tutorials.Understanding-Semantics.md

Understanding Semantics

This tutorial introduces basics of vocabularies to describe things in a certain domain. These vocabularies ensure that applications speak the same language (are semantically interoperable) and are one of the prerequisites to publish Linked Data. For NGSI-LDF, Linked Data aligned with vocabularies enable interoperability with other Digital Twin components.

We will introduce the RDF data model and investigate how these are applied in the Schema.org, NGSI-LD, OSLO and W3C ecosystems. Then, we will explain what Application Profiles are and how they can help developers achieving data compliance.

In the next tutorial, we will create a OSLO-compliant data snippet of a Building which we will use further in the tutorials.

RDF

RDF or the Resource Description Framework is a standardized model to describe knowledge. Knowledge is represented as a set of facts, which are described with triple-statements.

For example, if I want to express that the housenumber of my house is 14, than you describe this with an RDF triple:

<myHouse> <hasHouseNumber> "14" .

In a RDF triple, you have three parts:

• the subject (<myHouse>)
• the predicate (<hasHouseNumber>)
• the object ("14")

In this example, the subject and predicate are nodes while the object is a literal. It's also possible to have a node in the object, for example:

<myHouse> <hasNeighbouringHouse> <myNeighoursHouse> .

Nodes are the glue for Linked Data: other datasets can also reference <myHouse> when they want to state facts about my house (e.g. how good is the air quality on a working day). Also, it's important that datasets use the same reference for predicates, like <hasHouseNumber>, to allow automatic integration of different datasets.

There are two types of nodes: named nodes and blank nodes. Named nodes are resources that are referenced by an URI. This is an unique, global identifier on the Web and as best-practice it should be dereferenceable allowing others to look it up. Blank nodes are unnamed resources and are locally identified within the datadump or Web document that created this node.

There are multiple syntaxes available to describe RDF (JSON-LD, Turtle, N-Quads...). In these tutorials, we will focus on the usage of JSON-LD. JSON-LD uses the JSON syntax, but maps JSON terms to URIs through a (JSON-LD) context. @id is a keyword to reference a resource and thus creates a named node for the subject of triple(s). Note 1: we added "http://example.org/houses/id/" as base URI for houses and "http://example.org/" for the predicate. This should be replaced with a real location on the Web that can return more information about those things (cfr. dereferencing). Note 2: myNeighboursHouse also has a @id. This way, a JSON-LD parser will know that myNeighboursHouse is a resource and not a literal.

{
  "@id": "http://example.org/houses/id/myHouse",
  "http://example.org/hasHouseNumber": "14",
  "http://example.org/hasNeighbouringHouse": {
    "@id": "http://example.org/houses/id/myNeighboursHouse"
  }
}

You can try the above example in the JSON-LD playground. Click on "Table" to retrieve an overview of parsed triples.

To ease the readability of this example, we can add a JSON-LD context (@context) that allows defining mappings for JSON terms. We created a mapping houses so we can compact the URIs of myHouse and myNeighboursHouse with the prefix houses. We also created a mapping for the predicates "hasHouseNumber" and "hasNeighbouringHouse", which don't need to be used as a prefix. We used an expanded mapping to not only map to a URI with @id, but also indicate what datatype it is using XSD datatypes or if its a resource "@type": "@id".

{
  "@context": {
    "houses": "http://example.org/houses/id/",
    "hasHouseNumber": {
      "@id": "http://example.org/hasHouseNumber",
      "@type": "http://www.w3.org/2001/XMLSchema#string"
    },
    "hasNeighbouringHouse": {
      "@id": "http://example.org/hasNeighbouringHouse",
      "@type": "@id"
    }
  },
  "@id": "houses:myHouse",
  "hasHouseNumber": "14",
  "hasNeighbouringHouse": "houses:myNeighboursHouse"
}

You can try the above example in the JSON-LD playground. Click on "Table" to retrieve an overview of parsed triples.