- From: Seth Russell <seth@robustai.net>
- Date: Fri, 4 May 2001 08:58:08 -0700
- To: "Jan Grant" <Jan.Grant@bristol.ac.uk>
- Cc: "Danny Ayers" <danny@panlanka.net>, "Www-Rdf-Logic" <www-rdf-logic@w3.org>, "RDF-Interest" <www-rdf-interest@w3.org>
From: "Jan Grant" <Jan.Grant@bristol.ac.uk>
> Yep; I think this is what both Danny and
> I intimated was a "naieve" approach.
Ok, fair enough. I ran into most of the problems you mention when I tried
to model SEM's implementation in RDBMS using SEM mentographs [1]. I had to
invent arcs for "fields", "related", "pointsTo", "collects" etc. I noticed
that "relates" should be more detailed, expressing one-to-one, one-to-many
kinds of concepts, but gave up because what I wanted to express in my
diagram didn't *need* that level of detail. So I would like to describe my
approach, not as naive, but as less detailed. For example I can say
{FieldA related FieldB} and then later elaborate that with something like
{FieldA (related one to many) FieldB} where {(related one to many)
subPropertyOf related} and allow both assertions to exist in the same model.
But If the graph cannot ever be elaborated with such subProperty arcs, then
the graph is just plain wrong ... and that would be something that I woud
need to know now. So is my SEM graph right or wrong?
[1] http://robustai.net/mentography/SemStructure.gif
Seth
Received on Friday, 4 May 2001 12:04:25 UTC