Re: s+as review (fwd)

>2.2 Facts
>
>"Normal Form C" - has this restriction been relaxed now? Check JJC.
I take it to be this line:
    lexicalForm ::= as in RDF, a unicode string in normal form C


No, this restriction is still in place - it is the restriction on URIs that 
has been removed.

>2.3.2.1 BNF for axiom
>[[
>         | 'EquivalentClasses(' description { description } ')'
>]]
>[Editorial] Other 'equivalentX' productions specify a minimum of two
>equivalent Xs.

This was deliberate - the motivation is that, when combined with the mapping 
rules in section 4 it permits the following to be in OWL DL.

_:b rdf:type owl:Restriction .
_:b owl:someValuesFrom owl:Thing .
_:b owl:onProperty eg:p .
eg:p rdf:type owl:ObjectProperty .

I argued for this being permitted in OWL DL since I found it unaesthetic to 
make 'orphaned' descriptions and restrictions syntactic errors rather than 
merely semantically not very meaningful.

The phrase in section 4 that combines with this is:
"When the transformation of a component is used as the subject, predicate, 
object of a triple, even an optional triple, the transformation of the 
component is part of the production (but only once per production) and the 
main node of that transformation should be used in the triple. "

The optional triple 
in
EquivalentClasses( description1 ) 
is

description1 owl:equivalentClass description1 .
which even when omitted gives rise to the orphan as above.

>4.1
>[note] While the abstract syntax naturally associates (via syntactic
>nesting) ontologies with all their directives, no such association is
>made in teh RDF graph expression of the ontology (apart from
>Annotations). I can see why this is the case.
>
>I'm not really sold on the translation table; I think the meaning of it
>is unclear. However, I'm stumped as to an alternative compact expression
>of the translation into RDF Graph form so feel free to ignore this
>comment.

Did you look at my formal objection?
http://lists.w3.org/Archives/Public/www-webont-wg/2003Mar/0264.html
in it I point at
http://lists.w3.org/Archives/Public/www-archive/2003Mar/att-0089/m
as an alternative articulation of the constraints on the graph. You still need 
the mapping rules to tie it up with the abstract syntax.
It's also worth noting that the mapping rule

EquivalentClasses(description1 … descriptionn)
==>
 T(descriptioni) owl:equivalentClass T(descriptioni+1) . 1≤i<n 
T(descriptioni) owl:equivalentClass T(descriptionj) . [opt] 1≤i,j≤n 

makes the syntax NP complete (see
http://lists.w3.org/Archives/Public/www-webont-wg/2003Apr/0003.html
)

In the formal objection I argue that there should be fewer restrictions on 
blank nodes used as descriptions. I wonder whether we might also be of the 
opinion that there should be fewer restrictions (e.g. none) on blank nodes 
used as individuals.
e.g.

currently OWL DL (and OWL Lite) does not allow a blank node denoting an 
unnamed indiviudal to be the object of two triples, nor does it allow cycles 
of blank nodes denoting unnamed individuals. Nor does it allow unnamed 
individuals to participate in owl:sameIndividualAs or owl:differentFrom 
owl:oneOf or owl:hasValue.
Since we have decided that the similar restrictions in RDF/XML were a mistake 
and we have removed them, might we be of the opinion that these restrictions 
are similarly a mistake in OWL DL?

Jeremy

Received on Sunday, 27 April 2003 16:53:28 UTC