Re: ontolex.owl

Dear Philipp,

I reviewed the final specification (
http://www.w3.org/community/ontolex/wiki/Final_Model_Specification) and the
OWL ontology (
https://github.com/cimiano/ontolex/blob/master/Ontologies/ontolex.owl), for
what concerns with the core module. The following paragraphs follows the
structure of the final specification; However, I interweave comments on the
OWL ontology as well.

*Comments on the ontology (ontolex.owl):*

The comments on the defined entities are represented as xsd:string typed
literal. In fact, they should be plain literals (or language tagged
literals, in RDF 1.1) with language tag en.


*Section "Conventions in this document"*

Since you advocate the use of RDF 1.1, then you could use the latest
specification of Turtle, which has been published alongside RDF 1.1:
http://www.w3.org/TR/turtle/

The LIME namespace has been erroneously associated with the prefix vartrans:

*Section "Core*"

In the previous section you associate the ontolex: prefix to the namespace <
http://www.w3.org/ns/lemon/ontolex#>.
However, in the vocabulary description, you use URIs such as <
http://www.w3.org/ns/lemon/ontolex/LexicalEntry>, which assumes a different
namespace <http://www.w3.org/ns/lemon/ontolex/>.

*Section "Core" / "Lexical Entries"*

In the specification there is only the axiom stating that a lexical entry
has a at least one lexical form. However, in the ontology there is also an
axiom stating that there could be at most one canonical form.

*Section "Core" / "Forms"*

In the specification of the class Form, you use a qualified number
restriction, while in the ontology you use an ordinary number restriction.
Moreover, the examples following the class definition don't explain to me,
when a form may have two or more written representations.

In the ontology, ontolex:representation does not have a range. Nor
ontolex:phoneticRep has a range (it would be redundant, anyway).
ontolex:writtenRep has range xsd:string, which is inconsistent with the
specification, which uses RDF 1.1 language tagged literals.

>From the specification: "It can also have different phonetic
representations corresponding to different ways of pronouncing it." I would
replace the pronoun with the explicit subject "Form"

Thank you for making me aware of language tags such as en-US-fonipa (for
the phonetic representation). Maybe you could highlight in the
specification how you used the sublanguage tag fonipa for the phonetic
representation of a form in a given language.

*Section "Core" / "Semantics"*

In the ontology, the definition for ontolex:isSenseOf erroneously uses "
owl:incompatibleWith", instead of rdfs:comment

In the ontology, ontolex:isReferenceOf is not explicitly declared inverse
functional.

*Section "Core" / "Lexicons"*

In the first example, the row

:lexicon a rdf:type ontolex:Lexicon;

is wrong, since it contains both "rdf:type" and its abbreviation "a".

In the ontology, ontolex:languageURI applies also to LexicalSense, while
ontolex:language does not.

The range of ontolex:language in the description is wrongly indicated as
langString

I would be more explicit about the possibility to mix language tags: for
example, "en" for the Lexicon and "en-GB"/"en-US" for morphological
variations of a lexical entry. I don't know if it is the case to explicitly
asserting that you should not have a lexicon for english containing a
lexical form in French.

>From the specification "This shows how the English lexicon and the Spanish
lexicon are linked via the concepts that the different words in the lexicon
denote, thus integrating the lexica at the conceptual level." I don't know
if it is appropriate to use the word "concept", since it may be ambiguous
with the following notion of Lexical Concept.

*Section "Core" / "Lexical Concept"*

In the ontology, there is no axiom relating ontolex:LexicalConcept to
skos:Concept.

In the ontology, ontolex:isEvokedBy has no inverse property axiom.

In the ontology, there is no inverse property axiom relating
ontolex:lexicalizedSense and ontolex:isLexicalizedSenseOf



2014-06-23 8:28 GMT+02:00 Philipp Cimiano <cimiano@cit-ec.uni-bielefeld.de>:

> Dear all,
>
>  I have been working on ensuring consistency of the ontolex.owl module
> with the spec.
>
> I have:
>
> 1) made sure that all axioms in the ontology correspond to those in the
> spec; the ontology is consistent ;-)
> 2) ensured that all domain / ranges match
> 3) introduced examples illustrating the use in the git under directory
> "Examples"
> 4) included the semiotics.owl ontology as we decided some time ago; we
> agreed to not have this in the spec, but in the actual ontology, see
> ontology file
>
> I attach the current version of the ontolex.owl module (see attached).
>
> I would kindly ask you to help me to ensure that the example in Examples
> are fine. I did not find a service to validate the files (they are in
> Turtle syntax). Can someone please check them and modify them
> appropriately, modifying also the spec where these example are given?
>
> I would like to finalize the model on our telco on Friday, so please raise
> any concerns this week.
>
> Please carefully review the ontology and the text; if you spot any issues,
> please let me know.
>
> I have added a pointer to the GIT in the main wiki page.
>
> Best regards,
>
> Philipp.
>
>
> --
>
> Prof. Dr. Philipp Cimiano
>
> Phone: +49 521 106 12249
> Fax: +49 521 106 12412
> Mail: cimiano@cit-ec.uni-bielefeld.de
>
> Forschungsbau Intelligente Systeme (FBIIS)
> Raum 2.307
> Universität Bielefeld
> Inspiration 1
> 33619 Bielefeld
>
>


-- 
Manuel Fiorelli