[Paper] Towards Formalising Stakeholder Context using SysML v2

Source: arXiv - 2604.19390v1

Overview

The paper proposes a structured framework that translates the often‑vague “stakeholder context” into a rigorously defined system architecture using SysML v2 and Soft Systems Methodology (SSM). By leveraging the formal semantics of the Kernel Modelling Language (KerML) and ISO 42010 alignment, the authors demonstrate a traceable path from stakeholder concerns straight to architectural elements—promising fewer misinterpretations in complex projects.

Key Contributions

• Hybrid methodology that couples SSM’s qualitative stakeholder analysis with SysML v2’s formal modeling capabilities.
• Reference architecture that maps SSM artefacts (e.g., rich pictures, CATWOE analyses) to SysML v2 concepts such as Stakeholder, Concern, and Requirement.
• Demonstration case study showing end‑to‑end traceability from stakeholder statements to concrete system components.
• Discussion of trade‑offs, notably the learning curve introduced by SysML v2’s textual notation versus the gain in semantic precision.

Methodology

1. Elicit stakeholder context using SSM techniques (rich pictures, root definitions, CATWOE). This step captures the “soft” aspects—values, motivations, and constraints—that are usually expressed in natural language.
2. Formalise the context by translating SSM artefacts into SysML v2 elements. The authors employ KerML’s precise syntax to encode stakeholder roles, concerns, and the relationships among them.
3. Construct a reference architecture that aligns the formalised context with ISO 42010’s architecture description standards. This architecture serves as a scaffold where each stakeholder concern is linked to a specific system view or component.
4. Validate through a case study (a mid‑size logistics information system). The study walks through each transformation step, producing a traceability matrix that can be inspected by both domain experts and engineers.

The approach is deliberately tool‑agnostic; the authors use open‑source SysML v2 parsers to illustrate how the textual models can be processed programmatically (e.g., for automated consistency checks).

Results & Findings

Practical Implications

• Improved requirements engineering: Developers can generate machine‑readable specifications directly from stakeholder workshops, feeding them into CI pipelines for early validation.
• Automated impact analysis: When a stakeholder’s priority shifts, the traceability matrix instantly highlights which system components need re‑evaluation, saving time on manual reviews.
• Compliance & auditability: Aligning with ISO 42010 makes it easier to produce documentation that satisfies regulatory standards (e.g., aerospace, medical devices).
• Tool integration: The textual nature of SysML v2 means models can be version‑controlled (Git) and processed with standard parsers, enabling “infrastructure‑as‑code”‑style governance of system architecture.

Limitations & Future Work

• Empirical breadth: The framework has only been validated on a single case study; broader testing across domains (e.g., finance, IoT) is needed to confirm generality.
• Learning curve: The textual notation of SysML v2 may deter teams accustomed to purely graphical tools; the authors suggest developing higher‑level DSL wrappers or IDE plugins to lower this barrier.
• Tool ecosystem maturity: Current SysML v2 tooling is still evolving, which could limit immediate adoption in fast‑moving development environments.

Overall, the paper offers a promising bridge between the human‑centric world of stakeholder analysis and the rigor required for modern system architecture—an approach that could streamline development pipelines and reduce costly misalignments.

Authors

• Matthew Harrison
• John Carlin
• Chengyuan Liu
• Sarah Dunnett
• Siyuan Ji

Paper Information

• arXiv ID: 2604.19390v1
• Categories: cs.SE
• Published: April 21, 2026
• PDF: Download PDF