Independent technical research

Own the meaning behind your systems.

Ontology-Native Software Engineering explores the hypothesis that ontologies could serve as an authoritative semantic foundation for software systems, guiding their design, implementation, and evolution.

The research investigates how explicitly modeled meaning might remain consistent as technologies, implementations, and requirements change.

Research HypothesisThe ontology is the software.

Code is its compiled form.

01 / Research Motivation

The translation tax

Why do complex systems repeatedly reconstruct the same meaning?

In complex organizations, business meaning is often distributed across software, policies, data structures, documentation, and human expertise.

When systems evolve, the intent behind their behavior may need to be reconstructed from these fragmented representations.

This recurring translation of meaning motivates the research behind Ontology-Native Software Engineering.

01

Change starts with archaeology

Understanding a change may require reconstructing intent from code, policies, data, documentation, and the knowledge of experienced people.

02

Context is translated repeatedly

Different disciplines may interpret the same business requirement through separate technical, operational, and organizational perspectives.

03

Assurance arrives too late

Relationships between approved intent, implementation, dependencies, and verification evidence may only become visible after changes have been implemented.

Research ObservationSoftware systems encode business meaning, but that meaning is not always explicitly represented, independently accessible, or preserved across implementations.
02 / The Research Hypothesis

From implementation to intent

What if meaning became the authoritative foundation of software?

Ontology-Native Software Engineering investigates whether essential system meaning can be represented explicitly, maintained independently of implementation, and used to guide software evolution.

The central hypothesis is that an ontology could become the authoritative representation of intended system behavior, while implementation artifacts are generated, adapted, or validated against it.

This is an evolving methodological proposition, subject to experimentation and technical validation.

01

Own the meaning

Explore how critical system intent can be represented explicitly rather than remaining implicit within particular implementations.

02

Make implementation replaceable

Investigate whether business meaning and behavioral constraints can remain stable while underlying technologies change.

03

Keep change accountable

Examine how approved intent, implementation changes, and verification evidence might remain connected throughout software evolution.

A bounded, well-defined system change provides a useful starting point for investigating this hypothesis.

03 / Research Scenarios

Three research questions

Where could an ontology-native approach make a difference?

The hypothesis can be explored through scenarios involving system modernization, regulated change, and AI agent behavior.

These scenarios help identify open technical questions and potential experimental directions.

System modernization

What must remain true when the technology changes?

Can explicit semantic models help preserve essential business behavior when an implementation is replaced or modernized?

Regulated change

Can systems, controls, and obligations be traced before implementation?

Could ontology-based representations help identify dependencies and consistency requirements associated with a proposed change?

AI agent authority

What defines the boundaries of an agent's permitted actions?

Can independently defined semantic rules and constraints help preserve an agent's intended authority when prompts, models, or tool configurations change?

90-second research self-assessment

How well is business meaning represented across systems?

Consider the following questions about semantic consistency, traceability, and system evolution. Mark those that would be difficult to answer in your current environment.

  1. 01

    Can you point to where a critical business rule is authoritatively defined?

  2. 02

    Can you see its affected systems, teams, and controls before changing it?

  3. 03

    Could you move a capability without rediscovering its business logic?

  4. 04

    Can you explain why production behavior conforms to approved intent?

  5. 05

    Could a prompt or tool configuration quietly change an agent's authority?

04 / Exploring the Hypothesis

From hypothesis to experiment

How could the hypothesis be investigated?

The ideas behind Ontology-Native Software Engineering can be explored through conceptual models, controlled experiments, and illustrative software scenarios.

A possible experimental approach begins with a bounded system capability and investigates whether its essential business meaning can be represented explicitly, connected to implementation artifacts, and preserved through a controlled change.

Such experiments may examine:

  • How business concepts, rules, and constraints are represented.
  • How semantic models relate to code, APIs, data structures, and tests.
  • How proposed changes affect dependent system components.
  • How implementation behavior can be compared against explicitly defined intent.
  • Which limitations, ambiguities, and verification challenges remain unresolved.

The purpose is to investigate the technical feasibility, limitations, and implications of ontology-native software engineering.

The hypothesis remains open to technical experimentation, critical discussion, and further validation.

05 / EXPERIMENTAL RESEARCH PROTOTYPE

From semantic extraction to ontology refinement.

To investigate the Ontology as Software hypothesis, we have developed a working experimental prototype that explores how domain knowledge can be extracted, represented, and refined from existing software artifacts.

The prototype combines automated concept identification, human-guided semantic clarification, and ontology review within a unified research environment.

01 — Ontology Extraction & Semantic Clarification

The first stage focuses on identifying domain concepts and relationships from source code and documentation, while detecting ambiguities that require human interpretation.

The prototype enables researchers to examine conflicting meanings, review supporting sources, and provide clarification that guides the resulting ontology.

In the example shown, the distinction between Merchant and MerchantAccount is resolved through human input, allowing the concepts and their relationship to be explicitly defined.

Experimental ontology clarification dashboard reviewing Merchant and MerchantAccount as separate classes connected by a hasAccount relationship, with human feedback and source references.
Ontology extraction and semantic clarification interface from the working research prototype.

02 — Ontology Review & Refinement

The second stage focuses on examining and refining the resulting semantic model through an interactive ontology graph.

The prototype enables researchers to explore concepts, properties, and relationships, review suggested changes, investigate uncertain mappings, and trace semantic elements to their supporting sources.

This review process helps identify inconsistencies and refine the ontology before proceeding with further experiments in ontology-native software engineering.

Ontology review dashboard showing a graph of payment concepts and relationships, accepted concepts, suggested changes, low-confidence mappings, and source-backed evidence.
Ontology graph, semantic review, and source traceability interfaces from the working research prototype.

RESEARCH STATUS

Working experimental prototype · Ongoing, non-commercial technical research.

The prototype is being developed to investigate the feasibility and limitations of ontology-native software engineering. It is not a production deployment or a commercial product offering.

06 / Continue the inquiry

For researchers and practitioners

Explore the research hypothesis.