Re: RDF Interface specification

On 03/07/14 16:22, Miguel wrote:
> I think that finishing, publishing and possibly evolving this spec
> would be useful not only for Javascript developers.
I agree it would be good to evolve the spec.

In particular, I'm missing a query API. The interface, as it is now, is
useful for only small datasets. As soon as I need to make a query, which
in my experience is most of the time, I am outside the spec.

Jakub

> In fact I think that the ongoing effort of harmonization of Java RDF
> interfaces cited by Nandana should be connected to the RDF Interfaces
> work.
> The existence of a common standard inter-language interface (as for
> XML DOM) is good for example when a developer needs to switch from a
> language to an another (he does not have to learn again the
> interface).
>
> Some ideas of possible changes to the RDF Interfaces Spec:
> 1) As the RDF 1.1 spec defines the RDF Dataset, I think that it makes
> sense to add a Dataset interface to the RDF Interfaces spec.
> 2) In RDF 1.1 spec a RDF Graph (and thus a RDF Dataset) is immutable,
> so I would move the methods with side effects from the Graph interface
> to a GraphStore interface. Does it make sense?
> 3) Again RDF 1.1 spec informally defines a RDF Source as a "persistent
> yet mutable source or container of RDF graphs", basically a time-based
> sequence of snapshots, represented by RDF graphs. For example a Graph
> Store can be seen also as a (mutable) set of RDF Sources (the default
> one and the named ones). A RDFSource interface could be used as the
> subject of an Observer Pattern delivering change events when its
> content changes (a change event can be represented as the set of added
> triples and the set of removed triples with respect to the previous
> snapshot). So in the end there would be different interfaces of
> (loosely speaking) growing richness: immutables (Graph and Dataset),
> mutables (RDFSource) and writables (GraphStore).
>
> Best Regards,
> Miguel
>

Received on Thursday, 3 July 2014 21:57:37 UTC