The author describes some of the classic problems involved
in formulating a knowledge base. For instance, the choice of
a domain is difficult where the problem is too complex or
experts disagree. In addition, acceptance is difficult to
obtain universally across many organizations. Uncertainty
is always a factor in formulating an expert system. Updating the
knowledge base is another issue because the current applicability
of facts and data will always be a critical concern. Every
expert system has some limitations. These need to be set forth.
The development and refinement of algorithms involves the
constant induction and perfection of rules and rule structures.
In a structured approach to induction, it is critical to identify
problems and attributes while optimizing the rule structures.