RDFS Entailments - how sound is your Graph Engine really?

Producers of RDFS?enabled Graph Stores claim their Store does have a (functioning) ?truth maintenance“ system, which correctly computes the RDFS entailment for RDF(S) Graphs. The (standard)?RDFS entailment rules?instruct the engine in your RDFS Graph Store how to perform (standard) deductions to discover (infer) implicit assertions in RDF graph(s). Unluckily and despite W3C’s Graph Store recommendations, this is not always the case.

Why is RDFS Entailment so important?

RDFS allows you to derive intensional RDF data (e.g. RDF data which is not explicitly written) out of the RDF statements of your Graph. As a very simple example, consider the two simple RDF Statements:

(:Tim :eats :Apple) and (:eats rdfs:subPropertyOf :ingests)

RDFS entailment allows to automatically infer (entails) via rule rdfs7 that:

(:Tim :ingests :Apple)

Even if this is not explicitly written in the previous considered example RDF statements. RDFS entailment hence derives intensional properties out of (opportunely shaped) RDF data. Once further RDF statements are inferred, we often say they are "materialised". This said, it seems straight-forward that RDFS can be an important "companion" for complex materialisations along a linked-data space in an RDFS enabled repository.

How can I test, whether my RDFS Store is really W3C compliant (*)?

Semweb has published a (free) service to test whether your RDF(S) Store is W3C compliant with respect to RDFS entailment or not, giving you the information, which single entailment was detected to be not working.

Here you find the service.

Thanks a lot for reading

Best regards

Yours Fabio Ricci from Semweb

---

(*) Or if it is simply correct configured for RDFS entailment.