Knowledge Management (KM) is a field that began as a means to consciously
collect, manage, and share the key knowledge of many people in a systemic manner. By bringing
common ad hoc activities into a more systemic process, the goal is to provide both wider sharing
of useful knowledge among large groups of people and to pinpoint the knowledge people need
during their work to greatly improve their productivity and effectiveness.
Knowledge differs from information and data in a critical way. Data is
transactional information which, when presented independently, provides little contribution
to decision-making and action-taking. In contrast, information is content that is necessary
for people to have in order to perform their jobs. Information is typically data that has
been assembled in some meaningful way. Finally, knowledge is a step more value-added than
information in that knowledge is content which contains user experiences around it.
Knowledge provides insights that move the user of that knowledge to make decisions and take action.
The same text item can simultaneously be data, information, and knowledge
depending on the situation's context and the person's experience and perspective. The distinction
depends on prior knowledge, and task focus. For example, the phrase “Set the soldering iron to 200 degrees”
can be information from manual for general use or knowledge from expert for specific manufacturing process.
Similarly, the phrase “10000 units shipped yesterday” can be simple data for logistics, information
for a shipping manager, or knowledge for competitor monitoring market share.
As with all large scale management fields, KM must overcome the greatest
single challenge to effective management, namely, organizing a very large amount of continuously
increasing information into a small enough group that a variety of different people find useful
and easy to use while being up-to-date and accurate. A big part of this challenge can use the
same good management practices of any large office environment to distribute tasks and
use teams to define requirements. Yet, the remainder of the challenge requires activities and
tool specific to the nature of KM.
Knowledge differs from simple information or data since it conveys the
context, timeliness, confidence, and relationships among the individual pieces of
information. In particular, the contextual and confidence nature of knowledge is the fuzzy
difference between when words can change from being information for one person to knowledge
for another. People "know" something when they believe confidently that they can use the
knowledge in a specific context to make decisions or take actions.
Yet, this need for context and confidence is what makes using KM in
practice so difficult since people "know" things at a personal level. Trying to understand and
capture the personal views and needs of everyone in a group is extremely difficult, but doing that
and then packaging and tagging the bits and pieces for others to find and use at a later
date is almost impossible. KM as a management method therefore does not try to get all
knowledge but only the key knowledge that is relevant and useful to the organization, most
people, and in the most important activities. Even with this distillation and prioritization, it
is a daunting challenge to package and tag knowledge into understandable, findable, and
reusable units. Solving this challenge requires good housekeeping, like organizing a warehouse or
even a kitchen pantry.
The way this is done is to define the way you want to organize
things (e.g. by type, age, cost, etc) and the titles you will use to group
like items (e.g. fruit, meat, dairy). The way to organize things is defined by an ontology which is
a conceptual map of the main ideas and the relationships among the ideas. Once this conceptual map
is made a set of titles can be created within each idea to be more specific. This structured set of
titles constitutes the taxonomy. This paper describes the basics of ontologies and taxonomies for
KM and how to develop and implement them.
Ontologies and taxonomies provide a structure to the concepts and
language used to organize knowledge. Without them, the knowledge will inevitably be difficult to
find and reuse as people have very different perspectives on how the knowledge is related in the
context of their situations.
Ontologies specify the primary concepts and the relationships among the
concepts in a particular domain. The term means several things depending on the field in which
it is used. In philosophy, ontology is concerned with the metaphysical nature and relationships of
being. In contrast, computer science uses ontologies to describe specific conceptual terms and
relationships in a standardized machine readable format. Any KM effort must grapple with the
challenge that there are several viable and valid perspectives on any given topic or business
domain. To make the knowledge useful and an effective enabler of organizational success, the
KM manager must create a single shared understanding among people of what the knowledge means to
the organization within the context of its business domain and how it is intended to be used.
An ontology provides this unifying map of concepts and relationships. The ontology can be
represented either graphically or in a structured text format. The former is usually used
when the primary goal is to forge a shared understanding of the domain and provide guidance to
the members of the group. The latter approach is most often used for computer applications
that perform language analysis and concept matching, such as the goal of greater automated
semantic capabilities on the Internet (i.e. the Semantic Web).
Taxonomies are the classification scheme used to categorize a set of
information items. They represent an agreed vocabulary of topics arranged around a particular
theme. Although they can have either a hierarchical or non-hierarchical structure, we typically
encounter hierarchical taxonomies such as in libraries, biology, or military organizations. This
type has a tree-like structure with nodes branching into sub-nodes where each node represents a
topic with a few descriptive words. For example, the familiar Dewey Decimal System was
introduced in 1876 as a general catalog of knowledge and is the most common system used in libraries.
The need to classify information is not new. One of the first large
organized cataloguing and classification projects was in the center of ancient knowledge at the
library in Alexandria, Egypt. Its first bibliographer Callimachus compiled the Pinakes, a 120
volume subject catalog of all the library’s books. He is considered the founding father of
librarians since he did not just list the books, but included the author, data on the text, and
comments on authenticity to guide users . However, many others throughout history solved the classification
problem by strictly limiting the number of books by religious, political, or economic reasons, and
then organizing the set by acquisition date, size, or other simple criteria. Thus, classifying
information becomes more important as the number of items increases and people have more trouble
remembering what they have and where to find it. This is now crucial as we buckle under the
immense volume of information available to everyone by the electronic networking of the world.
We have become the fabled man dying of thirst while at sea as we search for the one or two items
that answer our needs from within this sea of information. Indeed, KM is specifically focused on
not only giving people the right information, but going to the trouble of distilling it into
validated contextually connected knowledge that fuses information and data from a variety
of distinct topical areas. In order to classify information a framework must be defined. There
are many existing standards from the Federal Government, consortia, and professional societies. For
example, the Defense Technical Information Center (DTIC) has a technology taxonomy while the
Standard Subject Identification Code (SSIC) is the standard for all DOD information including
memorandums and records management. Similarly, the Library of Congress Classification (LOCC)
is a commonly used general purpose system. However, taxonomies inevitably have a central theme that
guides how the tree structure is arranged. For example, the LOCC and Dewey Decimal System are
built from a perspective of classifying knowledge itself in a general purpose manner. Thus, the
major LOCC headings include topics such as: Philosophy, Psychology, Religion; Auxiliary Sciences of
History; History (General); and Fine Arts. In contrast, DTIC’s major headings are more focused
on technical issues and include: Aviation; Agriculture; Chemistry; and Electrotechnology
and Fluidics. Clearly, trying to find a technology issue will be easier with DTIC than LOCC.