OM_Observation used as property values
There are a couple of things that needs to be set in the way we encode OM_Observation, specially when we use them as property values.
There are two cases when we use OM_Observation as values
- When using relatedObservation on any sub types of SF_SamplingFeature
- Using a specific property (such as GW_FluidBody/gwBodyVolume)
OM_Observation has a mandatory featureOfInterest that is supposed to point back to the feature that is the bearer of the property.
In case 2) , the bearer of the property is obviously the feature having this property, so technically, we should constrain featureOfInterest to link back to that feature
Eg: we should technically have this
Issues: (Question in red)
Cases where the bearer of property is not a feature
(GW_WaterBudget for instance, is a Type, not a Feature), so we cant use featureOfInterest to link it back
In this case, we should point back to the refers to this Type ? In case of WaterBudget, who does ?
Same Observation shared by multiple properties
nothing prevents more than one property to point to the same Observation (by reference), which becomes a problem for the featureOfInterest (foi is 1..1). It can be done schematically even if does not make sense logically.
Should we have a clause that says OM_Observation that are values of explicit properties will use featureOfInterest to identify the legitimate bearer of properties. Reference to OM_Observation from an explicit property when the property bearer is not the owner is invalid (should it be invalid ?, or can it have another meaning)
OM_Observation also has an observedProperty. Technically, this should be a reference to the property (the property type, not a specific instance of the property) this observation is about. We have several options (
which one should we use)
- Use a reference to schemas - like this "http://ngwd-bdnes.cits.nrcan.gc.ca/service/gwml/schemas/2.1/gwml2-well.xsd#GW_WellType.gwWellConstructedDepth (there are no good way to link to a property, it has no identity - (apparently, new ISO 19109 will change this)
- Use a good old vocabulary unrelated to domain schema - with the benefit of having more precise definition of the phenomena than the schema provide
--
EricBoisvert - 07 Jul 2015