I have a question regarding ontology definition within topic maps. Say I have
a topic map that addresses two or more ontologies that intersect. By using the
concept of topic map templates, I can create topic classes to define an
ontology, and then add topic instances. (i.e. essentially a topic map template
per ontology? – does this make sense?).
However, I’m not sure how topic types are represented in the RM or the HG4TM.
In the proposed hypergraph model, 12/3/2002 (Bernard Vatant, and originally
from Pascal Auillans, et. al.) a triple is introduced consisting of:
S = set of subjects
H = 5 tuple hypergraph
r = representation function
The hypergraph sets of vertices, edges, and incidences relate nicely to the S
set subjects, assertions and role casting:
Association = r(edge)
Role casting = r(incidence)
Subject = r(vertex)
Within S, class typing is also provided for:
T(a, t), or assertion, a, is of assertion type t
T(c, r), or role casting, c, is of role type r
I did not see subject typing however, which I would’ve guessed to be:
T(s, sc), or subject, s, is of subject class sc
I’m probably missing something very fundamental here. Does the absence of this
last formula mean that a subject must be reified before it can be an instance
of a class? ..and if so, how are topic class represented?
The RM is consistent with the HG4TM, so the same question arises – where is the
topic class-instance represented?
Thanks, Rik
|
