Hi Alex,

>> Ok but in that case, how a concept will be related to other concepts? By
>> this I mean: each tag will be a concept without any relationship with
>> other concepts. So, the graph of tags from this ontology will create a
>> graph with unlinked nodes, thousands of them?
> Actually my goal is not to link tags together, so I will not adress
> that issue in this ontology.
> But in that case, I think we can rely on the SCOT ontology, that have
> properties to link tag objects.
> I have to see how it fits there (maybe the moat:Tag class as a
> subclass of tag:Tag and scot:tag)
But in that case, how to make sure the related tags are rigthly related?

By this I refers to:

I have a tag "tag_a" that has a meaning "meaning_a" and "meaning_b"

Then another person has a "tag_a" too with meaning "meaning_c" and 

Now what if such a system link "tag_a" with "tag_b" and that "tab_b" is 
in conflict with the meaning given by me (so "meaning_b").

I fear that such a syste would link the links only according to the 
literal of the tag. When MOAT is interesting since it relates a literal 
to some of the meaning of that literal, in a context:

- the context where the tag is used
- the user that used that lateral to categorizes that thing. (so the 
cultural baggage and experience of the tagger will affect the meaning he 
gives to the tag)

> I thaught that's optimal for the ontology not to restrict the range to
> be (or "become") a skos:Concept but to be the more open as possible.
> Not especially with dbpedia, but any knowledge base, since I first
> used this for a internal project with its own ontologies.



Received on Monday, 21 January 2008 14:38:19 UTC