search

NIEM / NIEM-Releases

Repository for releases of the National Information Exchange Model
https://niem.github.io/niem-releases/
87 stars 23 forks source link

a JSON-LD serialization is a good step, but an ontology RDF representation of NIEM to match it would help more #191

Closed eric-jahn closed 1 year ago

eric-jahn commented 4 years ago

FHIR possesses a RDF ontology (for health care) https://www.hl7.org/fhir/fhir.ttl, and if NIEM had one as well, mapping between FHIR and NIEM could be easily done (perhaps using SSSOM https://github.com/OBOFoundry/SSSOM). To get other domains to join NIEM, would first involve them developing their own ontology, mapping to NIEM's ontology, then harmonizing them over time.

webb commented 3 years ago

We plan to generate an ontological / RDFS+OWL output from a NIEM model, based on the NIEM metamodel.

EmidioStani commented 2 years ago

What is the status of this issue ? Are you still planning to generate a RDF output ?

TomCarlson-NTAC commented 2 years ago

We are still planning RDF as one of the outputs from the metamodel, along with XML Schema and JSON in this initial phase.

Currently we're in an iterative process of writing proof of concept code which leads to changes in the metamodel, which leads to more code, etc...

Some sort of draft RDF output should be available by the end of 2021, albeit not the final version of the metamodel.

AtesComp commented 2 years ago

Adopting an RDF version of the ontology would be the most helpful. A JSON-LD format follows naturally from it as it will be readily translated to whatever transport format you like...Turtle, RDF XML, JSON-LD...the specification would no longer be bound to a specific format and become far more useful. Applications could request results from, say, a SPARQL endpoint in whatever transport format they want while still conforming to the established ontology.

eric-jahn commented 2 years ago

@AtesComp Right, since JSON-LD is an official standard RDF representation. And I think we can just skip the Common Model Format, and just use OWL to store the NIEM model and metadata (so OWL is not just another output, but the model). Then we just need to use SPARQL rules (maybe TARQL project's subset of work) to apply the NIEM Naming and Design Rules to get JSON (without the LD), CSV, legacy XML (not RDF XML), etc.. This is a common path, instead of a unique approach (CMF).

AtesComp commented 2 years ago

Exactly...and as I've commented in the NIEM Metamodel repo, OWL is the ultimate Metamodel. The RDF/OWL community is not format specific...Manchester Syntax anyone? It contains all the axiom rules needed to represent any other model format...transitivity, symmetry, asymmetry, functional, inverse functional, cardinality, equivalence, disjoint...and much more including all the complex type modeling built from it. Top this all off with INFERENCE models.

And the tools! Protege from Stanford University is a good open source tool for modeling ontologies. OpenRefine has RDF plugins. There are a ton of solutions in this domain that can be immediately applied to the NIEM model. There are even UML RDF tools.

NIEM should live up to it's by-line and truly share data by joining the Linked Open Data community.

AtesComp commented 2 years ago

I'll throw the SHACL standard out there as well.

eric-jahn commented 2 years ago

Yes, SHACL offers NIEM the validation functionality it currently enjoys with XML Schema, but SHACL is more flexible.

cdmgtri commented 1 year ago

Issue migrated to niemopen/ntac-admin#30