Some remarks on LIME

Dear Philipp, John, Armando, All

I looked both at the Final Model Specification (
http://www.w3.org/community/ontolex/wiki/Final_Model_Specification) and at
the ontologies published to GitHub (
https://github.com/cimiano/ontolex/blob/master/Ontologies/lime.owl and
https://github.com/cimiano/ontolex/blob/master/Ontologies/ontolex.owl), in
order to evaluate the status of the metadata module.

You can find my remarks below.

Remark #1
The figure on the wiki (
http://www.w3.org/community/ontolex/wiki/Final_Model_Specification#Metadata_.28lime.29)
is outdated. However, I do believe this is due to the fact that the
vocabulary is still being discussed.

Remark #2
The class lime:LexicalizationSet appears to require exactly 1
ontolex:Lexicon. However, I remember that we allowed the description of
"legacy" lexicalization sets (e.g. SKOS, or RDFS) that do not refer to a
separate lexicon. Therefore, the association between a
lime:LexicalizationSet and a ontolex:Lexicon should be optional rather than
mandatory. Additionally, lime:lexicalizationModel (which is supposed to
distinguish between different lexicalization approaches) has no domain, nor
it is used in the definition of lime:lexicalizationSet.

Remark #3
In the formula defining lime:percentage (
http://www.w3.org/community/ontolex/wiki/images/b/bb/Percentage_formula.gif),
it appears something like "entity ∈ reference", which is not entirely
obvious to me.

Remark #4
lime:conceptualDataset or lime:conceptDataset? I remember we have discussed
it, but I am not sure if we agreed on a choice. The intended meaning should
be "a dataset containing lexical concepts".

Remark #5
The description of ontolex:ConceptSet in the Wiki is not consistent with
the OWL definition. The description in the OWL ontology (
https://github.com/cimiano/ontolex/blob/master/Ontologies/ontolex.owl#L391)
seems to contain some mistakes:

   - there is no rdfs:subclassOf void:Dataset axiom
   - the axiom concerning skos:inScheme is wrong, because the property
   links concepts to concept schemes, rather than the opposite. It seems to me
   that there is no property relating a concept scheme with its concepts (not
   only the top concepts). In OWL2, you could use an "inverse of" property
   expression. Also, I am entirely sure about the use of owl:equivalentClass
   .

Remark #6

ontolex:Lexicon is not declared to be a void:Dataset. Remember that we
agreed to combine the data-level class ontolex:Lexicon with the metadata
level class lime:Lexicon.

Remark #7

There is no class ontolex:Conceptualization, which associates an
ontolex:Lexicon with a ontolex:LexicalConceptSet. With respect to this
class, I wonder whether we can find a less ambiguous name. Indeed, it
recalls to may mind the famous definition "an ontology is a formal,
explicit specification of a shared conceptualization" (Studer et al.,
1998), in which the word conceptualization is used in a rather different
sense.


-- 
Manuel Fiorelli