[19]
Ontological
engineering is now a mainstream activity practiced by some of the large e-commerce enterprises and
dot-coms on the Web. This subject is important enough to warrant a chapter by Leo Obrst and Howard
Liu, Knowledge Representation, Ontological Engineering, and Topic Maps (
Chapter 7
). The chapter
presents a historical, theoretical, and practical sketch of the subject. An entire book-length treatment
will eventually be needed, but a notion underlying this book's presentation is that ontological
engineering is what you are doing when you construct XTM documents, and it is important to
introduce that topic early. Bernard Vatant suggests in
Chapter 5
that the use of PSIs is germane to the
process of sharing knowledge, and constructing representations of knowledge is, at once, an art and a
science, as explained in the Obrst and Liu chapter. Later in this book, we return to knowledge
representation using semantic networks (in
Chapter 13
by Eric Freese) and using topic map schemas
(in
Chapter 14
by Holger Rath).
[19]
My first exposure to the term ontological engineering was in a book by Douglas Lenat and R.V.
Guha, Building Large Knowledge-Based Systems: Representation and Inference in the Cyc Project,
Addison-Wesley, Reading, MA, 1990. It is entirely possible that appropriate attribution should lie in
sources much earlier than that.
But wait! There's more--ontological engineering just whets your appetite. So, we follow the gentle
introduction to ontological engineering with a chapter that develops intermediate-level topic maps. For
this, we turn to another notion that underlies this book: topic maps belong in the classroom. In fact,
three different chapters speak to classroom issues--
Chapter 8
, Topic Maps in the Life Sciences;
Chapter 16
, Prediction: A Profound Paradigm Shift; and
Chapter 17
, Topic Maps, Semantic Web, and
Education--where topic maps can add great value. John Lassen Park and Nefer Lin Park, with a bit of
help from me, created
Chapter 8
,Topic Maps in the Life Sciences, which discusses the construction of
several topic maps. Mind you, these are not simple topic maps. Rather, they form the beginnings of an
extended kind of topic map, one that we call a drill-down topic map (that is, one that has the ability to
reference an entire topic map from a topic in a different topic map). Building a drill-down topic map is
a rather new enterprise, one not that well understood.
Chapter 8
presents just one approach to an
implementation of the drill-down feature.
In
Chapter 8
, one topic map serves as a very high level index into several other topic maps, each of
which presents information in a more detailed fashion and serves as an index into even deeper
presentations in the form of more topic maps. This application of topic maps satisfies part of what
Kathleen Fisher (the author of
Chapter 16
) and I characterize as constructivist learning, a learning
process in which children construct their own knowledge primarily by way of personal discovery
during projects, some of which include the construction of concept maps and topic maps.
Chapter 8
begins the process of applying some of the ideas expressed in
Chapter 7
. In the final section
of the book (see below), we pursue these knowledge representation ideas further.
You might be wondering, "How complex can a topic map be?" My immediate answer to that question
is that we just don't know yet. We have intuitions, some backed up by some early observations, but,
judging from efforts to surf Web sites that accumulate taxonomic information on living things, we
already know that some sites, when fully downloaded, accumulate many tens of megabytes of
information. Well, that's a huge download for kids in school, but for governmental agencies involved
in large data management problems, that's small. As a small illustration of the complexity issue, the
opening pages of
Chapter 10
, Open Source Topic Map Software, present two screen images of the