Re: Semantic Web standards

The first time I looked at the Semantic Web standards documents, which
was a while ago, I didn't know of any APIs. (Probably there were some
that I just hadn't heard of!) With an intense mark-up language like
RDF-XML, the API is going to be pretty important.

As for the real benefits or costs of this approach, that's
complicated. I suppose the main benefit might be that using existing
tools will facilitate more rapid development. Since "inference" is
supported by OWL, I'd be curious to know what sorts of inferences you
might like to see computed.

Certainly the OWL data structures can be transformed into triple
structures like I talked about and vice versa. The actual mechanics
of further annotating an edge might be complicated in the OWL model
however; and if you build large complexes with successive annotations
attached to edges, it seems like addressing would get gnarly. Having
first-class edges makes annotation basically "first-order".

On the other hand, I don't have any concrete evidence to indicate how
useful this feature is; just the nice clean statement that "everything
in Arxana is annotatable".

(Another thought: One thing that seems nice to have is "baseless
theories", i.e., sets of relationships among undisclosed, unspecified
objects. While I think any system will need to use place-holder
objects, it would be convenient to be able to swap these place-holder
objects out for other more interesting objects. It seems to me that
an object-centric theory would make this harder to do.)

I don't think I had heard of Chris Hanson's "Web-scale Environments"
project, but it sounds phenomenally interesting. Maybe it will
eventually subsume everything else we are doing here. (At least: if
you look in the last paragraph where he talks about enhancing
"Lisp-based programming systems with persistent storage that is better
matched to Lisp than relational databases or file systems", that is
what Elephant already does, although it does not currently do it in a
distributed way.)