Comments on RDF Schema proposed recommandation

Dear RDF enthusiasts, 

RDF Scheme proposed recommandation looks very fine. The aims and 
concepts are now clear.
I have few questions question and remarks. I hope I didn't
misunderstood some points.

1- I am very confused with some used terms, especialy with "subClassOf".
Refering to RDF M&S, an arc labelled C from A to B means "B is C of A".
Applying this rule to the figure 2, this gives for example 
"rdf:Property is subClassOf of rdf:subPropertyOf".
The same result for Class with Resource, ConstraintResource with Class,
 etc.
Should not "subClassOf" be replaced with "superClass", in the same 
manner than orginal "instanceOf" has been replaced with "type" ?
In this case, should not "subPropertyOf" be replaced with 
"superProperty", and "isDefinedBy" with "definition" ?

2- In 2.3.1 and below, the notation "A is X of rdf:type rdfs:Class"
is a bit confusing too.
Why should not use "A is X with rdfs:Class as property rdf:type value" ?

3- In 2.3.2.1,  should not the figure be labeled "Figure 3" ?

4- In Figure in 2.3.2.1 should not an arc from xyz:MiniVan to
rdfs:Class,
labelled with "rdf:type" be  added ?

5- In 2.3.2.1, what is the meaning of "In general, the XML ..." ?
Should we understand that this is always true, or what are the cases for
which this assertion is false ?


6- In example  2.3.2.1, it is not clear why class MotorVehicle is
declared as subclass of rdf:Resource.  If this is not implicit, why
in example 1 in section 7, the class "MaritalStatus" is 
not also declared as subclass of rdf:Resource ?
The same question  applies for example in annex C (classes
p:Multivalue, 
ColorValue, and for XML serialization (Literal from others, while
Property is declared as subclass of Resource).
If this declaration is always needed, what should be the behavior of 
an application finding a class declaration ruling out this ?
In definition of rdfs:Resource class (in 2.2.1) it is stated that
"All things being described by RDF expressions are called resources..."
Since the properties itself can be - in my understanding - be described
by RDF expressions, why properties are not subclasses of rdfs:Resources
too ?

7- In 2.3.5 if I understood the need of such property, the example is
not
very clear for me. Why is it stated that "[the construct does] not 
indicate the URI of the schema..." ? If it is the form of the URI,
it must be clearly explicited. Anyway, I understood that giving a URI
for namespace does not imply that the schema can be found at this URI.
If this last is true, the need of definition property is clearly 
established. But do we need to consider that if there is not a such 
property, a RDF parser can resonably consider that the schema could 
be found a the namespace URI ?

8- In figure in section 3, should not a vertex "rdf:object" be added ?
Are  the domain value and the absence of range for rdf:value correct ?

9- In section 6.1  in definition of rdfs:Literal, it is stated this 
is a set.
In XML serialization it is a object of type "rdfs:Class", but in
figure 2
in section 2.1.2, its type is "rdfs:Resource". 
What is the correct definition ?

10- In section 6.5, in definition of rdf:object, there is only a domain 
defined. I understood that its range is either a resources or a
literal, 
which can be expressed with a range of resources. Am I wrong ?

11- Why rdf:value has no range and no domain ?

I hope that I am not too far from a correct understanding of this
proposed
recommandation and I hope this could help.

Regards


==
Olivier MARCE

_________________________________________________________
ĘTES-VOUS YAHOO!?
Votre e-mail @yahoo.fr gratuit sur http://courrier.yahoo.fr

Received on Monday, 22 March 1999 15:31:23 UTC