Re: Blank Node Identifiers and RDF Dataset Normalization

On Tue, 26 Feb 2013 17:46:13 +0000, Andy Seaborne <andy.seaborne@epimorphics.com> said:

    > 1/ When there is one graph being published
    > 2/ As the union of the named graphs
    > 3/ As a single place to put the manifest

    > Conclusion: you don't have to use it if you don't want to.

All of these can, as Kingsley pointed out, be accomplished by
configuring the query engine in a certain way. And that is all it
is. There is no "it" to want to use or not.

SPARQL can talk about (implementation/configuration dependent) default
graphs. RDF need not. SPARQL can still perfectly coherently talk about
ways that a user can write queries that don't specify which graph they
are interested in and what an implementation might do in that case.

In other words, the default graph is not some sort of thing that has
any meaning or existence outwith the SPARQL query language, it is just
a syntactic construct to make composing queries more succint.

Wherever the RDF specifications talk about the default graph, they are
actually talking about something that doesn't need to exist in the RDF
data model. It's unnecessary (and confusing) cruft.

Cheers,
-w