I am reasonably sure that one can implement topic map
functionality in a set of relational tables and SQL queries, and then,
as with RDF, export a compliant SGML/XML instance of that
information. Do you know of any reason why this would not
work?
I ask because as Rex notes, some groups are attempting
to build infrastructure for VRML communities that will be
realized as a database. While topic maps are attractive,
I know of only two working implementations (it's been awhile so
if there are more, great) and neither of these are in widespread
use yet. So to meet shortterm goals, a relational DB may
be the best approach. Still, it may be useful to ensure
that topic maps can be produced to provide the information
to those systems that can use topic maps.
Therefore the question: if one is to implement relationally,
what should one consider with an eye towards eventual use
of topic maps?
len
>Ralph Ferris wrote:
>
> Topic Navigation Maps (TNMs) provide a means of structured navigation
> through a body of related information. TNMs represent a generalization of
> concepts such as indices, glossaries, thesauri, and tables of contents.
> They follow a pattern of:
>
> Subject
> Subject definition
> Links between related subjects
>
> Link elements are used to create associations between related subjects;
> metadata can be used to describe the relationships.
>
> TNMs also provide a means of expressing inter-relationships between bodies
> of information. They are a key potential application area for
> XML/XLink/XPointer. Or, said another way, TNMs represent a means for
> applying XML/XLink/XPointer to KM.
>