Dear Dan,
On 2010/11/12, at 20:08, Dan Brickley wrote:
Dear all,
The FOAF RDFS/OWL document currently includes the triple
foaf:name rdfs:subPropertyOf rdfs:label .
This is one of several things that OWL DL oriented tools (eg.
http://www.mygrid.org.uk/OWL/Validator) don't seem to like, since it
mixes application schemas with the W3C builtins.
[snip]
3. would you consider checking for ?x rdf:type foaf:LabelProperty or
other idioms instead (or rather, as well).
No.
4. would you object if the triple "foaf:name rdfs:subPropertyOf
rdfs:label " is removed from future version of the main FOAF RDFS/OWL
schema? (it could be linked elsewhere, mind)
Please Dan. I really don't like the idea of removing this triple. I am
not an expert on ontology but we are talking about Linked Data here in
mailing list. Linked Data means linking data across domains and
ontologies are data. We have to encourage people to make links between
different ontologies so that we can do more inference, not only OWL-
based, in the future.
People can hard-code {foaf:name rdfs:subPropertyOf rdfs:label} into
their system but I really don't think this is very elegant and
subpropertying rdfs:label is the only clue that these label properties
are related.
This is also pure fact-finding. If we add {rdfs:label a
foaf:LabelProperty} to RDFS, will that break even more systems?
I see only two options:
1. Do nothing
2. Delete foaf:name and ask people to use rdfs:label instead
--- Theoretical Discussions I am not good at ---
If you deleted {foaf:name rdfs:subPropertyOf rdfs:label}, the means it
was a lie. And it is very sad now that lots of people are relying on
this lie.
What exactly is the reason why DL tools don't like this triple? Is it
logically valid or it is some bug in the DL tools. I am never good at
descriptive logic but I can try to learn it.
Cheers,
Kenny
