RE: Objects as RDF subgraphs

>a) 'folding' is a nice word for it. I might steal that ;-)

By all means - I've been eyeing up the pretty shapes in RDFAuthor ;-)

>b) This 'folding' (cheers) algorithm is (I think) exactly what one
>group has
>been using here at HP. Very close, at least. Looks promising.

A small correction - I should have said that the
terminuses(/termini/terminals?) of the clusters could also be literals as
well as resources, just not properties or bNodes. Please let me know how you
get on anyway, with this or anything similar you come up with.

(btw, I didn't realise you were at HP - glad to see you weren't got by the
Email Inquisition)

Cheers,
Danny.