ogcscotts / TC-Meeting-topics

place to discuss topics raised by Working Groups
10 stars 0 forks source link

Principles for JSON based data OGC model implementation standards: Coupling with particular APIs, common linking patterns, guidance on if and how to combine GeoJSON with JSON-LD, mapping UML feature types into JSON objects: one solution (complex vs. simple) does not fit all #173

Open ogcscotts opened 3 years ago

jechterhoff commented 3 years ago

Regarding mapping UML feature types into JSON Schema: If you have a data model defined in UML then there are ways to automatically encode that model as JSON Schema. The OGC UGAS-2020 Pilot investigated and implemented such JSON Schema encoding rules. They are applicable to ISO 19109 compliant application schemas as well as some ISO schemas (e.g. ISO 19115, 19157).

If the JSON encoding shall represent the original data model to a large extent, then the automated encoding is useful. If the JSON Schema representation shall be a simplified, specific encoding that merely realizes the conceptual model (to a certain extent) - an example would be a geometry or coverage encoding - then such a JSON Schema will need to be defined manually. Still, some of the principles defined by the JSON Schema encoding rules might be useful building blocks for such work.

The UGAS-2020 Engineering Report is currently available on pending docs and will be published soon. The open source tool ShapeChange implements the encoding rules.