background image
tudies have shown that 3D interfaces
can be more efficient and more power-
ful than classical 2D interfaces at managing
overlapping windows.15 This efficiency can
be reached when objects and navigation/
interaction metaphors are designed depend-
ing on the particular context in which they
should take place. For 3D digital libraries,
there is no reason to reproduce faithfully a
real library building and the various steps
needed to select and read a book in real
life.16 The only significant positive point in
this `real' to `virtual' mapping is to ease the
recovery of books for users who are famil-
iar only with the real library. However, for
users who are familiar with interaction on
the Internet, this approach doesn't help. We
believe that it is more appropriate to pro-
vide abstract navigation and representation
and offer services like reordering the collec-
tion in space depending on search criteria.17
n agreement with A. Manguel, who
thinks that a book is distinguishable
from others by its cover or shape as much
as by its authors or title,18 we decided to
use pictures of books to offer visual infor-
mation. Indeed, it is widely known that
libraries' patrons can discover interesting
bibliographical information by casual glanc-
es at the book collections. This `visual heu-
ristic' is only one example of the `many
complex information tasks [that] can be
simplified by offloading complex cogni-
tive tasks onto the human perceptual sys-
bstract navigation is also important
to minimise moves in a 3D context
and therefore ease the training of inexperi-
enced or occasional users. In a real or `vir-
tually real' library, moves are often necessary
to find the shelves containing the book of
interest, to pick it up and to go back to a
reading desk to work on it. These steps can
be simplified in virtual worlds.
The research tool
In the book selection process, a 3D inter-
face avoids browsing through long lists of
textual information and also offers addi-
tional visual information via images of the
books themselves. With the right metaphor
to represent the collection, a 3D interface
also makes it possible to connect browsing
in the library with the widening or refine-
ment of search criteria. We set up an `on
the fly' generation of a new way of organis-
ing the library at each user's request. To do
so, books are classified in different groups:
those that are relevant to all the search cri-
teria, those that answer some criteria but
not others, continuing to the group that
are relevant to none of the criteria. We
then organise the groups geometrical-
ly and offer the user the means of switch-
ing from one group to another, either to
refine or to widen their query. This manner
of building and submitting the results to
the user enhances the `visual heuristic' for
finding books. Indeed, it makes it possible
to gather books
not relevant to all criteria
besides the relevant group, allowing users
to see an interesting or complementary
book just by taking a quick glance around.
This serendipitous discovery of books is
often employed in real libraries but is not,
unfortunately, possible with standard dig-
ital libraries.
ne of our main concerns beside usa-
bility is the flexibility of our inter-
face. It is therefore necessary to allow
display using different representations. For
example, we could show a cylindrical or
sphere shaped collection as well as other
metaphors like semantic maps, cone trees,
Fig 2: The cylinder-shaped collection
15 See S. Card, G. Robertson and W. York, `The WebBook
and the Web Forager: An Information Workspace for
the World-Wide-Web' in Proceedings of ACM CHI`96,
Vancouver, April 1996.
16 For a comparative view, see E. Fox, N. Kipp et al.,
`Networked Digital Library of Theses and Dissertations:
An International Effort Unlocking University Resources'
in D-Lib Magazine, September 1997, and P. Plénacoste, E.
Lecolinet, S. Pook, C. Dumas and J-D. Fekete, `Zoomable
and 3D Representations for Digital Libraries' in
of IHM-HCI 2001, Lille, September 2001.
17 See P. Cubaud, C. Thiria and A. Topol, `Experimenting
a 3D Interface for the access to a Digital Library' in
Proceedings of ACM DL'98, Pittsburg, USA, June 1998.
18 A. Manguel, A History of Reading (Viking Penguin
Group, 1996).

19 M. Hearst, G. Kopec and D. Brotsky, `Research in
support of digital libraries at Xerox Parc. Part II',
DLib Magazine, June 1996.