Page 449 -

440 CHAPTER 14 Online and ubiquitous HCI research

The importance of contextual factors and the novelty of the technologies involved
make up-front qualitative investigations vitally important, as misunderstandings of
contexts and how tools might be used can often compromise the goals of the system.
Interviews with users, particularly when conducted in context (Vaisutis et al., 2014;
Mynatt et al., 2001; Rowan and Mynatt, 2005) can identify key requirements and
guide design. Similarly, field studies are often vital for understanding impacts of
the tools in context (Houben et al., 2016), particularly in terms of the many privacy
and control concerns associated with unobtrusive sensors and monitors that might
capture activities that some individuals might not want recorded (Mynatt et al., 2001;
Rowan and Mynatt, 2005; Kärkkäinen et al., 2010). Despite the strengths of these
field studies, they remain expensive and challenging, leading some to look for alter-
native approaches such as scale models useful for inexpensively demonstrating and
exploring possibilities (Chatzigiannakis et al., 2014).
Other efforts have explored the use of more quantitative versions of diaries. One
exploration looked at the frequencies and locations of interactions with all things in
the home—whether digital or not—during intervals lasting several hours. Histograms
indicating the frequency of use of various objects, along with maps describing loca-
tions, provide detailed understanding of interactions with objects that might con-
ceivably be integrated into an Internet of Things (Crabtree et al., 2006). Similar
approaches to studying the use of everyday, noncomputer onjects and how those uses
might inform the Internet of Things have led to the development of “object-oriented
ethnography”—the study of how objects interact with our lives and how this under-
standing might inform the design of devices augmented with computing capabilities
(Nansen et al., 2014).
Design sessions, similar to those used in participatory design, can also be use-
ful for developing ideas. One effort examined devices, interactions, and roles to
develop a set of graphical building blocks describing roles of individuals and de-
vices in various ubiquitous computing scenarios, using these blocks to both convey
elements of design and to identify difficulties with proposed designs (Kim et al.,
2016b). Although the development of methods of this sort quickly becomes its own
research project involving multiple iterations and revisions, the results can often be
very informative.
Sensor data presents its own challenges. Inexpensive sensors for detecting motion
and sound might be available commercially, either as “ready-to-deploy” products or
as components suitable for assembly or control via Arduino or Raspberry PI hard-
ware (Diez and Posada, 2013). Transferring data from sensors to remote servers,
through wireless or USB connections will be a requirement. You will also need a plan
for storing or processing sensor data, which can be quite voluminous. A willingness
to tinker and to consider outrageous ideas can be helpful in tackling these projects,
as unexpected approaches can often be quite helpful. In response to the potential
expense and complexity of sensors used in projects like digital family portraits, one
study found that microphones taped to pipes in a basement could inexpensively and
accurately identify the use of bathrooms (sinks, showers, and toilets) and kitchen ap-
pliances (Fogarty et al., 2006).