W3C home > Mailing lists > Public > public-ontolex@w3.org > February 2013

Re: next telco and todos

From: Aldo Gangemi <aldo.gangemi@cnr.it>
Date: Sun, 3 Feb 2013 20:00:31 +0100
Cc: Aldo Gangemi <aldo.gangemi@cnr.it>, "'Philipp Cimiano'" <cimiano@cit-ec.uni-bielefeld.de>, "'Aldo Gangemi'" <aldo.gangemi@cnr.it>, <public-ontolex@w3.org>
Message-Id: <27AC63B1-58A9-4636-9411-F9B01C55A446@cnr.it>
To: <stellato@info.uniroma2.it>
Dear Armando, Philipp,

the examples from my notes in the wiki were not supposed to be definitive, but just a placeholder: they probably contain typos, and there can be better ways to implement those mappings, as Armando was correctly spotting.

Concerning the possibility to create a lexical vocabulary for them all resources, I'd be really glad to contribute to such a thing  (I have made some work  in this direction in the past as you probably know ;)), anyway let's be assured that it is quite a research-oriented task. 
In principle, research is not considered very well as an activity to be perfomed as a W3C group outcome. 

On the other hand, there is at least one major counterexample, i.e. SKOS, which can be considered a research achievement, and it's 99% outcome of W3C groups. 
Also RDF, OWL and RIF contain pieces of novel research work after all.

So, if we want to go for a sort of SLEX (Simple LEXical resources), I suggest to start from existing lexical ontologies and standards (those mentioned in the requirement page should be enough to start with), and select all the top and certain mid-level concepts and relations that we want to include in such a vocabulary. After that, the task will be to check whether our selection can actually constitute a viable and shared vocabulary, and test it for the (schema-level) alignment of resources.

It's be then interesting to test if SKOS mapping relations are enough to make sense of (instance-level) lexical mappings, or if there are cases of lexical mappings that need an extension of it.

My 2c
Aldo

On Feb 3, 2013, at 5:56:02 PM , "Armando Stellato" <stellato@info.uniroma2.it> wrote:

> 
>> Dear Armando, Aldo,
>> 
>>    of course, every contribution in any section is welcome.
> 
> Thx!
> 
>> On what you say: I understand that many resources have been migrated into
>> RDF, but one issue I see is that they all use a different vocabulary.
>> Would it not be could to have one vocabulary that is general enough to
>> represent all these lexical resources? One to bind them all so to speak
> ;-)
> 
> Not sure, but probably here you are addressing Aldo's response. Btw if I got
> his answer well, both Aldo and me agree on a binding vocabulary, which may
> be used to tie definitions from any resource under our vocabulary. Obviously
> a common modeling framework to be used directly in modeling existing
> resources is not bad, though in my opinion, I see very easily
> rdfs:subClassOf relationships between specific theories and our binding
> vocabulary. And then, we can use ontolex properties to bind these elements
> to the ontology, thus having a common hat.
> 
> For instance, the first triple in:
> http://www.w3.org/community/ontolex/wiki/Specification_of_Requirements/Linke
> d_Data
> wordnetschema:WordSense owl:equivalentClass ontolex:Sense ;
> 
> then, we could say
> wn20instances:wordsense-bank-noun-1 ontolex:hasReference  myontology:Bank
> 
> thus by using our ontolex vocabulary to bind the wordsense in wordnet to an
> entry in the ontology.
> 
> Two possible typos found when roaming around our wiki:
> 
> In
> http://www.w3.org/community/ontolex/wiki/Specification_of_Requirements/Linke
> d_Data
> If I'm correct, it is not:	wordnetSchema:Sense, as written there, but:
> wordnetschema:WordSense
> 
> Also, when looking for which name to use for the property
> ontolex:hasReference, I found this sentence:
> 
> "The relation between the lexical entry and the ontology element should be
> represented by a path involving two object properties relating the lexical
> entry to one of its (many) sense and one relating the sense to the
> corresponding LEXICAL entry"
> 
> At:
> http://www.w3.org/community/ontolex/wiki/Specification_of_Requirements/Lexic
> on-Ontology-Mapping
> 
> Suppose the correct one is:
> "The relation between the lexical entry and the ontology element should be
> represented by a path involving two object properties relating the lexical
> entry to one of its (many) sense and one relating the sense to the
> corresponding ONTOLOGY entry"
> 
> ---
> Finally, one remark on the rels between our binding vocabulary and the
> existing vocabularies for lexical resources.
> 
> Maybe the owl:equivalentClass stated in the specific case of the example
> binding wordnet word senses to ontolex senses, is appropriate, but in
> general, a rdfs:subClassOf relation would allow us to obtain the
> interoperability we desire, without committing too much to the specific
> theories and subtle differences that each specific lexical resource may
> expose. This would also be a +1 on seeing the vocabulary more as a binding
> knot between resources' vocabularies (even pre-existing ones, such as
> Wordnet) than as a basic modeling vocabulary to be used for writing them
> from scratch.
> 
> Best regards,
> 
> Armando
> 
> 
> 
Received on Sunday, 3 February 2013 19:01:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 3 February 2013 19:01:07 GMT