Labeling of entities in DiCon
In DiCon, all entities can be labeled with multiple different identifiers and classifications, following the philosophy of gist ontology: the labels are represented as data, instead of defining large numbers of new data properties or classes for them. Flexible labeling is important capability since labels can communicate the users as well as external systems what the entity in question is and what kind of entity it is. A simplified diagram of the most important classes and properties is shown below.
Identifiers
URIs as primary identifiers
Each independent entity represented according to DiCon ontologies has a URI as its primary identifier (Please note that there can also be subordinate entities that are anonymous and represented as blank nodes at the RDF level). Apart from identifying the entity, the URI can be used to retrieve the description of the entity.
Each URI is unambiguous (refer to one real-world entity) but not necessarily unique (there can be different URIs that refer to the same real-world entity). For example, http://company1.com/ElmStreet1 and http://project1.com/Building1 could be two URIs that refer to the same real-world building. Each of the URIs may return information
http://company1.com/ElmStreet1 -->
http://company1.com/ElmStreet1 a dice:BuiltAsset ;
dica:hasOwner http://example.com/Owner1 ;
...
http://project1.com/Building1 -->
http://project1.com/Building1 a dice:Building ;
dice:hasAgent http://company1.com/Contractors/Contractor1 ;
...
These URIs can be declared with owl:sameAs axiom as references to the same entity. In a RDF database with proper support for reasoning, the result is a merging of properties from the two sources:
http://company1.com/ElmStreet1 owl:sameAs http://project1.com/Building1 .
http://company1.com/ElmStreet1 -->
http://company1.com/ElmStreet1 a dice:BuiltAsset, dice:Building ;
dica:hasOwner http://example.com/Owner1 ;
dice:hasAgent http://company1.com/Contractors/Contractor1 ;
...
The description above is meant to remind that in Linked Data systems URIs are the primary identifiers of entities. They can be made retrievable, they are required by the Linked Data Principles, and the reasoning works based on equivalence defined by them. However, they are not sufficient for all practical settings related to user interaction or cross-system access.
Additional identifiers
Any real-world entity can have distinct identifiers in different systems and contexts. Often it is essential to know these identifiers to be able to refer to the entity in user interfaces, when accessing entities in those systems or using them in the specified contexts. Some of the identifiers are globally unambiguous (such as GUID or DID) and some locally unambiguous. Examples of local identifiers are room numbers in a building, control point numbers in a building automation system, user names in a communication application, identifiers in a local database, and so on. Many of the commonly used identifier types are not URIs; in general, they can be any strings.
In detail, any instance of the class dice:Entity (or some of its subclasses) can have additional identifiers of the class dice:Identifier connected to through the object property dice:isIdentifiedBy.
An example of a labeling is the following. The primary identifier of an entity is http://company1.com/ElmStreet1 and it has an IFC GUID and an asset identifier as additional identifiers. The residental unit http://company1.com/ElmStreet1Apt1 in the building has the apartment number as a local identifier in the scope of the building. It should be noted that the identifiers are always with respect to some scope that can be an identifier category or an entity which hosts other identified entities.
@prefix : <http://company1.com/> .
@prefix dice: <https://w3id.org/digitalconstruction/0.5/Entities#> .
:IfcGlobalId a dice:Category ;
rdfs:label "Ifc global identifier (GUID)" .
:AssetId a dice:Category ;
rdfs:label "An asset identifier in the Asset Management System" .
:ElmStreet1 a dice:BuiltAsset ;
dica:hasOwner http://company1.com/Owner1 ;
dice:isIdentifiedBy [ a dice:Identifier ;
dice:hasLabel "21EC2020-3AEA-4069-A2DD-08002B30309D" ;
dice:hasScope : IfcGlobalId ] ,
[ a dice:Identifier ;
dice:hasLabel "987654" ;
dice:hasScope :AssetId ] .
:Elm1StreetApartments a dice:EntityScope ;
dice:hasEntitiesFrom :ElmStreet1 .
:ElmStreet1Apt1 a dice:ResidentialUnit ;
dice:isBuildingUnitOf :ElmStreet1 ;
dice:isIdentifiedBy [ a dice:Identifier ;
dice:hasLabel "101" ;
dice:hasScope :ElmStreet1Apartments ] ,
[ a dice:Identifier ;
dice:hasLabel "987654-1" ;
dice:hasScope :AssetId ] .
Classifications
In ananalogous fashion to identifiers, classifications using dice:Category can be associated to entities through the object property dice:isClassifiedBy. There are many different classification or coding systems - OmniClass, ETIM, CoClass, Talo2000, or GS1 codes - that specify the types of entities. The following example shows the use of classification with product data:
@prefix : <http://company1.com/> .
@prefix dice: <https://w3id.org/digitalconstruction/0.5/Entities#> .
:ETIM a dice:Category ;
rdfs:label "ETIM classification for technical products" .
:BSDD a dice:Category ;
rdfs:label "bSDD classification" .
:GTIN a dice:Category ;
rdfs:label "GS1 Trade Item Number for a class of trade items" .
:InternalProductId a dice:EntityScope ;
dice:hasEntitiesFrom :InternalProductDB ;
rdfs:label "Internal product database" .
:Lamp1 a dice:BuildingObject ;
dice:isIdentifiedBy [a dice:Identifier ;
dice:hasScope :InternalProductID ;
dice:hasLabel "pr0233.98723878822" ] ;
dice:isClassifiedBy [a dice:Category ;
dice:hasScope :ETIM ;
dice:hasLabel "EC000819" ] ,
[a dice:Category ;
dice:hasScope :BSDD ;
dice:hasLabel "1YlGK0inSHuO00025QrE$V" ] ,
[a dice:Category ;
dice:hasScope :GTIN ;
dice:hasLabel "012345.67890" ].
Different classification systems can contain thousands of detailed categories for different kinds of entities. The association of these categories to entities in a generic fasion enables to keep the DiCon ontologies simpler. The ontologies can focus on the complexity of the domain, while the classifications take care of the variety in it.
Last modified: Tue Sep 7 16:53:22 EEST 2021