Web-based Learning Solutions for Communities of Practice

tion of such paths delivers a longer path that can
still be a learning experience.
The graph-theoretic model also allows us to
build in temporal information in the navigation
paths. As a matter of fact, relative temporal information
is inherently available in a path (sequence
of node visits). Furthermore, the annotation of
edges in terms of actual time spent in an activity
before moving on to the next is a straightforward
enhancement.
The detour ends here by noting that the above
considerations simply suggest that, after we get
the initial graph-theoretic model fixed, there exist
a set of computational processes that will allow us
to define arbitrarily complex layers of information
based on the ground data. We elaborate on that in
a following section.
DISCUSSION ON LEARNING DESIGN
The main purpose of the above schema is not to
propose a new learning design and is certainly not
intending to form a new specification. We are using
the idea of a graph as the foundation of a learning
environment and therefore we need to identify the
modules that can be the nodes of the graph, given
that a student will need certain resources and use
certain material. We simply define the modularity
of the material and the resources in accordance to
the activities a learner will perform, so that we can
clarify what the nodes in the graph will be.
There are no expected outcomes and objectives
that necessarily have to be met and that will define
the future progress of the course. Every activity is
intended to offer specific gains to the student, but
they are not explicitly measurable and the student
is free to navigate and choose his own course of
studies. That is exactly what we want to examine,
how the user forms his path and not whether he
can be successful or not in a predefined one where
the structure determines the path. We are more
concerned in pinning down the modules that a
learner might need for a complete educational ex-
60
Using Graphs in Developing Educational Material
perience and we want to observe his walk through
that graph. His path will reveal the usage of the
resources and whether the proposed granularity,
structure and assembly of different resources is
used to its full extent and is assistive or whether
the users’ patterns reveal a different structure and
a behavior that was not foreseen.
On a first look there is a similarity to the
IMS learning design model and in specific to its
conceptual model. In the IMS design information
model (IMS Global Learning Consortium, 2003)
website, the specification includes learning activities,
a learning environment, learning objects but
they are explicitly defined for a specific model. For
example the environment in IMS is defined as “…
a structured collection of learning objects, services,
and sub-environments. The relationship between
an activity and an environment can be derived
from the linguistic description of the activities.
…”, whereas we consider it to be an underlying
graph connecting learning activities. We can see
a resemblance in the definition of learning activities
and learning objects, however we use a more
plain definition since we are not trying to capture
a wide variety of pedagogical concepts. Our main
focus is to the process followed by users and not
the design we offer to them.
TOOL DEPLOYMENT ISSUES
We start by noting that the theoretical model can
be in principle implemented using rudimentary
technology, such as hyper-linked files of conventional
office-type applications, where educational
assets can be grouped together in repository-type
worksheets. Assets can then be drawn to compile
learning activities. Such tools offer relatively
smooth short learning curves for data collection
and web publishing too.
As an example, Figure 1 shows how MS Excel
could be used to design a learning environment.
A learning activity is composed by an asset and
by a learning task (allowing, of course, for some