[open-bibliography] Library support and REST

[Ben O'Steen]

On Tue, 2010-10-26 at 06:26 -0700, Karen Coyle wrote:
> I don't know how we solve this dilemma, but to me this puts OKD-Open  
> in a "pre-cloud" position 

I wouldn't say that this puts the OKD effort on a timeline, from legacy
to 'cloud' - we are looking for datasets that have yet to publish
openly. 

I don't think that per-triple provenance solves current problems - there
are many much much bigger fish to fry. The problem it would solve is to
ascertain provenance for triples from joined datasets. A quad store (or
greater) index would let you straightforwardly query this information
out. 

We can certainly attribute, provide provenance and characterise the
named graphs, the 'documents' that contain the data and a quad store
would provide the additional SPARQL keywords of FROM and GRAPH with
which to augment the queries.

The bigger problem is getting data as PDDL, CC0 or similar licence so
that you can merge datasets for more than just personal inspection - at
least in the EU.

Ben