In complex systems engineering—ranging from aerospace to legal compliance—a fundamental "Great Divide" often exists between high-level requirements and technical implementation. Traditionally, requirements live in static documents (Word files or spreadsheets), while implementation lives in code or architectural models. This separation makes it nearly impossible to verify that every design decision is justified or that every requirement has been met.
To bridge this gap, researchers have developed mechanisms for formal bidirectional traceability. By transforming requirements into structured, machine-analyzable formats and linking them directly to technical artifacts, teams can ensure consistency, automate audits, and manage changes across the entire development lifecycle.
From Documents to Ontologies
The shift toward traceability begins by moving away from "document-centric" approaches. As Adedjouma et al. observe in the context of automotive systems, traditional methods struggle with the complexity of modern units like Selective Catalytic Reduction systems. Instead, a model-based approach allows for automated traceability between abstraction levels, linking system functions directly to software components.
The intuition behind this works by treating requirements not as prose, but as a structured ontology. Litao Li argues that in high-stakes fields like aerospace, requirements should be modeled with explicit logical links—such as "allocation" and "satisfaction" relationships. This creates a mathematical matrix where every design element is justified by a requirement, and every requirement is addressed by a design element.
Sun et al. take this further by mapping visual SysML models into a Web Ontology Language (OWL 2) framework. By defining logical axioms for how individual hardware or software blocks modify parameters, the system can use automated reasoning to verify that safety-critical constraints are satisfied without manual intervention.
Bridging the Semantic Gap
Traceability is not just a technical challenge; it is a communication challenge between multidisciplinary teams. Haesen et al. propose a framework to bridge the divide between User-Centred Design (UCD) and Software Engineering (SE) by linking user personas to task models and interactive prototypes. Similarly, Luyten et al. suggest using metadata-enriched storyboards to link narrative scenarios to technical device specifications, providing a "common language" that reduces ambiguity.
This logic extends to the legal domain, where the "requirements" are statutory texts like the GDPR. Torre et al. and Soltana et al. demonstrate that by translating legal rules into UML class diagrams and Object Constraint Language (OCL), organizations can automate compliance checking. The key to their approach is a traceability matrix that maps every technical constraint back to a specific legal article. This ensures that a non-technical legal expert can verify the model’s correctness, maintaining what the authors call "legal defensibility."
Maintaining Consistency Through Change
The true power of formal traceability emerges during change management. In a traditional system, changing a single requirement might necessitate a manual search through thousands of pages of documentation to find affected components.
Modern frameworks solve this through automated synchronization:
- Gateway Interfaces: He and Zhang describe using digital gateways to link textual requirements to architectural elements in aircraft electrical systems, facilitating instant impact analysis when requirements shift.
- Functorial Relationships: Bakirtzis et al. apply category theory to cyber-physical systems, using functors to map behavioral models to architectural implementations. This formalizes how a change in behavior at one level mathematically propagates to another.
- Live Crosswalks: For rapidly evolving fields like cybersecurity, Tupsamudre et al. propose AI-driven "live crosswalks" that detect updates in regulatory standards and automatically map them to technical implementation scripts (Infrastructure-as-Code).
By establishing these explicit, bidirectional links, systems move from being a collection of disconnected artifacts to a unified, synchronized graph. As Ciaghi et al. note, this level of integration allows the system to automatically identify necessary amendments in laws when a process model changes, and vice-versa, ensuring that the implementation remains legally and technically sound throughout its lifecycle.
Key Foundational Work
Incorporating work, process and task analysis into commercial and industrial object-oriented systems development (1998) — contributed 1 concept elements to this direction.
MuiCSer: A Process Framework for Multi-disciplinary User-Centred Software Engineering Processes (2008) — contributed 1 concept elements to this direction.