Liam, hello.
On 5 Oct 2016, at 18:01, Liam R. E. Quin wrote:
> It turns out even RDF has moved away from triples - an RDF "named
> graph" associates (implicitly) additional information with each
> triple.
>
> For example, consider needing to track the date on which each fact was
> entered into a database. If you store that as a separate triple (on
> such-and-such a date triple-id foo was added by bar (oops that's
> four))
> you need to record when that information was added and by whom.
I think that might be an exaggeration. Some _implementations_ of triple
stores might usefully store such extended provenance information, in the
same way that some XML parsers or databases might document that they
preserve attribute order. That doesn't meant that 'XML has moved away
from unordered attributes'.
'Named graphs' are different. There, there's a mild extension (formally
standardised, I _think_) of the RDF model, to include the name of the
graph in which the statement 'Alice knows Bob' is asserted. Such graphs
are supported in 'quad stores'. Such provenance turns out to be rather
painful to express in bare RDF, to the extent that the extension is
often useful.
All the best,
Norman
--
Norman Gray : https://nxg.me.uk
SUPA School of Physics and Astronomy, University of Glasgow, UK
