Lime: short check list of pending decisions / small fixes

Dear all,

 

Don’t be scared by the length, these are mostly things which can be decided quite easily, and that do not change the overall specification. We just preferred to make this summary to go through it very easily during the call.

 

The only thing requiring a little effort are those properties in a limbo between: split them/widen their domain. I’m sorry but, since the decision was pending, we couldn’t fix the specification. The position of both Manuel and me is to widen the limbo, and keep a meaningful name for them to be used where needed. However, we have no issue in splitting them, just, in this case, we need the further property to be generated, as not adding it would create a gap in the specification.

 

*         Add namespace of (Lemon/all or Ontolex) in the definition of lexicalizationModel to represent when ontolex is used as a lexicalizaiton (which of the two? Lem/all or ontolex?)

*         Properties to be split/range to be widened:

o   ReferenceDataset: Philipp removed LexicalLinkSet from the domain, but LexicalLinkSet needs something like this, be it the same or another one. Note also that an axiom on LexicalLinkSet still say to require: =1.referenceDataset.

o   References: same as for referenceDataset

o   Concepts:  missing ConceptualizationSet from its domain (to be homogeneous with LexicalizationSet featuring also lexicalEntries)

o   Lexical entries: 

*         Name of avgXXX props (both of them: avgSynonymy, avgPolysemy). See relevant email sent by Armando yesterday night

*         URI representing a lexicalization based on URI of references themselves (in case no other is available, good to consider if the localnames are more or less readable, and to know in which language). A proposal could be:  <http://uri4uri.net/vocab> http://uri4uri.net/vocab or we invent 

*         Statistics Counts: We can decide if the metaelements (e.g. ontology itself) count or not. After all, maybe the ontology itslef is not always expected to be counted. This is "not so" important for the specification, though we might need to fix the examples (in one comment on the lexicalized FOAF example, the coverage is not 100% because of the lack of the lexicalization on the FOAF URI itself).

 

Cheers,

 

Armando