background image
6 DigiCULT
seem relatively well served, in contrast the expecta-
tions of educational users expose heritage institu-
tions to more fundamental challenges in providing
online access to their resources. Taking the lead from
the entertainment industry, educational users seek
compelling and interactive resource discovery envi-
ronments. The fact that users benchmark their edu-
cational offerings against the industry standards in
the `edutainment' sector needs to be recognised and
designed into heritage online services.
T
hree interviews provide different perspectives
in overcoming the still considerable barriers to
meaningful search and retrieval of heritage resourc-
es. Jussi Karlgren offers a unique approach on how to
recapture the added value provided by librarians and
curators, which he believes has been lost in the digit-
al revolution. He expresses the need for research into
technologies that enable these information and sub-
ject experts to fill the role of information brokers
online. This form of stewardship would ensure that
users have authoritative information and would better
assist them in the formulation of their search needs.
Such support, if performed correctly, can eliminate
the feeling of isolation and prevent search fatigue,
which ultimately results in users accepting unsatisfy-
ing or, worse, misleading results.
Andreas Rauber begins from the standpoint that
there may never be a perfect way to represent a piece
of content, so now may be the time to redirect some
research effort towards new approaches. His approach
is to put objects in an information space rather than
define a representation schema for each class of
objects. He describes the use of an unsupervised neu-
ral system to classify and cluster objects according to
content. The research has also expanded into music
files with clustering techniques able to organise music
collections into genres of similar music. The capacity
to fine-tune the classification is built into the system.
Pia Borlund is a firm believer in user-driven tech-
nological development. Her research focus is to
adapt retrieval systems to the user and not the oth-
er way round. To better understand user needs she has
developed a new method for evaluating information
retrieval systems. Initial trials indicate that the more
cognitively different the representations are that point
towards a certain document, the higher the probabili-
ty that the document is relevant to a given set of cri-
teria an approach she believes will have application
value for the cultural heritage sector.
J
ose Joemon in his contribution begins with the
dilemma that users are poor at formulating a query
that fulfils the needs of a machine-based query sys-
tem. He therefore sees the value in continued devel-
opment of retrieval tools based on personalisation
methods. He presents unique approaches to refine
the search based on implicit feedback techniques. The
techniques include bundling of sources to overcome
loss of context and a query-less interface based on
object selection to define a user's information need.
Whilst not as accurate as explicit feedback, it has
been demonstrated that implicit feedback can be an
effective substitute for explicit feedback in interactive
information seeking environments.
M
ichael Steemson summarises the Forum's
expert discussion, which asked, on behalf
of cultural heritage clients everywhere: `How can
we find something if we don't know it exists?' The
experts were also asked not to forget the backdrop of
powerful commercial services that give the appear-
ance of being comprehensive and offer instant grati-
fication to their users. Consensus was reached on the
need to think beyond existing general-purpose search
engines to tools for different kinds of search behav-
iours capable of finding the rich cultural heritage
resources online. Future research challenges, as a key
focus of the discussion, revealed agreement on the
need to channel effort towards a modular approach
in building user-driven search tools and services, able
to integrate the learning and research behaviours of
users.
T
his Thematic Issue also contains a case study,
which concentrates on how to make use of
Knowledge Organisation Systems (KOS) such as the-
sauri and ontologies in Web-based resource discovery.
Douglas Tudhope and Ceri Binding highlight the fact
that there is a vast existing legacy of such intellectual
systems, and collections indexed by using their con-
cepts and controlled vocabulary, to be found within
cultural heritage institutions. This rich legacy of KOS
makes it possible to offer more intelligent search
options than the current generation of Web search
engines. However, today such KOS are often only
made available on the Web for display and reference
purposes. There is a need to integrate them more
fully in indexing and search systems, in particular
through the development of standardised application
programming interfaces (APIs) and access protocols.
F
inally, regular readers of our Thematic Issues will
find that this time we did not use images from a
heritage institution to illustrate the publication. Rath-
er, Birgit Retsch from the DigiCULT Forum secre-
tariat shot a series of `resource discovery' photographs
at the Archivio di Stato di Roma, where the sixth
expert forum was held.