This document summarizes the insights gathered during the seminar. We first provide an overview of the motivation for this seminar before presenting an overview of the activities that occurred during these five days. We then provide a series of outputs that we gathered in addition to the list of abstract provided on the website.

1. Motivation

Interactive systems are becoming increasingly complex and diversified, often comprised of multiple interconnected devices, with many different functionalities. They are slowly merging within our everyday objects. Such systems are becoming ubiquitous. Ubiquitous computing, or ubicomp, is a multidisciplinary field of study that explores the design and implementation of such embedded, networked computing systems. Due to the novel aspect of the technologies involved and the multidisciplinary nature of skills needed to design such systems, teaching and training new innovators in this field are not well addressed through traditional programs and instruction. Consequently, it is important to ask several questions about the training and education needed to help students become valuable members and leaders of ubicomp teams. Three central questions about ubiquitous computing education emerge: why, what and how, with the goal of enhancing ubicomp education through interdisciplinary perspectives:

• WHY is training in ubicomp needed? Is it enough to train experts in narrow domains (e.g. those who can create low-power embedded circuits, or those who can make usable applications), and then bring them together in teams that will tackle ubicomp problems? Or do we need specialized training that targets ubicomp in addition to domain expertise? There is broad consensus that we do need specialized training, but often this argument is based on intuition and anecdotal evidence. We approach this question by first asking: what are the grand challenges that we expect our students to tackle in the world (e.g. privacy, sustainability) by inventing and developing ubicomp solutions? Next, we ask: who can better address the challenges: teams of domain experts, or teams where at least some team members have specialized ubicomp education? Answers to these questions will clearly identify problems that might exist with current ubicomp educational approaches.
• WHAT should constitute training in ubicomp? Once we identify the grand challenges, we need to ask further questions. What are the values, knowledge, and skills we should train students in ubicomp? What are the topics that should be covered? How do these depend on the background of students or their degree program? Answers to these types of questions will allow us to set goals for ubicomp education.
• HOW should we teach and engage a diverse body of students? Once we identify specific goals for ubicomp education, we need to ask ourselves how those goals can be achieved. How does the unique nature of ubicomp challenge the current pedagogical approaches? How can we create new pedagogical approaches for teaching and training in ubiquitous computing? Answers to these types of questions will help create the appropriate tools to reach our ubicomp education goals.

2. Overview of the activities

Our goal was to create a community to support new forms of teaching, training, and learning in ubiquitous computing. Our activities were centered on our main questions:

• Day 1, we explored the WHO and WHY. Each participant presented briefly their research and current teaching, and highlighted what they see are the main challenges for teaching ubicomp in the morning. We then brainstormed and discussed why is is important to rethink the way we teach ubicomp material and what are the grand challenges associated to this change.
• Day 2, we explored the WHAT. In groups, we defined the curriculum for Ubicomp education for different types of students, different degree levels, as well as identified what are the learning goals. One discussion that came up relating to the limits of Ubicomp material, specifically how complex it currently is to define what is ubicomp.
• Day 3, we explored the HOW, and particularly brainstormed about the challenges related to ubiquitous education. Participants generated a list of their current active learning methods or tools and exchanged them in a speed dating fashion with each other.
• Day 4, we explored further the HOW. In groups, we developed and experienced new active learning pedagogies on ubiquitous computing topics. We also discussed pedagogies for academic ubicomp programs.
• Day 5, we wrapped up the seminar and plan for concrete actions for the future, in particular, ideas for the next Dagstuhl seminar.

3. The challenges of teaching Ubicomp (WHY)

Figure 1 illustrates the grand challenges of teaching Ubicomp from a motivation point of view. We have identified several themes including (1) who is the audience in terms of diversity, motivation, population; and how (2) these aspects particularly impact their engagement and what methods can we use to better engage with students. We also talked about the difficulty that Ubicomp brings in terms of being a multi-disciplinary field and we highlight the fact that it is difficult to choose (3) which topics should be covered and which ones should not be covered in a particular case. What are the boundaries of Ubicomp? In fact, our discussions highlighted that there is not a clear (4) definition of Ubicomp. We talked about (5) issues with the high workload of both teaching and learning about ubicomp, and how research-led teaching could alleviate some of these issues. We discussed (6) scale issues, i.e. how to teach to a large number of students (and provide feedback) when it seems that certain aspects of Ubicomp teaching (e.g. workshop activities) can only be taught to smaller groups. We pointed out the issues of (7) space and that Ubicomp teaching is based on traditional classroom but also new types of spaces such as workshops, hackerspaces, and maker spaces. Furthermore, we discussed other media types such as (8) online lectures. We also discussed more general topics such as (9) the impact of ubicomp (e.g. on business and industry) and the future of universities and how this relates to ubicomp education.

