Boundary Objects (BOs) in Systems Engineering

• Boundary objects (BOs) are artifacts that adapt to local needs while preserving a common identity, facilitating coordination across diverse organizational and disciplinary boundaries.
• They are employed in systems engineering as high-level requirements, architecture models, and verification documents to support cross-disciplinary collaboration.
• Empirical studies reveal that effective management of BOs improves knowledge transfer and mitigates coordination challenges in complex, multidisciplinary environments.

Boundary objects (BOs) are artifacts that simultaneously maintain a consistent identity across organizational and disciplinary boundaries while permitting interpretive flexibility and adaptation to local contexts. Originally formulated in the work of Star and Griesemer, boundary objects are central coordination devices in complex, multidisciplinary environments such as agile systems engineering, large-scale systems development, and collaborative cyber-physical systems. They serve as critical translation and negotiation mechanisms among heterogeneous stakeholder groups, facilitating shared understanding, knowledge management, and alignment without necessitating complete semantic consensus.

1. Formal Properties and Definitions

Boundary objects are defined as artifacts that are “plastic enough to adapt to local needs and the constraints of the several parties employing them, yet robust enough to maintain a common identity across sites.” Their principal properties—derived from empirical and theoretical studies—include interpretive flexibility, identity preservation, stability, modularity, and the capability to support abstraction or concreteness as required by context (Wohlrab et al., 2019, Wohlrab et al., 2020, Ramli et al., 2021).

Within the context of systems engineering, BOs commonly manifest as high-level requirements, architecture models, reference architectures, and variability management documents. These artifacts do not impose a singular perspective but instead support multiple concurrent interpretations, allowing diverse groups (“social worlds” or “methodological islands”) to coordinate on the basis of a shared but pluralistic reference (Wohlrab et al., 2019).

2. Typologies and Dimensions

Analysis in the engineering domain distinguishes between “vertical” and “horizontal” boundary objects:

• Vertical boundary objects: Span multiple phases of the engineering lifecycle (for example, linking functional requirements, architectural formulation, and verification artifacts). These objects maintain continuity and traceability across temporal and procedural boundaries.
• Horizontal boundary objects: Shared concurrently among teams or groups operating within the same phase, supporting coordination on interfaces or shared components (for example, signal specifications or design interface standards).

A further refinement is present in the Boundary Objects and Methodological Islands (BOMI) metamodel (Wohlrab et al., 2020), which classifies BOs by:

Attribute	Description
Purpose	Intended function within or across groups
Level of Detail	Degree of granularity (high/low)
Frequency of Change	Rate at which the object is updated
Modularity	Degree to which the object is decomposable
Maintainability	Ease of updating or evolving the object
Prescriptive	Whether it defines mandatory methods/structures

This structured typing, along with supertype/subtype modeling, enables systematic classification aligned with both theoretical and empirical requirements.

3. Identification and Analysis Approaches

Identification of boundary objects proceeds via qualitative and tool-supported schema:

• Residual Categories: Using actor–network theory, BOs are detected by identifying artifacts that do not fully align with a single community’s classification yet are referenced by multiple domains.
• Reference Analysis: Empirical evaluation of artifact usage (for instance, in tools like SystemWeaver) reveals which items are accessed or modified by multiple organizational units, suggesting their status as BOs (Wohlrab et al., 2019).
• Metamodel Instantiation: The BOMI metamodel formalizes the relationship of BOs to methodological islands (MIs) and roles, supporting instantiation as UML models or domain-specific diagrams. The method includes explicit relationships such as Usage (accessibility, stability, fit for purpose) and Governs (governance team, update frequencies).

Automated detection of dysfunctional patterns (“bad smells”) in boundary object usage is facilitated by constraints such as:

context BoundaryObj inv DetailedHighChange: self.LevelofDetail = High and self.FrequencyofChange = High

This formalization allows direct integration with tool-based analysis (Wohlrab et al., 2020).

