-- Alan > Also, mime types tell you the format of the resource, but > not the type of content. So compare a definition and a policy > statement metadata for a uri. Both might be text/rdf+xml, no?
I am not sure what is your point here? What do you mean "compare a definition and a policy for a URI"? How can you compare such things? > It seems to me that content negotiation is a pre-semantic web > way of handling a bit of semantics. Why not adopt a uniform > way of handling these things? You could certainly support > content negotiation discovered information by translating it > to to the more expressive rdf/owl. Content negotiation is not "pre-semantic web way". You have to remember semantic web is an extension of the current web, not a new web. Per Godel's incomplete theorem, there is no such thing of a completely self-descriptive system, you have to build your system on certain foundation that cannot be answered by the system. I am just trying to help you not to waste your time. If you can come up with an approach that is generalized to all cases, I am the first one to adopt it. But if it is only a partial solution to a few use cases, it becomes a "hack". If it is a hack, there is a lot of cheaper ways to do so. For example, by social conventions and guess the nature of URI. Why goes to the extra miles of developing ontologies? Xiaoshu