Erratum clarifying sh:declare and owl:Ontology

[TallTed]

addresses #130 and #131

[wouterbeek]

The recommended subject for values of sh:declare is the IRI of the named graph containing the shapes that use the prefixes. These IRIs are often declared as an instance of owl:Ontology, but this rdf:type owl:Ontology declaration is not required.

For the second sentence I propose: "Notice that by asserting sh:declare, the subject IRI is also asserted to denote an OWL ontology."

My motivation:

• "These IRIs are often declared [...]" is an empirical observation that may be true or false, but whose truth is difficult to verify in practice. This empirical observation may also change over time, as practice changes. In my proposed formulation this empirical assertion is removed for this reason.
• While "this rdf:type owl:Ontology declaration is not required" is correct, it is also an obvious statement. Because any implicit assertion is not required to be expressed explicitly in RDF, I removed this from my proposed formulation.
• Since the second sentence acts as a notice to the reader, I use the word "Notice" explicitly.
• The thing the reader should be noticed about is that by asserting <s> sh:declare <o> they are also asserting that <s> denotes an OWL ontology. The latter is not obvious from reading the rest of the standard, so can act as a warning to readers who would otherwise have used sh:declare incorrectly. (Specifically, it would have warned me that sh:declare should not be used to relate prefix declaration to queries and other things that are not OWL ontologies.)

[HolgerKnublauch]

If you're OK I suggest closing this PR in favor of #133 - it is also easier to change the TTL file than the text of the spec.