[RFC] Data Hierarchy Refinement

While we’re at it:

I’ve been planning for this.

Do you have some suggestions on particulars of that design?

2 Likes

I’m sorry but I don’t have much. I know it will be necessary for the thing to make sense (why store stuff if we can’t do anything with it) but that’s all.

1 Like

It might be beneficial to aim for supporting RDF* (RDF Star) as it is currently undergoing an accelerated standardisation process, is central to Graph data standardisation. It is a pretty big deal and also a prerequisite for graph databases living on the Safe Network. A high level overview of what it’s all about plus pointers to the standardisation efforts underway here.

In this case it would require the support of “nested triples” so extra attributes can be assigned like provenance information which will be especially required on a decentralised permissionless network like Safe.

Luckily there are quite a few research/projects claiming various levels of success at decentralising and distributing triplestore query loads. With luck one or two of these methods work very well with the strengths of Safe Network and allowing graphs to have as you say “a more native feel on Safe”. A good contender for the killer app.

On the topic of scalable cloud based RDF storage @oetyng you might find this paper, “Towards a Scalable Cloud-based RDF StorageOffering a Pub/Sub Query Service” of interest due to your previous CEP work. In this case we would throw away the “compound event” quadruple hack work around to plain old RDF limitations and use (soon to be) standardised nested RDF* nested triples instead.

6 Likes