Lists Home |
Date Index |
Elliotte Rusty Harold wrote:
> You are missing something fundamental. We exchange an agreed upon
> syntax. You send it with your semantics. I receive it with mine. Your
> semantics are not my semantics. There may be some overlap, but it
> doesn't have to be a lot. I may not be interested in the same thing
> you're interested in. For example, you say potato and I say potato,
> but to a store "potato" may mean inventory unit. To a shopper it may
> mean "tasty lunch". The store and the shopper do not have to share
> their internal models of a "potato" in order to do business.
The idea of "shared semantics" is not to imply that the shopper and the
store need have the same exact semantics, rather that there is an
overlap in each's understanding of what a potato is i.e. if I order a
potato on the Web, I expect to get _a potato_. That is the idea. The
store need not think it is a tasty snack, nonetheless, we need to share
enough semantics so that I will be satisfied in what I get.
> The effort to standardize semantics and data models is an effort to
> require all parties to a transaction to have a single way of looking
> at the same content.
No! At least not in the case of RDF/OWL whose semantic models are
defined around the concepts of sets where shared semantics can be
defined as the intersection of any two or more party's semantics. That
is the whole point.