semantic
search tools and applications
2007: Theory: `Open World' query languages and
systems (cf. M. Doerr, p.38)
2008: First results: Some unified or harmonised
top-level ontologies of relationships (and class-
es) for descriptions of heritage, Arts & Humanities,
and other disciplines' knowledge resources
2008: First steps in making ontologies usable for
truth warranty in integrated systems (that would
need to consistently relate factual and categori-
cal data)
2008: Further developments in semantic databases,
maintenance algorithms and query systems
2008 (ongoing): High demand for theoretical
advances in coherent ontological frameworks (e.g.
relations between classes meta-classes meta-
2008: First results (M. Doerr: if addressed strategi-
cally, otherwise maybe never): Analysis of real user
DCTHI7_271104.indd 39
06.12.2004 8:37:42 Uhr