Page 33 -

18 CHAPTER 1 Introduction to HCI research

in this topic shows the importance of the research area and strengthens your findings.
Perhaps you should be more worried if no one else is interested in your research.

1.9 INHERENT TRADE-OFFS IN HCI
It would at first seem that, with enough research, you could simply decide which
design is best by optimizing some specific measurement, such as task performance
or time performance. First of all, as discussed earlier in this chapter, socio-technical
systems can rarely be reduced to two or three measurements, and there are many fac-
tors to be controlled for. We can do comparison studies of small differences in menu
structure or some detailed aspect of interface design, but it is much harder to com-
pare fundamental recastings of tasks. In addition, there are inherent conflicts in HCI
research and design. We make trade-offs and accept “better solutions” rather than
optimal solutions. We have multiple stakeholders and not all of them can be satisfied.
Design is not simple and it's not an optimization problem. Good HCI research allows
us to understand the various factors at play, which design features may work well for
which users, and where there are potential conflicts or trade-offs.
For example, we can learn how to make interfaces that are far better than our
current interfaces. However, users may not prefer those interfaces because they are
so different from the current interfaces. So maybe we should modify our interfaces
gradually, making only minor changes each time? Keyboards are a perfect example
of this. We know how to make keyboards that are more ergonomic, with key layouts
that allow for much faster typing. However, the keyboard layout predominantly used
with the Roman alphabet is still the QWERTY key layout. Why? We have far supe-
rior designs. However, people have been comfortable with the QWERTY layout for
years and the other key layouts have not caught on (despite their clear superiority
from a design and usability point of view). So we still use the QWERTY layout. It's
a trade-off. You want to make interfaces that are much better but users want consis-
tency. In the short-term, a totally new interface lowers user performance, increases
user error, and lowers user satisfaction. In the long-term, a modified interface may
improve performance and result in higher satisfaction. This focus on very minor
tweaks can be seen in the attention currently being paid, in industry and government,
to the idea of A/B testing, where you test very minor interface changes, barely notice-
able by the user, and then roll out those that are deemed to be successful, increasing
traffic, increasing sales, and reducing costs (Wolfers, 2015). Of course, there are
sometimes new interfaces, new devices, that just leap ahead with a totally different
design and users love it, such as the Apple iPad tablet device. You shouldn't create a
totally new design, apparently, unless it's something so cool that users want to spend
the time to learn how to use it. Well, how do you measure that? How do you decide
that? How do you plan for that? It's not easy.
Other examples of trade-offs in HCI also exist. For instance, the intersection of
usability and security (Bardram, 2005; DeWitt and Kuljis, 2006). In HCI, we want
interfaces that are 100% easy to use. People focused on computer security want