So, both Gregor and Marcus are suggesting reducing the number of properties for an InfoItem to the following four (related properties in parentheses):
- organismInteraction (associatedTaxon)
- subjectTag (category, context, contextValue)
- region (contextOccurrence)
- info (hasContent, hasValue)
I would prefer to reduce it to a *single* highly generic property like "info", "data", "generics", and introduce the different classes as as the possible types of this property:
<SPM>
<info> <Text><content> sdfsdfsdf</content></Text> </info>
<info> <OrganismInteraction> <interaction resource="pollination"/> <taxon resource="first taxon"/> <taxon resource="second taxon"/> <geoContext resource="Germany"/> </OrganismInteraction> </info>
<info> <Distribution> <status resource="neobiota"/> <geoContext resource="Germany"/> <geoContext resource="Netherlands"/> </Distribution> </info>
</SPM>
A single attribute term seems to simplify extensibility in the future (other classes, e.g. a specific to inform about available Genbank accessions, etc.).
Distribution: I suggest that Occurrence should be limited to the context of species distribution; I think it would be better to reserve Occurrence for actual occurrence as in the other TDWG standards using TaxonOccurrence. Distribution information of a species is a hypothetical, or synthetical information that could be accompanied by estimates of frequency, OccurrenceDensity etc. We could use the same term, of course, but my feeling is that we avoid misunderstandings if we separate terms here.
Furthermore, I would appreciate to allow the occurrence of RDF collections, i.e. the info/data/generics attribute could be directly in SPM or in a collection container (SEQ, BAG). Again, this may be an argument for a single attribute, although a sequence of different attribute may also be possible in RDF, I don't really know.
Gregor