Page 120 -

Knowledge Capture and Codifi cation 103

involves reducing a vast volume of content from diverse domains into a precise, easily
usable set of facts and rules.
The idea of acquiring knowledge from an expert in a given fi eld for the purpose of designing a
specifi c presentation of the acquired information is not new. Reporters, journalists, writers,
announcers and instructional designers have been practicing knowledge acquisition for years
. . . system analysts have functioned in a very similar role in the design and development of
conventional software systems. ( McGraw and Harrison-Briggs 1989 , 8 – 9)
The approach used to capture, describe, and subsequently code knowledge
depends on the type of knowledge: explicit knowledge is already well described, but
we may need to abstract or summarize this content. Tacit knowledge, on the other
hand, may require much more signifi cant up-front analysis and organization before
it can be suitably described and represented. The ways in which we can tackle tacit
knowledge range from simple graphical representations to sophisticated mathematical
formulations.
In the design and development of knowledge-based systems, or expert systems,
knowledge engineers interviewed subject matter experts, produced a conceptual model
of their critical knowledge and then “ translated ” this model into a computer execut-
able model such that an “ expert on a diskette ” resulted (e.g., Hayes-Roth, Waterman,
and Lenat 1983 ). The global aim of such systems was to extract and render explicit
the primarily procedural knowledge that comprised specialized know-how — typically
in a very narrow fi eld. Procedural knowledge is knowledge of how to do things, how
to make decisions, how to diagnose and prescribe. The other type of knowledge,
declarative knowledge, was used to denote descriptive knowledge or knowing what as
opposed to knowing how . It soon became apparent that certain types of content were
easily extracted and modeled in this manner — anything that was similar to an interac-
tive online manual or help function in such fi elds as engineering, manufacturing,
decision support, and medicine.
A wonderful by-product of the work in artifi cial intelligence was the array of inno-
vative knowledge acquisition techniques that were created. The interactions with
subject matter experts that were needed to render tacit knowledge explicit made up
the knowledge engineer ’ s toolkit. Quite a few of these techniques are imminently
relevant and applicable to the process of tacit knowledge capture in knowledge
management applications. The major tasks carried out by knowledge engineers
included:
• Analyzing information and knowledge fl ow
• Working with experts to obtain information
• Designing and implementing an expert system