Closed baskaufs closed 8 years ago
Since indication of whether an RDF document is the normative document for a standard would be included in this issue, I'm blocking it on Issue #5
Reference Recommendation 5.5. of the VoMaG report, which strongly recommends use of minimal metadata element set for LOD vocabularies at http://lov.okfn.org/dataset/lov/Recommendations_Vocabulary_Design.pdf and listed in Table 1 in the report.
Reference Recommendation 5.6 of the VoMaG report about properties for parent vocabulary, listed in Tables 2 and 3 of the report.
http://lov.okfn.org/dataset/lov/Recommendations_Vocabulary_Design.pdf presently returns 404 FileNotFound and I couldn't find a replacement in a few minutes of looking on lov.okfn.org. There is a copy on the author's github https://github.com/pyvandenbussche/lov/blob/master/public/Recommendations_Vocabulary_Design.pdf, but if this document was never adopted by okfn, or someone influential, it is questionable whether TDWG should bless it..... Hopefully there is a way to find it or a replacement on okfn.org or someplace worthy of citing...
Refer to Data Catalog Vocabulary (DCAT) for vocabulary used to describe Datasets, Catalogs, and CatalogRecords.
At the 2015-07-15 TG meeting, we discussed a possible hierarchy model that outlines the components of standards and vocabularies. The terms dcterms:hasPart/dcterms:isPartOf, owl:imports, and rdfs:isDefinedBy are used in the model to inter-relate the components in a way that facilitates their discovery. Resource versioning and status is discussed in a possible versioning model. Steve has been working on this.
I think this issue is pretty much covered now in the draft document. Most of the properties in the Metadata Recommendations for Linked Open Data Vocabularies have been followed, as well as the structure of DCAT. However, I did not include typing of vocabularies as voaf:Vocabulary because its definition specifically says that the vocabulary is used in the Linked Open Data cloud and voaf:Vocabulary is a subclass of void:Dataset, which is a set of RDF triples. I felt that TDWG vocabularies were broader than that and instead recommended typing them as dcterms:Dataset.
This issue is related to Recommendation 2.12. of the VoMaG report.
Experience with the TDWG Ontologies has shown the need for guidelines for creating links to authoritative information about vocabularies and their terms. The RDF/XML documents existed in multiple locations in differing forms that had the same version numbers and dates. The documentation was scattered around the TDWG website and Google Code site. Recommendation 2.12 suggests that the Documentation Specification Standard should specify the properties that should be used in RDF documents to specify the nature of the term or document (current, superseded, deprecated, etc.), the version number, the last modified date, licensing information, links to human-readable documentation, links between the term URI and vocabulary URI, etc. A client dereferencing a term URI should be able to follow its nose and discover all of this information.
The draft Standards Documentation Specification provides guidelines only for the formatting of human-readable documents, so we will need to look elsewhere for precedent (e.g. DwC, AC, DCMI).