Executable Ontologies: Event Semantics and Dataflow Architecture for Dynamic Systems

Traditional modeling and process execution systems often separate semantic descriptions from control logic, leading to brittle architectures that are hard to adapt, extend, or reason about. In this paper, Alexander Boldachev introduces an alternative approach where semantic models become inherently executable, driving both data and behavior within a unified framework. 

The core contribution of the work is the boldsea architecture — a system that synthesizes event semantics with dataflow principles so that models themselves are dynamic structures directly controlling execution. Instead of treating ontologies as passive metadata or static schemas, this approach uses them as first-class executable artifacts that blend data, logic, and process execution seamlessly. 

Key aspects highlighted in the paper include:

• How executable ontologies act as dynamic controllers rather than static descriptions. 

• The definition and formal grammar of BSL (Boldsea Semantic Language) used to encode these models. 

• Architecture of the boldsea engine which interprets semantic models at runtime without a separate compilation step. 

• Runtime modification of event models with temporal transparency, enabling flexible adaptation of system behavior. 

• A unified semantic framework merging data and business logic, simplifying integration and traceability. 

This paradigm is relevant for complex enterprise systems, semantic knowledge graphs, dynamic workflows, and environments where adaptability and semantic coherence are critical.

Read the full academic paper on arXiv: https://arxiv.org/abs/2509.09775