Parmenides of Elea was a Greek philosopher who lived 2,500 years ago. His primary surviving work is a poem often referred to as “On Nature,” the objective of which is to separate truth from opinion. “What is” and “What is not.” Parmenides is considered one of the founders of the study of metaphysics and ontology: Ontos (ontoß) meaning “truth,” and logos (λόγος) related to symbolic conceptualizations, e.g. “words”. Subject to interpretation, of course.
What does this have to do with data and system integration? Well, the study of nature and truth has continued through the ages and is crucial when it comes to data, since the meaning of data is of paramount importance and is the central goal of an ontology. Ontologies define a language by which definitions of and relations among concepts are formalized, and by doing so allows data to be understood, providing the semantics, or meaning, behind the data.
There are several technologies that can be used to specify ontologies: RDF, OWL, SPARQL, and many more. Why use poetry? Being married to a poet, @mishiepoet, we discuss and experiment with this form of communication often, so writing more poetically is a natural outcome of my life experience. Poetry promotes a careful brevity, a focus on clarity, and a reverence for a subject that works well for complex topics. Poems do not have to rhyme, but devices such as meter, repetition, and form add interest, beauty, and fun to a subject for which you might not expect it.
When I became involved in a project to define an ontology using IEC CIM as the basis for an analytics platform, understanding and explaining these concepts became necessary, and the assertions started taking shape.
Without further ado, here is the first revision of “Assertions: On the Nature of Ontology.”
Assertions: On the Nature of Ontology
First principles are fundamental axioms
that provide the basis for reasoning
Assertions are clear, attributed statements of belief,
giving rise to confidence in truth
Love, concern for our world, is our highest authority,
and is a precondition for all activity
Models are simplified representations of reality
comprised of conceptual symbols
Objects express collections of organized,
structured information about the world
Concepts define, describe, and relate entities and events,
across time, space, and context
Ontologies are models of concepts in a domain
that describe objects and their relationships
Ontologies support logical deductive and inductive reasoning
of conclusions using inference
Ontologies include data and metadata:
types, properties, relations, and instances
Requirements are expressions of fundamental,
existing, and aspirational objectives
Requirements are prioritized by value, effort,
urgency, and dependency
Designs distill and translate requirements
into modular, implementable specifications
Schemas are ontologies that define data elements
as symbolic tokens within a namespace context
Definitions specify the type, context, relations,
domain, and range of model elements
Rules articulate assertions about the content, conditions,
and expectations amongst concepts
Taxonomies define abstract-to-specific type categorizations
within a domain
Relationships convey taxonomy,
aggregation, composition, containment, and causality
Category dimensions provide the basis
for summarization of event metrics
Definition of ontology provides semantic clarity
for information models, enabling computation
Rules specified by assertion are transparent, certified,
and governed by human stakeholders
Governing bodies form around topic collections
with strong correlation of interrelationships
Modular design encapsulates functionality
within versioned service interface boundaries
Service interface compatibility manages dependencies
to enable independent enhancements
Implementations codify the translation
of requirements and design into execution
Ontologies expose our understanding
of the metaphysical nature of the universe
Assertions are ratified through publicly anonymized,
fully auditable consensus
Through computation and disambiguation
of ontologies, understanding can emerge
Poems use natural language. Natural language is necessary for facilitating discussions and specifications at the level of requirements and business rules. These specifications can then be expanded and translated into implementations. Large neural network natural language processing (NLP) technology is advancing rapidly, and someday, compilation of ontology assertions in carefully crafted English into a model that can support reasoning might be a task that a computer could perform. It could already be possible.
Are these statements cohesive, minimal, and sufficient?
No – far from it. But they are a foundation for a beginning.
Please join the conversation! Which assertions do you agree or disagree with?
What would you add? Let us know your thoughts in the comments!
Xtensible Solutions supports energy utilities leveraging industry standards including the IEC CIM to build information models and platforms to solve today’s data integration challenges. Ontologies represent the next frontier in integration, and Xtensible is leading the way, applying machine learning, graph technology, agile methodologies, data governance, and distributed platforms to reduce integration friction and build the semantics management platform of the future.
Interested in learning more about the use of ontologies for data integration? Speak to a member of the Xtensible team.