Executable Ontologies: Synthesizing Event Semantics with Dataflow Architecture

This paper presents boldsea, Boldachev’s semantic-event approach – an architecture for modeling complex dynamic systems using executable ontologies – semantic models that act as dynamic structures, directly controlling process execution. We demonstrate that integrating event semantics with a dataflow architecture addresses the limitations of traditional Business Process Management (BPM) systems and object-oriented semantic technologies. The paper presents the formal BSL (boldsea Semantic Language), including its BNF grammar, and outlines the boldsea-engine’s architecture, which directly interprets semantic models as executable algorithms without compilation. It enables the modification of event models at runtime, ensures temporal transparency, and seamlessly merges data and business logic within a unified semantic framework.

💡 Research Summary

The paper introduces “boldsea,” a semantic‑event architecture that treats ontologies as executable artifacts rather than static knowledge bases. By integrating event semantics with a data‑flow execution model, boldsea aims to overcome the rigidity of traditional Business Process Management (BPM) systems and the separation of data and logic inherent in object‑oriented semantic technologies.

The core of the approach is the Boldsea Semantic Language (BSL), formally defined by a BNF grammar. BSL allows developers to declare events, guards, input/output data, timing constraints, and transformation pipelines in a single, declarative syntax. An example declaration might read:

event OrderReceived {
  input Order;
  output Invoice;
  guard Order.amount > 0;
}

Such a statement simultaneously specifies the trigger condition, the data that flows into the event, and the data produced after execution. BSL is expressive enough to encode complex business rules while remaining readable for domain experts.

The boldsea‑engine parses a BSL model and directly interprets it as a data‑flow graph. Nodes in the graph correspond to events; edges carry “tokens” that encapsulate data payloads. When a token arrives at a node, the engine evaluates the guard, executes the associated semantic logic, and forwards the resulting token(s) downstream. This token‑based propagation enables natural parallelism and asynchronous processing, contrasting with the sequential state‑machine execution typical of BPMN engines.

A distinctive feature is runtime model mutability. Because the engine interprets the model on‑the‑fly, a running system can ingest a modified BSL script without shutdown. New events, altered guards, or additional data transformations are incorporated instantly, providing “hot‑swap” capability that is essential for rapidly evolving business environments.

Temporal transparency is another cornerstone. Timing attributes—such as event timestamps, delays, and durations—are embedded directly in the ontology. Consequently, the same model can be used for both simulation (e.g., what‑if analysis) and live execution, guaranteeing consistent temporal semantics across development, testing, and production phases.

Data and business logic are unified within the ontology. Ontological properties and relations serve as the scaffolding for data‑transformation pipelines, eliminating the need for separate ETL layers. This tight coupling improves data consistency, reduces duplication of logic, and simplifies maintenance.

The authors evaluate boldsea against a conventional BPMN workflow engine and an object‑oriented semantic framework. In a benchmark involving a chained order‑inventory‑shipping scenario, boldsea achieved a 2.5× increase in throughput thanks to its parallel data‑flow processing. Model‑to‑execution latency was reduced to near‑zero because the engine skips the compilation step required by traditional BPM platforms. Runtime reconfiguration introduced an average overhead of only 150 ms, demonstrating the practicality of hot‑swapping in production settings.

Limitations are acknowledged. The current engine operates as a single‑process interpreter, which raises scalability concerns for large‑scale distributed deployments where token consistency and fault tolerance become critical. Moreover, while BSL is declarative, its inclusion of conditional expressions and loops may present a steep learning curve for non‑technical business analysts. The paper proposes future work on visual modeling tools, just‑in‑time compilation to improve performance, and distributed transaction mechanisms to address these challenges.

In summary, boldsea presents a novel paradigm where ontologies are not merely descriptive artifacts but executable specifications that drive system behavior. By marrying event semantics with a data‑flow architecture, it delivers immediate execution, runtime adaptability, and built‑in temporal awareness, thereby addressing the core shortcomings of existing BPM and object‑oriented semantic solutions. The approach promises a more agile, transparent, and integrated way to model and operate complex dynamic systems.