- From: Simon Spero <sesuncedu@gmail.com>
- Date: Fri, 16 Jan 2015 15:22:35 -0500
- To: Dan Brickley <danbri@google.com>
- Cc: W3C Web Schemas Task Force <public-vocabs@w3.org>, Jarno van Driel <jarnovandriel@gmail.com>
- Message-ID: <CADE8KM55ph8ybQavHy1u9jMF-zOZt_mgT1gtdoCHSq=5H0n90w@mail.gmail.com>
For some enumerations, the instances are defined as part of schema.org. The values of these enumerations can be found by looking for individuals in the canonical RDFa whose type is the enumeration, rather than being classes or properties [ignore True and False]. Other enumerations do not have values defined in schema.org, but instead make use of external sources of values. In some cases, the IRIs of the values are contained in the rdfs comment. In other cases, the comment contains the URL for a website that can be used as a source of values. There is no algorithmic mechanism for obtaining the values of this kind of Enumeration. Most enumerations classes are direct subclasses of Enumeration. The main exception is MedicalEnumeration, which is a subclass of Enumeration, but which has no direct instances. Enumerations can be closed (only listed values may be permitted), or semi-open (values other than those listed may be permitted but may not be understood). This is not always formally stated. Enumeration values are instances of the given enumeration class (and are thus implicitly Things and Intangibles). In some cases the values might be classes; however these classes are not subclasses of the Enumeration class. Thus, http://schema.org/Virus is an intangible, but if it were made a real class, it would not have instances that were intangible. [Ignore True and False]. Simon P. S. My reason for ignoring True and False is that they are the only things in schema that are not declared to be classes, but which have subClassOf assertions (and which are the only triples that my sdo parser doesn't consume, so they keep generating warnings).
Received on Friday, 16 January 2015 20:23:02 UTC