In a connected world with systemic relations and complexity, design needs new methods and approaches to create and design for systems that grow dynamically, relate to human users or inhabitants temporarily or over time, and expand their capabilities with emerging technologies. The openness of such systems challenges designers and conventional design approaches. The IoT Sandbox is a novel design environment that allows individual designers and design teams to experience and create given synthetic, adaptable constraints of a scripted system-through cooperation and competition. The IoT Sandbox exists in conceptual, physical and networked space, clearly linked to an educational methodology. We showcase this system in a group design project that targets a smart home design space, and we conclude the paper with a discussion of our ifndings and position for future work.
• Human-centered computing → Interaction design.
Systems and systems thinking are widely recognized as the next
challenge for (interaction and product) design (e.g., [
Despite the natural afinity of design to complexity and the
openness of systems, we observe a gap in design research and practice
when it comes to designing for systems beyond the conceptual.
The openness of systems challenges the design as their
conventional approaches often do not work. Here we focus on two aspects
of systemic openness, (1) distributed scope and interaction and
(2) functional composition. Looking at the functional scope,
standalone interactive products have clearly demarcated physical and
interaction boundaries. It is clear where they start and end, which
opportunities for accessing functionality through interaction they
ofer and how many users are involved to access this
functionality. Systems of products do not have these clear boundaries; they
may flexibly extend to the currently active products in the system.
Second, a system’s composite functionality changes with its
structure. The functionality depends not only on the functionality of
the individual products in the system, but also on functionality
that emerges as part of the system, that is, functionality that
different individual products can only ofer together or in interplay.
This form of emergence is partly undetermined and open at the
time of design, challenging the designer to deal with a multitude
of dependencies in the design process that can only be resolved
through a dual perspective on product and system level–this is
what we refer to as systemic openness. Systemic openness is a
particularly wicked phenomenon [
The IoT Sandbox builds on more than a decade of experience with teaching design students to design connected and smart products in a growing systems context. We first summarize our experiences briefly after which we present the IoT Sandbox together with methodological considerations. 2.1
Over the past decade we have introduced the systems perspective to our Industrial Design students. This has been an insightful journey where we have learned about the dificulties in adopting a systems perspective in design. One of the main hiccups proved to be systemic openness, students have consistently struggled with the lack of boundaries (or constraints) that was caused by the concept of systemic openness. A persistent tendency in our students was the desire to ‘think the solution through’ rather than to start making solutions: the mind was trusted over the hands. Even more than for traditional design endeavors, this approach is bound to fail for systems and even more so as the system is not static but evolves dynamically through the designing and acting of others. Consequently, this created long unproductive struggles that took up significant time in the semester, designerly exploration sufered. Another prominent observation was that students created ‘walled gardens’ that reduced their design solutions to essentially interaction design solutions rather than ‘systems’ design solutions–data, connectivity, emergence or openness were markedly scarce in the solutions that the students delivered. After all, designing systems is not compatible with a divide-and-conquer approach or mindset— and the same holds for teaching how to design for systems. So, we conceptualized the IoT Sandbox. 2.2
The IoT Sandbox turns the act of designing for systems into a
systemic activity itself. It explicitly limits designing in compartments.
With the IoT Sandbox we took both pragmatic steps towards
lowering the technological threshold as well as methodological steps
helping with taking a systems perspective. Physically the IoT
Sandbox is a 1:20 scale model of a family home [
In summary, the combination of enabling layers and constraining rules shape the systemic openness of the systems design challenge by ‘forcing’ students to bring their designs together and evaluate the consequences–showing the dynamics of a growing system. The IoT Sandbox makes the systemic openness concrete and manipulable while still allowing the students to explore what would happen if their designed artifact is introduced in a diferent (functional) context. 3
When the students start with their IoT Sandbox based design project they are given a presentation that both sensitizes them to the systems design challenge as well as familiarizes them with the IoT Sandbox. Former work is presented, and they are told that the previous mundane characters have decided to move and that a new family is moving in. They are given the names, ages and family relations between the mundane characters and given the task to elaborate a day in the life of those families. Using the IoT Sandbox as a three-dimensional pinboard students work through their combined scenarios of the day in the life the family, leaving traces by means of sticky notes in the process. This 2-hour workshop lays the foundation for the semester to come. After the sensitizing workshop the students start their designerly explorations while simultaneously furnishing the home and elaborating the mundane characters. In the following, we showcase a concrete group design project that developed from the basis of the sensitization and introduction to the IoT Sandbox.
