Re: lime module

Hi Philipp, All

please read my further answers.


2015-07-15 8:27 GMT+02:00 Philipp Cimiano <cimiano@cit-ec.uni-bielefeld.de>:

>  Hi Manuel,
>
> thanks, see below ...
>
> Am 13.07.15 um 18:22 schrieb Manuel Fiorelli:
>
>  Dear Philipp, All
>
>  Following our discussion on the LIME module during the last telco, here
> are some updates on the specification:
>
>
> https://www.w3.org/community/ontolex/wiki/index.php?title=Final_Model_Specification&diff=2289&oldid=2250
>
> The spec has been modified to address some of the issues I have raised in
> previous emails (see details below within the quoted text).
>
>  The diagram on Draw.io has been modified, considering the current state
> of the Lime metadata vocabulary. Further modifications could be required
> once you decided what to do with the properties to renamed or split.
>
>  Some examples were added to the end of the metadata module, but we will
> revise them in the next days. We modified some definitions, but others have
> not been modified because of the possibility they could be split or
> renamed. Specifically, here are some definitions (or axioms) to be modified:
>
> *lime:lexicalEntries*
>
> - The domain of this property should be Lexicon or LexicalizationSet or
> Conceptualization and the definition should be changed accordingly, unless
> we want to split this property into two or more properties.
>
>
> I changed the property definition to also include ConceptualizationSet as
> domain.
>
> You mean ConceptualizationSet, right?
>
>
Yes, I meant ConceptualizatioSet.


>   *lime:referenceDataset*
>
> - the definition should be reviewed
>
>
> For me the definition is fine, what exactly should be reviewed?
>

I think Armando has already replied on this point. The fact is that we
should decide if to use this property to count ontology elements linked to
lexical concepts (via a LexicalLinkset) or not. In the affirmative case, we
should make it clear in the definition, otherwise we should coin a new
property (I think).

>
>  *lime:lexicalizationModel*
>
> - the domain should not include ontolex:Lexicon (this could be a refuse
> remained after the introduction of lime:linguisticModel)
>
>
>   OK, fixed...
>

In the wiki, I still see the class ontolex:Lexicon in the domain (as an
argument of the OR).


>
>   *lime:references*
>
> - Not sure if this will be split or renamed
>
>
>   See my other email on this, I propose that for the sake of clarity and
> avoid overloading we keep this property as denoting the number of distinct
> ?o in triples (?s,reference,?o)
>
>   *lime:percentages*
>
> - in the definition, we should add the mention to lexical linksets
>
>
> I changed this as follows:
>
> The '''percentage''' property expresses the percentage of entities in the
> reference dataset which have at least one lexicalization in a
> lexicalization set or are linked to a lexical concept in a lexical linkset.
>
> Fine?
>

It is fine to me.


>
>
>  *lime:partition*
>
> - the definition of partition is wrong, as it only refers to
> lexicalization sets
>
>
>  *lime:resourceType*
>
> - as before, it only mentions lexicalization sets
>
>
>
> OK, thanks. I changed the definitions. Are they fine now?
>
>
Concerning the definition of lime:resourceType, I am not sure about the
part "or a partition thereof", because a partition of a lexicalization set
or lexical linkset is, respectively, a lexicalization set or a lexical link
set.


>   *lime:concepts*
>
> the introduction to the definition of lime:concepts firstly mention its
> use in a concept set, although we are in the section about lexicalLinkset
>
>
>   OK, I introduced a pointer to the definition of ConceptSet in ontolex.
>
> Fine?
>

I think it is fine.


>
>   *lime:avgNumOfLinks*
>
> - the definition is wrong. This property should give the average number of
> links per ontology entity
>
> I changed the definition to:
>
> The '''average number of links''' property indicates the average number of
> links to a concept for each ontology element in the reference dataset.
>
>
I think that to we can apply the same observations that Armando make on
avgNumOfLexicalizations in another email.

Regards

Manuel Fiorelli