Page 294 -

Knowledge Management Tools 277

Taxonomies — hierarchical information trees for classifying information — act like
your library subject catalog. They can help overcome differences of language usage in
different parts of an organization and even the use of different languages. Traditionally
manually intensive, the growing problem of information overload means that they
are receiving signifi cant attention. But how do you cope with the evolution of terms,
whose meaning seems to change from one year to the next? Automatic (or
semi-automatic) classifi cation of information objects — natural language analyzers, text
summarizers, and other technology — helps to understand some of the meaning —
the concepts — behind blocks of text and to tag and index it appropriately for to aid
subsequent retrieval. Many take advantage of the organization’s underlying knowledge
taxonomy.

Folksonomies and Social Tagging/Bookmarking
Metadata is literally translated as data about data and refers to specifi c information
about content contained in books, reports, articles, images, and other containers so
that they can be organized and retrieved in an orderly fashion. Metadata is also
referred to as tags or keywords. Taylor (2004) notes that metadata comes in three
general fl avors: administrative, structural and descriptive. The Oxford Digital Library
(ODL) (http://www.odl.ox.ac.uk/metadata.htm) defi nes three types. Administrative
metadata is the information needed to manage the information resource over its life
cycle such as data about how it was acquired, where it came from, licensing, intel-
lectual property rights, and attribution (e.g., was it scanned, what format is it stored
in, etc.). This is sometimes referred to as preservation metadata. Structural metadata
relates to the actual computer elements involved such as tables, columns, and indi-
ces — all the logical units of the information resource. Descriptive metadata refers more
to the content or subject matter of the information resource to help users fi nd it (e.g.,
cataloguing records, fi ndings aids, keywords). Descriptive metadata is of greatest
concern in KM because we often need to expand this type of data about data greatly
in order to increase the usability (and reusability) of a given unit of knowledge.
Metadata is very formal and tends to be created and updated by dedicated person-
nel such as catalogers and other library and information science professionals. This is
the highest standard in metadata but is time consuming to produce (Mathes 2004).
An alternative is to have authors create and add their own metadata for their own
works. The Dublin Core best exemplifi es author-created metadata (Greenberg et al.
2001). Both of these approaches work well for the person who develops the metadata
but not necessarily as well for other users (often referred to as unknown or unantici-
pated users). A third option exists — that of user-created metadata. This bottom-up or