RDFS vs. OWL Reasoners - Design Considerations

Question

I read a number of RDF related questions here at StackOverflow that have an RDFS solution, but also an OWL solution. (Noteworthy, my own question )

I understand the reasoning concept and that OWL describes complex relationships in more detail, but this comes at a cost, which is often called update performance.

Can someone please explain what factors are involved in making a constructive decision about whether to use the OWL argument and how to evaluate whether a sufficient set of requirements is sufficient for RDFS analysis. (I am seriously considering AllegroGraph RDFS ++ now).

I already started working on the ontology and already defined the triple using things like owl:ObjectProperty , owl:FunctionalProperty , and has several owl:Thing subclasses. Does this mean that I have already passed the threshold and need quick-witted OWL? Where does this line cross and why?

Answer 1

When you start modeling using RDFS and OWL, you can use various types of constructs (or axioms) (for example, rdfs:subClassOf or owl:FunctionalProperty ). Each of these constructs adds some computational complexity as to what reasoners should do, such as consistency checking or classification. As a rule, the more expressive the constructions, the more time is required to calculate what they entail.

When you use the constructs available in RDFS , you limit the type of things you can express, but instead you can use highlighted efficient arguments, such as AllegroGraph.

However, the RDFS judge will not be able to calculate the axioms of OWL. OWL is more expressive than RDFS, and an OWL reasoner must deal with a large number of constructs. OWL concepts understand the axioms of RDFS (since all RDFS constructs are part of OWL), so you can use the OWL argument with RDFS.

So in short:

• The more expressive you want to be, the more difficult it becomes to compute what these constructions entail.
• RDFS is less expressive than OWL, so if you only stick to RDFS constructs, you can expect better results. But sometimes you need OWL to express what you want, the job of a designer is to find a good compromise.
• OWL2 presents some profiles ( OWL2 EL, RL, and QL ), focusing on a specific set of axioms to guarantee effective reasoning. You have a special argument about these profiles (e.g. ELK for OWL2 EL).
• You can read the W3C documentation to learn more about the computational properties of RDFS and OWL.