107 systemtype patterns

[dr-shorthair]

Patterns for System type descriptions, including complex sensors

Fixes #107

[dr-shorthair]

Previews:

• https://raw.githack.com/w3c/sdw-sosa-ssn/107-systemtype-patterns/ssn/index.html#ComplexSystems
• https://raw.githack.com/w3c/sdw-sosa-ssn/107-systemtype-patterns/ssn/index.html#SystemType

@ldesousa @sgrellet could you review?

[ldesousa]

@dr-shorthair I would like to clarify the question on rdfs:domain (and rdfs:range) before giving the green light. OWL defines those as constraints, and makes it clear:

An rdfs:domain axiom asserts that the subjects of such property statements must belong to the class extension of the indicated class description.

I understand this may be more nuanced with OWL 2, but SOSA only refers to OWL 1.

[dr-shorthair]

@dr-shorthair I would like to clarify the question on rdfs:domain (and rdfs:range) before giving the green light. OWL defines those as constraints, and makes it clear:

An rdfs:domain axiom asserts that the subjects of such property statements must belong to the class extension of the indicated class description.

Indeed this is correct: if there are global domain/range constraints associated with a property, then when the property appears in a graph, the subject/object must be a member of the domain/range classes.

However, because of the global constraint in the ontology, this class membership does not need to be asserted explicitly. This is the power of RDFS - a reasoner will infer the class membership.

OWL adds the ability to define guarded constraints. For example, 'allValuesFrom` constraints say that 'when this property appears in this context then the object of the property can be inferred to be a member of the specified class'.

In both cases, there is no need to assert the class membership explicitly, since it can be inferred from the ontology.

[ldesousa]

Alright. I wont block this any further. We can discuss inference some other time.

[sgrellet]

Apart from the small consistency question raised, looks fine to me