Applying Standards to Information Centric Operations
2 ONTOLOGY
The original definition of Ontology is the study of everything ; of what is . This has been around since the Ancient Greeks , but now we use the word in a subtly new way , to talk about the ontologies ( plural ) that our data represent . Each database , each messaging schema , even the variables in computer programs have some intended relationship with reality . They represent something . Whether they represent things completely or consistently , whether the data means the same thing going in as someone else thinks it means coming out the other end , is another matter . Everything in your computational environment has an ontology , whether it knows it or not ; the smart next move is to formalize those ontologies so that what people think a given piece of data means is not left to chance or to local interpretations but is set out logically .
A formal artifact that can be used to represent the meanings of real things , or the meanings that data is referring to , is also called an “ ontology .” Ontologies in this sense are models that use formal logic to make consistent , computationally verifiable assertions about things in the world .
An ontology uses logic to make assertions about the world in terms that can be understood in the same unambiguous way by both machines and people . But logic is only part of the story . Logic on its own doesn ' t make a model meaningful since you can have logically complete and consistent assertions that are utter nonsense , with no bearing in reality . For this you need to define some formal relationship between the model and what it is a model of .
The simplest kind of ontology is represented using a family of logic called Description Logic ( DL ) [ 4 ], a kind of knowledge representation language . This is a simple logic of what there is and what distinguishes things from one another . More complex features of reality would require higher orders of logic to represent them , but we start with DL , which is reflected in the syntax of the Web Ontology Language ( OWL ) [ 5 ].
We need to assert meanings against data to enable a data-centric environment . However , if you start from the bottom up , asserting meanings one by one against pieces of data , there is no guarantee of completeness or consistency and , more importantly , no actual grounding in reality . A more mature engineering approach is to work from the top down : to determine the meanings of things that are of interest to the organization , and only then link this to data , creating the necessary sea of linked data as an enterprise knowledge graph .
Some examples and case studies will help to clarify this relationship between ontologies that reflect reality , and the use of semantic technology such as OWL ontologies . These distinctions will become particularly important in considering digital twins .
3 ONTOLOGY EXAMPLES AND CASE STUDIES
The implications of data centric operations are more far reaching than most of us yet realize . A true inversion of data and operations means a true inversion of how people interact with data
Journal of Innovation 15