Page 223 -

206 Chapter 6

Barron (2000) summarizes the current state of the art of EPSSs and related approaches
in the following manner: “ take an e-learning course; chunk it into discrete learning
bites; surround it with technology that assesses a learner ’ s needs and delivers the
appropriate learning nuggets; add collaborative tools that allow learners to share
information. What do you get? Something that looks a whole lot like knowledge
management. ”
The best approach, then, requires a user model or trace — a record of the interaction
between the user and the system. The user model would capture the objects of interest
or focus — that is, what content was accessed, when, how often, in which sequence,
and so on. A log of user interactions can be abstracted to produce a user and task
signature. Together, these will yield a model of the user and the task that the user is
attempting to perform and these two sources of information can help in providing
the best possible support for knowledge application in that particular case. Figure 6.9
illustrates a sample user and task model.
It is assumed that episodes related to particular tasks usually share some common
features or patterns. Once these common features have been identifi ed for a given

Task Characteristics Features
Consequence Inter- Type of
Users Tasks Frequency Difficulty Complexity Desirability
of errors dependencies support
T(1) Weekly Low Low Low Low
U(1) Template
T(2) Monthly Moderate Moderate T(7), T(4) Moderate Moderate
Manager Example
T(3) Quarterly High High High High
T(2)
U(2) T(7)
Technical
T(8)
T(1)
U(3) T(2)
Sales T(3)
T(5)
Help Support Daily High Moderate N/A Knowledge Moderate High
Desk request repositories
Problem
IT Daily Moderate Moderate N/A Manuals Moderate Moderate
report
Tech. Strategic
Research Monthly Low Low — Moderate Moderate
Watch objectives
Strategic Business
CKO Quarterly High Moderate — High High
priorities units
U(n) T(n)
Figure 6.9
Sample user and task model