Json-LD instances test : Avoid having a graph as an object for properties.

[afeliachi]

Properties should have resources or literals as objects: Use list instead graphs if the property has many resources/literals.
Example : https://opengeospatial.github.io/ELFIE/FR/Piezometer/sgwlf/04398X0002-SONDAG.2.json
If we try in json-ld validator (N-quad view) : for the property "isFeatureOfInterestOf" , we’ll see that a blank node is created (the graph). That’s not what we want. We want to point to the observation itself or the list of observations

Should be declared
"isFeatureOfInterestOf": {
"@context": "https://opengeospatial.github.io/ELFIE/json-ld/elf-sosa-observation.jsonld",
"@id": "http://ressource.brgm-rec.fr/obs/RawOfferingPiezo/04398X0002/SONDAG.2",
"@type": "sosa:Observation",
"hasFeatureOfInterest": "http://ressource.brgm-rec.fr/data/Piezometre/04398X0002/SONDAG.2",
"hasResult": [
{
"@id": "http://ressource.brgm-rec.fr/data/Piezometre/04398X0002/SONDAG.2&responseFormat=application/json",
"@type": "wml2:MeasurementTimeseries"
}
],
"madeBySensor": "http://ressource.brgm-rec.fr/data/Sensor/should_put_serial_number_here",
"observedProperty": {
"@id": "http://id.eaufrance.fr/par/1689.xml",
"@type": "skos:Concept",
"name": "Goundwater Level (in meter)"
},
"phenomenonTime": {
"hasBeginning": "2016-01-05T11:00:00.000Z",
"hasEnd": "2017-10-10T10:00:00.000Z"
},
"resultTime": "2017-10-10T10:00:00.000Z",
"usedProcedure": {
"@id": "http://id.eaufrance.fr/met/403.xml",
"@type": "skos:Concept",
"name": "Electronic piezometric probe"
}
},

[dblodgett-usgs]

@sgrellet and @afeliachi -- has this been taken care of or are you still working on this stuff?

[afeliachi]

This has been taken care of from our side in the sample files. The best practices should be edited to reflect that.

[abhritchie]

Thanks for the @afeliachi - would you, or @sgrellet, mind updating the best practices? I'm not confident I understand well enough to do it myself.

[sgrellet]

Sure we'll do

[dblodgett-usgs]

Looks like this is started or done in #163 ?

[dblodgett-usgs]

@abhritchie -- if you have a sec, can you look at the update to best practice docs commited in #163? If happy, move to "Done".