Re: Should information be merged from several RDF files?

On Jul 5, 2014 5:14 PM, "Victor Porton" <> wrote:
> I work on this project:
> It involves downloading RDF files from the Web.
> When a "bundle" (for lack of a better word), that is a set of related RDF
triples, is incomplete it should be ignored.
> My question, if a bundle is split into several different RDF files, and
each part of it is incomplete, should it nevertheless not be ignored if the
union of all parts of the bundle is complete?
> A toy example (A, B1, B2, C be namespaces):
> In one file:
> <A> <B1> <C> .
> In an other file:
> <A> <B2> <C> .
> If both B1 and B2 properties are required, should this information be
ignored (as incomplete bundles)? or should the information from the two
files be merged and thus considered complete?

People are encouraged to dream up new governance rules for RDF simply
because it offers a new fluidity in merging data. I think that's a bit of a
red herring; we incorporate data based on whether it's accurate and
informative. RDF itself demands no particular rules with regard to
completeness or even accuracy. You may trust one source's assertions about
one topic but not another. You have to decide whether it's useful to
incorporate the B1 and B2 assertions from different files.

That said, if you are experimenting with some sort of validation rules, you
may be interested in Shape Expressions and Resource Shapes.

> --
> Victor Porton -

Received on Saturday, 5 July 2014 17:11:34 UTC