Collaborative Learning through Flexible Web CVE - Teaching through Multi-User Virtual Environments - page 472

Previous Approaches the system (no other users, or nobody playing

Leonardo in the virtual environment), the con-
nection time of the users was very low (typically
below five minutes). On the contrary, when forms
of collaborations where enacted, users remained
connected an average of 53 minutes, even engaged
on a single topic (Figure 1).
We had therefore the evidence that collabora-
tion within virtual environments had a clear,
outstanding and unexploited potential for attract-
ing attention and interest of a vast range of par-
ticipants, regardless of their level of education,
sustaining their attention for a long time.
“WebTalk-II” was the term for an initial frame-
work project whose goal was to tackle the issues
described above, by drafting a generic architecture
that was to be the foundation for all subsequent
research and development activities. The idea
was to lay out the groundwork for terminology
and architecture on which to coordinate all future
efforts and to offer a schematic vision of all the
features and main components of a possible “total
collaborative system”. The conceptual framework
describes two main abstract concepts for the man-
agement of the shared space: the Scene Builder
and the bi-dimensional GUI.
In conclusion, the WebTalk-II part of the re-
search was fundamental to organize the application
domain of Collaborative Virtual Environments
and to set the path and the direction of future
implementations and applications.
The experience with “ Virtual Leonardo ” and
the foundational work made with the WebTalk-II
framework enabled the team at the Polytechnic
of Milan to move toward an educational envi-
ronment: the idea was to exploit the potentiality
for collaboration (among students now) to the
maximum extent.
The occasion came with SEE, Shrine Educa-
tional Experience (Di Blas et al., 2003), devel-
oped in partnership with the Israel Museum in
Jerusalem. SEE is about the cultural, religious and
historical issues related to the Dead Sea Scrolls,
found in the Qumran (near the Dead Sea), in Israel.

Our first original aim was to build a virtual museum
where several visitors could “go together”. An
application that allowed a virtual visit to the “ Na-
tional Science and Technology Museum ” (Paolini
et al., 1999; Barbieri, 2000; Barbieri & Paolini,
2001a) of Milan, was for some time available to
the public, through the web site of the Museum
itself. It hosted a 3D virtual exhibition on the
machines “invented” by Leonardo Da Vinci or,
more precisely, of wood machines built according
to the drawings left by Leonardo. The application,
award winner at the 1999 Museums and the Web
international conference in New Orleans, allowed
the exploration of a building, vaguely representing
the actual museum. A “guide”, playing the role
of “Leonardo”, had the task of guiding visitors
through the virtual rooms. Virtual objects on
display were either reproductions of Leonardo's
machines (on display in the museum) or gate-
ways to web pages on the museum's website.
The reproduction of the machines, rather than
being realistic, was playful (also because most
of Leonardo's machines do not work in reality).
The prototype, called WebTalk-I and written in
1998, at the HOC laboratory of the Polytechnics
of Milan, was a client/server framework entirely
written in Java and VRML: it required a VRML
Browser and a Java Virtual Machine to run a Java
applet within the browser. The browsers at that
time (Netscape 4 and Explorer 4) had Java and a
VRML plug-in installed by default, therefore the
application was easily available for most PC users.
After six months of experimental usage, it
was evident that cooperation (a loose one) among
visitors was successful. Finding someone “there”,
visiting the museum from a far away place, was
an exciting experience: both if the meeting was
“prearranged” or it happened by chance. The
virtual museum, instead, by itself, was not very
attractive to the users. One important highlight
from the data we collected at the time, was that
when there were no collaboration activities within