4. The Ubicomp curriculum (WHAT)

Table 1 illustrates the topics central to Ubicomp Education brainstormed during the seminar. We split the participants (including the organizers) in four groups designing curriculum (standalone lecture or program) for different students (undergraduate UG or postgraduate PG) and technical (Computer Science) or non-technical (Interdisciplinary) background). We wish for this document (that we also plan to put onto our online web platform) to be used as guidelines for teachers in order to provide a better and unified Ubicomp curriculum across different institutions and countries.

5. Existing active learning methods for Ubicomp (HOW)

Figure 2 illustrates the grand challenges of teaching Ubicomp from a method's point of view. This was the result of a brainstorming with participants following the curriculum creation. We found that (1) managing the workload was a theme recurring again (as we also mentioned it in the initial brainstorming in Figure 1). We noted that one difficulty of teaching Ubicomp was (2) the lack of differentiation with other CS, HCI or Design teaching material. We also though this could create issue in (3) attacking certain types of students and that possibly, depending on the demographic, different terminology (Ubicomp, Interactive Systems, Interaction Design etc.) might be used. We raised issues in (4) evaluation and assessment potentially raised by (5) the interdisciplinarity of the community which makes it hard to assess student but also to teach so diverse material. We discussed issue in (6) engaging with students and enforcing skill acquisition (surface vs. deep learning). Finally we also add other issues such as (7) scaling of students, (8) project styles, (9) reaching to real end-users, (10) having input from industry and the (11) format of the lecture (e.g. online). We finally discussed about the issues raised by (12) admin and physical resources.

6. Innovative active learning methods for Ubicomp (HOW)

The next main activity focused on generating new educational material that may be difficult to generate, or missing, from a current curriculum. Participants formed six groups, they selected a topic, and investigated new active learning methods as well as initial teaching material related to the topic. The specific topics for each group were selected from topics and challenges highlighted earlier in the seminar. Next, groups formed pairs of groups, and each group tested their content and methods on other group and received feedback, before iterating on their design. Finally, teams presented a summary of their new materials to the group (Table 2).

From the discussions following up the presentation we also noted some actions to do:

• Create a repository or playlist (youtube) or videos that can be used within the community and define what is Ubicomp.
• Ask participants to upload a 2 minutes video of their definition of Ubicomp that can be used in class to show the variety of what people think is Ubicomp.

7. Future Steps

Although this seminar addressed many questions the organisers had originally highlighted, it also opened new exciting directions to explore and new challenges. To start addressing them we identified the main following avenues for future work and future events:

• Follow-up Dagstuhl seminar on writing a textbook
• Follow-up Dagstuhl seminar focussing on the industrial side, e.g. what skills do students need for the society we will built in 5/10/20/50 years?
• Using the website to keep the community alive as well as the access to material, and also create a video channel to create a repository of ubicom examples.
• We also have discussed about 3-4 follow up papers to be written among participants and organizers.

8. Reading List

We collected a reading list that addresses the why, what, and how of ubicomp education, designed for educators.

• Pervasive Computing Education, Audrey Girouard, Andrew L. Kun, Anne Roudaut, Orit Shaer, and Andrew L. Kun, IEEE Pervasive Computing, Oct, 2018.
• Teaching Pervasive Computing in Liberal Arts Colleges, Orit Shaer, and Evan M. Peck IEEE Pervasive Computing, Jul, 2018.
• The Fuzzy and the Techie: Why the Liberal Arts Will Rule the Digital World, Scott Hartley
• Fixing Tech's Ethics Problem Starts in the Classroom, Stephanie Wykstra, The Nation.
• The Pervasive, Embedded, and Mobile Computing Curriculum – Preparing Computer Science Students for the Technology of the Future, Jakob Bardram, 2012
• Bridging the Gap Between Teaching and Research: A Case Study for Engineering & Applied Science, Anne Roudaut, Higher Education Pedagogies 2019
• Krumm, J. (Ed.) Ubiquitous computing fundamentals. CRC Press, 2010 (fairly outdated by now)
• Rowland et al. Designing Connected Products. O'Reilly 2015 (on Design of IoT products, with a broad range of topics ranging from networking aspects, architecture to product design)
• Landay, J. A., & Borriello, G. (2003). Design patterns for ubiquitous computing. Computer, 36(8), 93-95.
• Electronics books from "Make".
• The radar diagram https://scottwhyoung.com/teaching/information-ethics-privacy-spring- 2018/

Creative Commons BY 3.0 Unported license

Audrey Girouard, Andrew Kun, Anne Roudaut, and Orit Shaer