DISQUS

Bozo Faust · 4 years ago

The formal logic which has been specified in the RDF semantics only works where identity is not sensitive to context. But identity is always sensitive to context. The W3C cannot legislate that away by writing specifications. I'm thinking that Semantic Web enthuiasts will slowly recognize that a strong logic, requiring the excluded middle, will never work on a global world scale. Oh well ...

John Black · 4 years ago

I think this is true. More and more people are saying this: Identity (and meaning as well) are in Agents[1] and are context dependent. I now think all ontologies are personal and are, in fact, problem specific. I have been trying to think through a system without "class" or "type" - only properties. Any notion of "concept" or "class" or "type" could only be calculated on ad hoc on the basis of a sample of items with similar properties. Such ephemeral "concepts" could be shattered at any moment by the addition of incompatible properties - only to be replaced momentarily with a new set of "classes" that fit the new dataset.

1. http://kashori.com/2004/12/it-takes-agent-to-be...

Marty Heyman · 4 years ago

Thanks. Nicely put and plenty to think about. I wonder if the real responsibility of the Aggregator isn't to collect the identity information in such a way that the collection of data has a minimum of ambiguities by design? From any richly understood identity, one can probably derive RDF representations of several "masks" worn by the same person or thing .

My recent projects in address book directory design for mixed business and personal use have me circling this subject and I appreciated the viewpoint you offer here.