task, but rather that it is the way in which
these rules and other representational formalisms are themselves represented that
is the limiting factor.
At first a simple rule-base
is relatively transparent, especially if properly documented. Certainly such systems
were easier to comprehend than procedural code and were subsequently easier to
update and amend. As such rule bases became larger and more complex however, a
simple syntax error, perhaps only involving one word, could prevent them from
operating correctly. The complexity of these rulesets also meant that it was difficult
to get an overview of what was intended, thus impeding their maintenance and extension.
Picture is Worth a Thousand Words
problem of rulebase comprehensibility, I would argue, is the fact that we have
primarily represented knowledge using text based structures rather than visual
ones. No matter how close to natural language a knowledge representation language
is, you cannot see at a glance what a complex system is trying to do.
Rule generation via
a graphical interface is a hot topic right now, with offerings from a number of
companies small and large. This is being driven in part by current interest in
so-called 'business rules management' which is arguably a reawakening of the KBS
paradigm we mentioned earlier.
Logic Programming Associates is an appropriate company to enter this market as
it has been producing rule-based software since the mid 1980s. Its latest product,
VisiRule, enables rule-based systems to be automatically generated from a flowchart
drawn on the screen.
Consider the following business rule (Ross 2003):
An order must be credit-checked if any of the following is true:
The order total is more than $500
* The outstanding
balance of the customer's account plus the
order amount is more than $600
* The customer's
account is not older than 30 days
* The customer's
account is inactive
* The customer is out of