Re: SV: RDF tutorial

Sari Hakkarainen wrote:
> I red the RDF Tutorial and liked it very much.

thanks :)

> In addition, it (unintentionally?) highlights the two, in my opinion, main reasons for "not getting it right", i.e. causes for <certain difficulty>, namely

this was all the more unintentional that I don't really see them as problems, actually :

> a) the treatment of the subject (S), predicate (P), object (O) as a natural language metaphor, and

sure, the NL transcription of RDF properties does not follows one single rule,
rdf:type(s,o)        must be read  "s HAS type o"
rdfs:subClassOf(s,o) must be read  "s IS subClassOf o"
tut:contains(s,o)    must be read  "s contains o"

but, well, NL transcription is syntactical "sugaring",
it makes RDF human-readable with as few as possible ambiguities,
does it have to be canonical ?

> b) the blurred line between types and instances.

> The type-instance-problem (b) is demonstrated by the reluctance to use the "normal" classification of RDF concepts into S&M and S and to give examples of Classes and types.

I dont think so. In my opinion, and as I mentioned in subsection "Schema definition concepts",
the notions of Class and type need not be known by any RDF agent.

> Following the spec:s, one inevitably defines instances and types in the same schema.

Why so ?? It is possible to define classes in one RDF file and instances in another one.

> It is a classical problem in application of typed knowledge representation languages to keep the type and instances separate. It is also where some ways of overcoming the problem may be found.

that is not true of F-Logic, which was used to process RDF statements in SiLRI

--- Quid quid Latine dictum sit, altum viditur
    Whatever is said in Latin sounds important.

Received on Monday, 27 March 2000 04:24:54 UTC