search

w3c / vsso

The Vehicle Signal Ontology (VSSo) based on COVESA VSS and making use of the SOSA/SSN modeling patterns
Other
15 stars 12 forks source link

Proposal Core Ontology: belongsToVehicleComponents for static property relation not according to OWL #8

Closed danielwilms closed 2 years ago

danielwilms commented 3 years ago

Thanks to the discussion today, we highlighted the vsso:VehicleComponent relation from the vsso:StaticVehicleProperty to the vsso:VehicleComponent. It was correctly argued, that this is not owl compliant. This brings me to the discussion, whether this information is actually needed.

The static properties define the vehicle itself and the connection to the component might just be overhead. There are two example queries (Q08, Q11), which would not be resolvable by dropping the relation. So, one thing is how to keep the information and how to model it formally correct. The other question is, whether the information is actually needed.

Happy to hear your opinion. #5

danielwilms commented 3 years ago

Apart from the question, if the information is needed, modelling the property as an ObjectProperty was indeed wrong and let to some confusion. The reason for the property was, that the branches as individuals are providing more context to the static and dynamic vehicle properties. I still think, that this approach is valid and that belongsToVehicleComponent could be actually modelled through a owl:AnnotationProperty. We'll push the code and the model later today, so that you can see how it looks like. What do you think?

rtroncy commented 3 years ago

I don't understand the paragraph in https://github.com/w3c/vsso/issues/8#issue-804965847. What is vsso:VehicleComponent? Certainly NOT a property as it is written. What is vsso:StaticVehicleProperty? None of those terms are defined in this turtle file.

What would be the definition of an hypothetical vsso:belongsToVehicleComponent property? A owl:AnnotationProperty conveys no semantics!

danielwilms commented 3 years ago

@rtroncy, it's about the core ontology proposal. Maybe I should keep the discussion in only one issue. See #5.

jdacoello commented 3 years ago

What @danielwilms is referring to is the following:

We made a proposal for the core Vehicle Signal Specification Ontology (VSSo) as an approach to keep the ontology fully synchronized and aligned with the continuous evolution of the Vehicle Signal Specification (VSS). The principal elements of the semantic model consist of:

With them, we can resemble the tree-like structure of VSS. The issue corresponds to a recent observation about one of the properties proposed (i.e., belongsToVehicleComponent). The issue was that we wrongly connected a the data property hasStaticVehicleProperty with the class VehicleComponent using a the object property belongsToVehicleComponent. Such a connection was not compliant to OWL. Therefore, we changed belongsToVehicleComponent from being an object property to an annotation property. All the queries were tested with this adjustment and are working. Pull request was already made and demo repository should be updated soon.

Another option we tried and that could be also a valid solution with more semantics is to keep belongsToVehicleComponent as an object property and add the class StaticVehicleProperty. In this way, each fixed attribute of a vehicle defined in VSS would be mapped to a subclass of StaticVehicleProperty.

felix-loesch commented 3 years ago

A solution proposal to make VSSo OWL compliant that we Felix, Anees (Bosch) discussed with @danielwilms is as follows:

This approachs makes VSSo OWL compliant as object properties, especially belongsToVehicleComponent is used according to OWL standard, i.e., relating the two classes StaticVehicleProperty or DynamicVehicleProperty and VehicleComponent

danielwilms commented 2 years ago

Included in current proposal