background image
16 DigiCULT
That way the object is linked to the different inter-
pretations rather than having the different interpreta-
tions assigned to the object.'
The procedure struck a chord with Douglas Tud-
hope, a researcher in hypermedia in the School of
Computing at the University of Glamorgan, Wales,
working with processes like interactive and automat-
ic query expansion. What he called the `raw object
and its metadata' would come from the museum col-
lection, he thought, `But then, it is the act of creating
this secondary information object which has got the
context and some perspective. People could add to it
and create layers. More attention could be given to
the tool to create this and to what metadata is need-
ed for this secondary information object.'
The conversation seemed to be getting a bit too
lofty for the likes of the Forum's sole woman expert,
Denmark's Dr Pia Borlund, an associate professor
in the Information Studies Department at the Roy-
al School of Library and Information Science in Aal-
borg. She wanted it brought down to earth again,
suggesting: `One thing we are also considering is
whether we want to go for ideal or what is prac-
tical and possible.' Her field of expertise is evalua-
tion of information retrieval and interactive systems
including, as she put it, `users and their use of infor-
mation retrieval systems leading to information-seek-
ing behaviour'.
But Moderator Jose had, apparently, rather liked
the turn of the debate and suggested: `Curators want
what the current technology allows them but we also
need to look beyond what the technology allows.'
Seamus Ross liked Dr Rauber's `information
space' process, too. Dr Ross is Director of the Uni-
versity of Glasgow's Humanities Advanced Tech-
nology and Information Institute (HATII) and the
EU's Electronic Resource Preservation Network
(ERPANET). He recalled undertaking a study in
the British Museum for which nineteenth-century
records had included drawings and a curator's notes
but little provenance. There had been nothing with
which to relate the object to material found in a sim-
ilar context and it had been very difficult to come to
an understanding of it. However, things improved, Dr
Ross remembered.
`Over time, by relating that particular object to
others we are able to build up a metadata picture of
the kind of contexts that it had to come from. We
attached the object to a whole set of metadata that
we built up about other objects of similar categories
and similar types. We need to recognise that metada-
ta evolve and in fact our ability to trust in their rele-
vance changes over time.'
K
NOTTY
METADATA
STANDARDS
T
hen the discussion turned to questioning a
requirement for standard metadata thesauri. Ital-
ian Paolo Buonora bemoaned the fact that `in Ita-
ly, the only point that common metadata committees
agree is administrative and management metadata. We
can not arrive at any common point about a descrip-
tion of a cultural object'. Mr Buonora is director
of the Rome State Archive's Image II digital imag-
ing project in whose reading room the Rome Elev-
en cogitated.
He was not, however, sure that common metadata
mattered. As he said: `The reasons why historians and
researchers in general love to come to the archives is
that they never find exactly what they were looking
for. All we need is a map so we do not get lost.'
Austrian Dr Rauber also wondered if the idea
of a common metadata model was even attaina-
ble let alone necessary. Could existing technologies
not adequately link differing representations of a sin-
gle object, he wondered, adding: `I am still pretty sure
that a lot of the tasks that we want to solve can be
done with state-of-the-art technology and without
investing great sums. So a lot can be solved with very
crude automatic approaches to a very high degree of
user satisfaction.'
He had a formula for the thesis: `My analysis leaves
me to ask whether they would not take us faster
towards quite reasonable results that may serve 80
per cent or our needs, and leave the remaining 20
per cent for discussions that we will probably never
be able to resolve anyway.'
Dr Karlgren was at odds with his 80 per cent fig-
ure, saying glumly: `I would say five per cent can be
structured with the technology we have today.' But
he readily acknowledged: `This is a rapidly evolving
field, finding contexts in texts, for instance, or recog-
nising objects in pictures or events in video frames.
This is something a lot of people are working on.
`Of course, most are working with national securi-
ty concerns or something, not cultural heritage. But
this stuff will be percolating into our fields, so we
can be certain that in 15 years from now there will
be fairly good video parsing sequences that are able
to extract information from video frames quite well.
And, we can use them to structure the information.
All we need to do now is to say that those things will
be available and now we do not need to draw a com-
plete metadata set.' Here, Traugott Koch preached
caution, warning: `We should be careful with our
present discussion not to be drawn into conflicts
about the view on metadata.'