Add wktLiteral @type for geojson:geometry

[sgillies]

Towards closing #32.

This helps RDF users get usable RDF by coercion of GeoJSON geometry to WKT.

It does nothing for RDF -> GeoJSON, but I don't think this lack has to be a blocker for a first release of GeoJSON-LD on the web.

Definition of a new container types specific to GeoJSON geometry objects and N-D arrays could assist with RDF -> GeoJSON conversions as well as serialization of geometries to pure RDF, but let's defer work on those.

Reviews welcome!

[pietercolpaert]

I'd vote against merging this as:

• The semantics are not right: what's given in the JSONLD is not a WKT string
• I'm still not convinced about fixing this with vanilla JSONLD parsers, as @dinizime puts it, in Coordinates to RDF doesn't work #32. I'm more in favour of fixing this by introducing something like @values in JSON-LD, which describes the values of multidimensional arrays.

[pietercolpaert]

I suggest waiting for json-ld/json-ld.org#397. If this proposal gets shot by the JSON-LD community, you can go ahead and merge this.

[sgillies]

@pietercolpaert thank you very much! I'll follow that issue.

[akuckartz]

@sgillies Thanks for trying to resolve these difficult and often controversial issues!

[pietercolpaert]

I don't understand how the @type is a WKT Literal, as it isn't: it's a JSON array. No?

[sgillies]

@pietercolpaert my intent is that @type would trigger coercion to WKT as in http://www.w3.org/TR/json-ld/#type-coercion. Am I misreading that section?

[pietercolpaert]

@sgillies No, you aren't. My bad!

[sgillies]

@pietercolpaert I'm glad to get second opinions. Sometimes I feel lost without a compass in JSON-LD!

[sgillies]

@pietercolpaert it looks like we should merge and continue without creation of a new media type. Agreed?

Related: I've read the discussion in w3c/activitystreams#52 and have a better understanding of where many commenters here are coming from.

[sgillies]

Closing.