Closed autermann closed 4 weeks ago
It was kind of intended :-) When I wrote the JSON schemas, I thought maybe we should simplify this part, but to be honest I was unsure about this. So I'm glad you brought this up.
I think the Text option can still be useful.
Certainly most cases of DataRecord
usages would now be handled by the new GeoPose
/ RelativePose
options introduced in JSON.
And my thinking was that now that we live in an API world, pointing to an AbstractProcess
or DataArray
does not make as much sense, and I would rather reference the datastream that provides the position data (whether it is provided by a process or a GPS sensor doesn't matter).
I'm open to suggestions though. Do you have use cases where you use the advanced options? Do you think we should try to stay 100% compatible with the XML encodings?
I don't have a strong opinion on this.
I think we could also keep all options and mark some as deprecated in the JSON encodings. And deprecated would mean implementations are not required to implement them.
Discussed during 02/15 Telecon. We will add back support for all SensorML 2.1 options into the JSON schema to maintain full compatibility
See 7e46046dbc1aa39e268b6c9981b4a8b453b4d4cc
Reopening to make sure I don't forget
Nevermind it's done
The SensorML 2.1 XML encoding allows the position of a physical process to be one of the following:
gml:Point
swe:Text
swe:Vector
swe:DataRecord
swe:DataArray
sml:AbstractProcess
While the JSON schemas only allows GeoJSON
Point
, GeoPose andswe:Vector
.Is this intended? Should be trivial to include the other definitions in the
oneOf
.