Skip to content

Decoupling kinds of evaluations and kinds of properties

SAREF4INMA defines an architectural ODP that consists in sub-classing s4inma:Measurement to specify the kind of measurement that is performed: s4inma:ActualMeasurement and s4inma:ExpectedMeasurement. Both subclasses end with string “Measurement” (naming ODP pattern). This could be generalized for other types of measurements such as temporal or logical aggregates, like “MinimumMeasurementOverTemporalInterval”, etc.

My proposal is to delete these two classes, as there is a more natural way to model things in SAREF V3.2:

the "expected" may be better modeled using a property value directly attached to the feature kind ("the ideal")

<item_type_A> a saref:FeatureKind; 
  saref:hasPropertyValue [ 
    saref:isValueOfProperty  <width> ; 
    saref:hasValue 12.0 ; 
    saref:isMeasuredIn unit:mm ] .

while the "actual" is modeled using a property value attached to the feature of interest,

<actual_item> a saref:FeatureOfInterest ; 
  saref:hasFeatureKind <item_type_A> ;
  saref:hasPropertyValue [ 
    saref:isValueOfProperty  <width> ; 
    saref:hasValue 12.02 ; 
    saref:isMeasuredIn unit:mm ] .

or evaluated using an observation

[] a saref:Observation ; 
  saref:observes <actual_item> ;
  saref:hasResult [ 
    saref:isValueOfProperty  <width> ; 
    saref:hasValue 12.02 ; 
    saref:isMeasuredIn unit:mm ] .