Provide best practice guide for using schema.org properties as annotations

[cmungall]

schema.org is increasingly used as a metadata vocabulary. What schema.org properties should be mandated, recommended, allowed for OMO? Are there any technical obstacles?

Proposal: existing OMO properties (from namespaces like IAO, oboInOwl, rdfs, owl, and dcterms) that are in established use should remain in place and used in place of schema.org equivalents. The gains from interoperability with the wider semweb are IMO not worth the churn in switching out. However, more bespoke schema.org properties for which there is no existing OMO equivalent may be used.

Technical issues:

1. schema.org is RDFS(like) not OWL, and hence doesn't commit to AP vs DP vs OP. This can in theory cause problems. We already had this discussion re skos with no completely satisfactory resolution Add guidelines for use of skos in OMO #65
2. schema.org is still undecided as to whether to use http or https. This sounds minor/obscure but could in fact lead to large problems if we are to adopt for important properties. If anyone does use schema.org I strongly recommend http namespaces as these seem to be more canonical. See schema.org http or https? solid/solid-namespace#21

[graybeal]

Agree with suggested approach. A third issue with schema.org (and bioschemas) in my casual review was that controlled vocabulary terms often were non-existent or insufficient, forcing additional profiling to ensure interoperable annotations.