opengeospatial / NamingAuthority

Primary repo for the OGC Naming Authority
6 stars 12 forks source link

Units registered in the OGC Rainbow #263

Closed cportele closed 7 months ago

cportele commented 7 months ago

The Features API SWG requires a language to unambiguously encode units of measurement. We also looked at the units registered in the OGC Rainbow and were quite confused by the fact that

After a closer look I understand that the multiple instances are for meter in different authorities (and for URIs that do not conform to the policy, but are probably maintained for legacy reasons).

However, it was still unclear to us what the status of the OGC definitions for units is. Especially since only very few units are registered. Is the conclusion correct that the recommendation is to use more widely used approaches (UCUM, QUDT Units) instead of an OGC-specific one? That is, the OGC units are basically deprecated and only maintained because they were registered at some point in the past?

dr-shorthair commented 7 months ago

FWIW QUDT is actively maintained. There is a public issue tracker, many contributors https://github.com/qudt/qudt-public-repo/graphs/contributors and PRs are processed frequently - https://github.com/qudt/qudt-public-repo/pulls?q=is%3Apr+

ghobona commented 7 months ago

The outcome of a similar discussion regarding GeoSPARQL is here.

In that situation, the decision of which UOM URI to use was delegated to the GeoSPARQL SWG.

The GeoSPARQL SWG decided to use QUDT. So there is a precedent for using QUDT.

cportele commented 7 months ago

Thanks for the reminder about the old discussion, @ghobona. My conclusion from that discussion is that we (Features API SWG) should use established approaches for expressing units, which likely means UCUM or QUDT Units. For now, we have settled on UCUM as the default, because it is a general grammar for expressing units and URIs for the units are not important.

I am closing this issue again, because the question about the units in the Definitions Server has been answered. Thank you.

dr-shorthair commented 7 months ago

Note that the guys at St Etienne have specified an RDF Datatype for a microformat using UCUM. This has some uptake. https://w3id.org/cdt/ucum

cportele commented 7 months ago

Interesting @dr-shorthair. This can be useful for cases where the unit cannot be constant and where filtering on the values is not important.