4. Management Practices and Governance

Empirically validated management guidelines delineate BO handling from locally relevant artifacts (Wohlrab et al., 2019):

• Joint Analysis and Evaluation: Multidisciplinary stakeholders periodically assess which artifacts function as boundary objects, employing cross-reference analysis in systems engineering tools.
• Coordinated Governance: Dedicated groups (e.g., artifact committees, Communities of Practice) are constituted to oversee the lifecycle of BOs, manage formal versions, address cross-boundary changes, and maintain traceability. Visualization of artifact relations is emphasized to enhance collective oversight.
• Flexible Upfront Definition: BOs are initially defined with lightweight, high-level descriptions suitable for early impact analysis. Continuous refinement adapts these artifacts to evolving project and organizational needs.
• Governance Model Integration: Governance is institutionalized through explicit assignment of roles, maintenance of update logs, and coordinated release management.

The application of these practices addresses challenges such as artifact degradation, coordination bottlenecks, and the reconciliation of agile with plan-driven methods.

5. Empirical Evidence and Practical Contexts

Empirical research across multiple large-scale companies confirms the critical role and effective application of boundary objects in industrial systems engineering, automotive development, and collaborative CPS (Wohlrab et al., 2019, Wohlrab et al., 2020, Ramli et al., 2021):

• Artifact Classification: Practitioners consistently distinguish between BOs (e.g., architecture models, high-level requirements) and team-internal artifacts (detailed documentation, low-level requirements, test cases).
• Coordination Effectiveness: BOs underpin alignment across organizational silos and support knowledge transfer, enabling robust discussions between product management, development, and testing—particularly in geographically distributed or dynamic team structures.
• Detection of Issues: Modeling BOs and their interactions with MIs enables organizations to identify “bad smells” such as high criticality/low stability artifacts, missing governance assignments, and disproportionate rates of change in legacy objects (Wohlrab et al., 2020).

Reference architectures in CPS are a specific example, functioning as BOs by capturing architectural knowledge, supporting knowledge transfer, and serving as translation layers between stakeholder perspectives. Challenges with accessibility, ongoing relevance, and inclusion of tacit knowledge are observed and form the basis for ongoing work (Ramli et al., 2021).

6. Challenges, Tool Support, and Formalization

Key challenges in BO management span:

• Interpretive Accessibility: Ensuring BOs (such as reference architectures) remain comprehensible and useful for all stakeholders, regardless of experience or background (Ramli et al., 2021).
• Tacit Knowledge Capture: Making implicit expertise (from senior practitioners) explicit within the BOs, enabling sustainable knowledge sharing and onboarding.
• Alignment Across Heterogeneity: Facilitating consensus and coordination amid methodological diversity, particularly where cross-domain or multi-discipline teams interact.

Tool-based support relies on formal modeling (e.g., UML, OCL constraints) and visualization. Tables map artifact classes to management dimensions (“current practices,” “relevance,” “affected areas”), while graphical representations (such as TikZ diagrams and UML associations) express the structure and lifecycle of BOs and their governance.

7. Prospects and Open Questions

Current and ongoing research seeks to refine the formal representation of BOs—especially in the area of trustworthiness attributes in CPS, where questions remain regarding the inclusion of ethical dimensions (transparency, oversight, explainability) (Ramli et al., 2021). Strategies for capturing and evolving tacit knowledge, and the incorporation of domain-specific visual languages for BO management, are identified as future priorities.

A plausible implication is that tool-supported, model-based approaches (e.g., the BOMI metamodel) will become an increasingly integral part of organizational process engineering to make coordination bottlenecks and governance challenges explicit and addressable. Further empirical validation may also clarify the boundary conditions under which BOs most effectively function, and which combination of attributes, modularity, governance, and refinement cycles deliver optimal coordination outcomes.

---

Boundary objects thus constitute a formalized, empirically grounded class of artifacts central to the coordination of knowledge, roles, and workflows in complex sociotechnical systems. Their rigorous identification, structured management, and continual adaptation are foundational for sustained cross-boundary collaboration and organizational agility.