[topicmaps-comment] RE: Ontological extravagance in the Topic MapGraph?

[Ivan Uemlianin <ivan@jurakm.com>]

Why are so many different types of nodes, arcs, non-node and non-arc objects
needed in the topic map graph?

According to tmpm4, the topic map graph has the following entities:  three
types of node (topic, association and scope) and four types of arc
(association-member, association-scope, association-template and
scope-component).  There are also the non-node subject identity points and
the non-arc links between topic nodes and subject identity points, both of
which are outside of the graph, but are nonetheless part of the ontology
defined by tmpm4.

It is possible to represent the same set of graphs with a smaller ontology,
specifically: three types of node (topic, association and literal) and one
type of arc (association-member).  Association-scope and
association-template arcs are here represented by association-member arcs,
labelled (i.e. with 'role' specified) as 'scope' and 'template'
respectively.  Subject identity points are represented as literal nodes
(with resource being a special case of literal), and topic nodes are linked
to these literal nodes via association-member arcs.

Presumably there were reasons to introduce the reified nodes and arcs (i.e.
scope nodes, and association-scope, association-template and scope-component
arcs) and to create ill-defined non-graph appendages (i.e. the subject
identity points and the links to them).  However, not only do they
complicate a simple ontology, they increase the distance between graphical
representation of topic maps and RDF.  Simple graph ontology and simple
RDF-TM links both seem like reasonably good ideas.

So, why such a complicated ontology?

Ivan

Ivan Uemlianin, PhD

Head of Topic Map Development
Jura Technology Limited

6 Tai Seion
Llanddeiniolen
Caernarfon
Gwynedd LL55 3AF
Wales, UK

Head Office:
35 Norroy Road
Putney
London SW15 1PQ
UK