2.1 Dimensions of Engineering Maturity

To make engineering maturity actionable, we introduce a set of Fab-ilities as a starting vocabulary for describing maturity in fabrication systems. Specifically, we sketch six dimensions (Table 1) that capture what a contribution makes buildable, runnable, and transferable beyond the original lab setting. This framing is inspired by a prior workshop focusing on an Interaction Engineering Glossary and a Repository for interactive software "Fab-ilities" [8], which aimed to elicit reference properties for assessing the quality of engineered HCI products. Our Fab-ilities are not a checklist or a compliance target. They are descriptors for articulating what kind of maturity a contribution achieves, what remains unresolved, and what that maturity enables for others. We do not claim these dimensions are exhaustive or final. We offer them to make maturity discussable and reviewable, and to prompt a broader conversation about what fabrication research in HCI should value beyond novelty.

Buildability asks whether others can physically reconstruct the artifact. Materials vary between suppliers and regions; components get discontinued within months of publication. The open-source hardware community has developed practices to address this [9]. Buildability requires identifying which specifications are critical, which components can be substituted, and which assembly sequences contain implicit constraints that only surface when someone else attempts reconstruction.

Executability asks: can someone who did not build the system actually operate it? A system may be reconstructable yet still require its original creators, because calibration is undocumented, error recovery depends on tacit knowledge, or the workflow assumes internalized expertise. This spans software (environment setup, dependency conflicts, implicit build assumptions) and physical operation (material placement, focus adjustment, speed/power tuning). The two interact: fabrication systems involve software controlling hardware, and failures can occur anywhere in that chain.

Reliability concerns behavior under real-world variability, for example: humidity affects adhesion, temperature influences curing, tool wear accumulates. Characterizing what fails first under what conditions requires sustained observation exceeding typical paper timelines. Recent work on FEDT [10] highlights that such characterization experiments are performed inconsistently, tools that support systematic experimental design offer one path forward.

Maintainability asks whether the system can be maintained without its original authors. Software engineering treats maintainability as first-class [11]; fabrication research does not. Yet physical systems face additional challenges: components wear out, materials degrade, and suppliers discontinue parts. A system that cannot be maintained becomes frozen at publication and is unable to evolve or adapt when dependencies change.

Transferability asks whether the system can travel to new contexts—different machines, materials, or users—and what must remain fixed. This is distinct from robustness: a system may be robust locally yet fail completely elsewhere. Replication studies show how contextual factors invisible to authors determine success in new settings [12]. Answering transferability questions requires systematic experimentation with variations.

Scalability confronts economic realities that prototypes leave unexamined. What does it cost to build ten versus a hundred? Research prototypes use components chosen for convenience—hand-soldered connections, expensive development boards, specialty materials—obscuring whether the contribution could ever leave the lab. Understanding which cost drivers dominate at volume, which steps need automation, and which choices trade convenience against viability is itself valuable knowledge.

3.1 Cases of valorisation attempts

JigFab [13] (Figure 1) was developed in close collaboration with Fablab Genk, an open access makerspace run by one of the authors. We worked with woodworkers, hobbyists, and engineering students who use the lab daily. The system helps users design and fabricate custom woodworking jigs using digital fabrication tools. When we pursued commercialization, we filed a patent application and engaged an external marketing agency to identify potential industrial markets, but we found little traction. Prospective commercial users wanted a more mature system. They asked for clearer failure modes, stronger documentation, and behavior that remained reliable across different machines, materials, and workflows. In addition, the project involved multiple universities, which made intellectual property valorization more complex. Shared ownership introduced additional coordination, legal constraints, and procedural barriers that effectively raised the threshold for moving from a research prototype to a commercial product.

StoryStick++[14] (Figure 2) is a phone clip on that guides measuring and marking with minimal reliance on numbers, units, or mental arithmetic. This reduces common errors. StoryStick++ is currently in market research. We repeatedly see the same pattern. Companies understand what the system does, but they struggle to imagine a finished product based on the bare bones prototype we can provide. The gap is therefore not conceptual. The gap is practical and economic. The knowledge needed to build, run, and adapt the system to their context is not written down and remains in the heads of the developers. As a result, basic questions about cost structure, manufacturing complexity, and market viability remain unanswered.

Silicone Devices[15] (Figure 3) presents a DIY workflow that uses a CO₂ laser cutter to fabricate stretchable electronic devices in cast silicone with embedded liquid metal traces, computation, and sensors. To support replication, we published a step by step tutorial as an Instructable [16]. In practice, the workflow requires careful execution, so workshops at our university and at the University of Chicago were facilitated by one of the authors. The project also received national valorization funding to attempt to set up a spin-off company on stretcheable electronics. We received funding for a two year trajectory to translate the technique into an industrial process. During that trajectory, the focus shifted from the original HCI goal of democratizing fabrication toward industrial applications, where it faced strong competition from established manufacturing techniques. The track ended when the lead researcher moved into a different role. The work, however, continues to inform research in both materials science and HCI.

PaperPulse [17, 18] (Figure 4) enabled non-experts to produce interactive paper devices using printed conductive traces. We explored commercialization together with a professional partner and investigated patenting options, including prior-art checks with an external intellectual property firm. However, the upfront costs and the uncertainty about a viable market path made it difficult to justify continuing the trajectory. The core novelty of PaperPulse aligns naturally with consumer oriented (B2C) markets, which are particularly difficult to sustain in Europe. This experience exposed a broader lesson. Many HCI system contributions are designed around specific user groups and contexts of use. When they are repositioned for different domains, such as expert, education or industrial users, the original contribution often loses relevance. This contrasts with general purpose techniques in materials science, in this case printed electronics, that tend to transfer more easily across application areas.

LamiFold [19] (Figure 5) exposed a different facet of the maturity gap, namely reproducibility. LamiFold’s lamination workflow selectively cuts and glues stacks of sheet material to embed functional mechanisms. In practice, the process is highly sensitive to machine and material parameters. Laser power, cutting speed, kerf width, and material thickness interact in ways that are hard to describe in general terms. Calibration settings that yield reliable results on one laser cutter often fail on another. As a result, the knowledge required to make LamiFold work reliably across different setups remained largely undocumented and stayed tied to hands on experience rather than transferable instructions.

3.2 Barriers to Uptake Beyond Publication

One pathway has worked consistently. Our projects have repeatedly influenced the classes we teach on designing and building interactive systems, as well as fabrication and prototyping, so the research spreads through education. The systems are complex, broad in scope, and deep in technical detail, which makes them difficult to master and therefore well suited as training material for students. However, that same complexity is also a barrier to use outside the lab. What is acceptable as open ended exploration in research and education becomes hard to operationalize in practice, where users expect stable behavior, clear procedures, and predictable costs.

When we pursued other pathways such as commercialization, licensing, spin offs, and other types of market adoption, we encountered various barriers that prevented uptake. The reasons differed across projects, but the pattern was similar. Reliability was not characterized, operational know-how remained undocumented, cost-benefit models were missing, designs were tightly bound to a specific (lab) context and did not transfer well to real life usage, and the documentation needed for others to continue the work was often absent.