In one of the first semesters that we used the IoT Sandbox, we assigned a group of master students the concrete challenge to design a family of systemic products for a house—inside the IoT Sandbox. The group chose to first work on a scenario where multiple connected products would be influenced by an overarching house ’system’ that would support a set of three family values: sustainability, comfort and social cohesion (see Figure 2). Then, each member of the group set out to design a single connected product with a specific functionality target: lighting, heating, security, media and communication (see Figure 3).
The first artifact (top-left) is a (series of) media player(s). There is a media player in each room, and they share playlists, ofer social listening (multiple people listening to the same songs) and they share presence data with the ‘system’. The second artifact (topright) is a communication mediator (unfinished) that streamlines casual communication (e.g., reminders, shopping lists, etc.) in the home. The third artifact (bottom-right) is a security system that takes presence data from the other devices as well as from dedicated sensors. It communicates potential security breaches by tilting. The fourth artifact (bottom-middle) is a (series of) light controller(s). There is at least one light controller in each room, and they ofer control over mood-lighting to support social cohesion. The light controllers use data from other artifacts to finetune the expressivity of the mood-lighting. The fifth and last artifact (bottom-left) is a (series of) smart thermostat(s). There is at least one in each room and it aims to balance energy control – it gives more rotational friction if much energy is used and it uses presence data from other artifacts to control the temperature in the rooms. 4
The IoT Sandbox and its mundane characters were perceived by many students as helpful. Yet there were also students who considered it to be an extra burden. The latter group only used the IoT Sandbox in the first weeks of the semester, mainly because we asked them to. In those latter cases the design challenge was consequently simplified to a ‘simple’ interaction design challenge and the systemic aspects were hardly touched. The successful projects, often group projects, ofered design proposals with clear systemic qualities. It was noticeable that these benefited from being group projects (tasked with designing product families, like in the example above) with multiple designs that worked as a system within the confines of the group (at the same time these were master students so that could also explain their better understanding of the challenge). While initially scarce, in the later semesters, cross-project initiatives became more numerous. While there is certainly room for improvement, important steps have been taken as we received design proposals that have clear systemic qualities and that are arguably open for growth.
The students who worked on the design case that we presented above reflected on the utility of the IoT Sandbox in their process report. They bring forward that the close proximity to the IoT Sandbox helped “to form an accurate model of orientation”, meaning that it helped ground and calibrate their design proposals. They also critiqued the IoT Sandbox and stated that “to create a specific context to design for, seems paradoxical when you simultaneously create that context yourself”. This feedback must be seen in the light of the early phase of the deployment of the IoT Sandbox in education – at that time we were still tuning the approach. Feedback such as that of this group of students, combined with our observations of the students’ design processes led us to a new implementation of the mechanisms of cooperation and competition. The students in the design case were working with more than one group in a house but there was enough room for them to be comfortably out of each other’s way. This can be read from their statements where they claim to have complete freedom of creating the environment in the IoT Sandbox. The important lesson there we learned there is that the IoT Sandbox needs to be ‘crowded’ with projects to make the mechanisms of cooperation and competition work. Pressure from other projects is needed so that there is a competition for (conceptual) space. Variation in projects is needed to create opportunity for meaningful cooperation.
Next to optimizing the implementation of the mechanisms of cooperation and competition we observed several additional points for improvement when working with the IoT Sandbox. A problem that became prominently visible during the first semesters of use is a consequence of the sandbox-like nature of the IoT Sandbox: it proved to be hard for the students to seek external validation for their design proposals ’in the wild’. In later semesters we coached our students to seek external validation, but it remains dificult to evaluate the systemic characteristics (i.e., the cross-project qualities) of the projects.
Another point for improvement concerns the use of data within but particularly between the design projects in the IoT Sandbox. In several of the projects (including the presented design case) the design proposals shared data as part of their functionality. The use of data as a material for design is one aspect of what makes these artifacts systemic artifacts. Yet, other aspects of systemic behavior such as co-dependence between artifacts, emergent functionality or even just simple forms of growth are harder to design for.
We have presented the IoT Sandbox as a tool and method for design in a systems context with explicit attention to its operation in design education. We have seen valuable work emerge from the hands of the students that worked with the IoT Sandbox and we feel that this has much to do with the mechanisms of cooperation and competition that the method builds on. The IoT Sandbox is instrumental in seeding, amplifying and enforcing mechanisms of cooperation and competition in systemic group design projects, and it is essential in making the design process systemic in itself. We will continue working with the platform and currently aim to better visualize the data that is shared in the physical model itself.