Re: RDF Terminologicus

Pierre-Antoine,

Thanks for picking up the baton.  I make no claim that my definitions were 
best, or even correct;  I was trying to respond to what I thought were good 
suggestions about locking down some of the terms we use.

I had been thinking some more about this issue, and wonder if it might not 
be a useful goal of the RDF-IG to prepare a NOTE covering "RDF terminology 
and concepts" -- that way any gains from these discussions can be "locked 
in".  Maybe an addition to the Semantic Web Activity proposal currently up 
for consideration?

Responding to your specific points...

At 11:47 AM 1/4/01 +0100, Pierre-Antoine CHAMPIN wrote:

> > Stand for:
> >    A labelled entity that is used in descriptions indicate some entity or
> >    concept.
>
>I don't get this one.

I was trying to cover my use of that phrase in the next definition...  I 
was probably being too picky.

> > Reification (of a statement):
> >    [See RDFM&S section 5]  A resource that stands for the statement
> >    together with the four statements that describe the statement.
> >    In my opinion, a reification of a statement is not unique:
> >    there may be more than one reification of any given statement.
>
>It is true that, according to M&S, the reification is the resource *and*
>the four triples. I would hence add
>
>Reified statement:
>     The resource which stands for a statement in a reification of
>     that statement.
>
>which may be confusing, but looks a useful distinction. To sum it up :
>reification : resource + 4 statements (s,p,o,type)
>reified st. : resource
>
>I agree with you about the fact that a statement (which is unique, cf M&S s5)
>may have more than one reification.

I think a problem with the term "reified statement" is that it suggests the 
reification is unique (e.g. the "statement" part _is_ unique, per M&S, 
hence the implication that "reified statement" is unique.).  I would 
suggest something like "reification resource" or "statement resource" for 
the purpose you describe.

> > Stating:
> >    An assertion that some statement is true in some context.
> >    (or should that be:
> >    An assertion in some context that some statement is true.
> >    ?)
> >    NOTE:  this assertion is a statement separate from the
> >    statement asserted to be true.
>
>Funny definition.
>At first sight, I would have said I disagreed with it,
>but the time I wrote it, I was not so sure...
>
>If I write
>
>   s1: [Bush won Elections]
>   s2: [s1   ist Context1]
>
>from your definition, s2 is a stating of s1.
>But we have no statings of s2, which has been clearly stated, though...
>To get a stating of s2, I have to write
>
>   s3: [s2   ist Context2]
>
>and so on...
>I like that idea of statings.

Hmmm... tricky one, this.

If I say:

   [Bush won Elections]

I think that is a "stating" of 's1' in the (implied) context of our 
conversation.  I think your

   [s1 ist Context1]

is another stating that describes a stating of "s1".

I think that "stating" is the basic mode of RDF:  every RDF statement is a 
stating.  A reification is also a stating, that some identified resource is 
of type rdf:Statement and has certain properties;  i.e. it merely asserts 
the existence of and an identity for a statement without asserting its truth.

Thus, I would re-word my definition to be:

Stating:
    The expression of an RDF statement [or set of statements]
    in some context of discourse that is taken to be an assertion
    of the truth of the statement[s] in that context.

NOTE: I don't think the notion of stating should, of itself,
invoke the formal language of contexts.

#g

Received on Thursday, 4 January 2001 07:58:19 UTC