Closed dr-shorthair closed 11 months ago
This requirement implicitly follows from the view that SOSA is primarily understood as an implementation of OMS.
Proposal is to create an example of a documentation link to OMS - if this is acceptable the next step is to formalise this pattern and enforce it - this can be done by applying a SHACL validation rule of the ontology (sosa.ttl) itself, in the same way we will validate examples.
We will need to revisit the supply of OM conformance class normative URIs to the OGC publishing infrastructure. These should be lodged to the SOSA repository as a resource to be integrated into validation process.
Here are three (nine?) options for a potential approach: https://github.com/w3c/sdw-sosa-ssn/blob/247a0dc1b3a5ab86ce8f9f1e86384a473141b242/ssn/integrated/sosa.ttl#L437-L439
As well as choosing a suitable predicate for the link, we need to understand which of these three potential requirements from the UML models (concept, core, basic) is actually being implemented here.
@Jano suggests that OGC/ISO should coin its own alignment axioms, with sub- or super-property relationships to the common vocabs to support inferencing.
@sgrellet suggests that the mappings should always be to the conceptual model.
on where to point to in oms I'd suggest the following sequence 1°/ point to oms-bs-cpt -> Conceptual model 2°/ point to oms-obs-core -> Abstract model, only when a 'class' appears for the 1st time (= not present at conceptual level) 3°/ not use basic because 'basic' intent is to have a 'ready to roll' UML implementation of OMS and should not create new semantic artefact
And apply the same logic for the Sample part
See #104 for completed version of SOSA-OMS graph with explicit mappings.
See #116
Use a uniform method to document SOSA-OMS alignments within the machine-readable artefacts (i.e. the RDFS/OWL).