Interview with Patrick Baudisch, by Doantam Pham, October
Doantam: What makes a good
Patrick: This is a tricky
question, because interaction design can be looked at from different
perspectives. Like many researchers in this field I am wearing two
When working on a project with concrete existing users I
often need to adopt techniques from usability engineering. Here user
centered design is the key component and the relevant skills are rapid
prototyping and evaluation. While a good amount of intuition can go a long
way, I think what makes an outstanding interaction designer is the ability
to successfully design for a target audience that one is not part of. This
requires knowing the limits of ones intuition.
Here is an example.
Many of us can fairly reliably distinguish a usable web page from a less
usable one. That's because we can use the intuition we gained from being
web users ourselves. When designing for a target audience we are not part
of, on the other hand, this intuition often fails. When designing a
customized application for a group of medical doctors, for example, as
good interaction designers we realize that the needs of these people might
be different from what we think. Consequently, we hold back our personal
intuition, and instead apply formalized methods, such as prototyping and
When working on my research projects, on the other
hand, things are often different. Now the perspective is 5-10 years out,
which means that I am trying to envision what _future_ users will need.
Many of user centered design techniques require talking to actual users,
but I obviously cannot go out and interview future users because they
don't exist yet. Take wall-size displays, for example. Yes, there are a
handful of users today, but these tend to be expert users, working with
very specialized applications. When I am envisioning a general purpose
user interface for a future where people use wall-size displays casually,
I consequently find myself in the room of speculation. I apply user
centered design techniques where possible, but basic research methodology
and intuition tend to play a bigger role here.
Doantam: How did you get started working on
knowing it. I was a Ph.D. student in Darmstadt, Germany and worked on user
interfaces for information filtering systems. A friend of mine saw my work
and said "oh, I did not know you were in HCI, too".
was the first time I heard of that field.
Doantam: What's your current research area?
Patrick: The first of my two main research areas is to
envision a future where users have access to very large personal displays.
I tend to think of them as high-resolution wall-size displays, but in many
cases multi-monitor setups, as they are already in use today bring up the
same questions and afford similar solutions. My first project in this
space was focus-plus-context screens: a large projection-based display
into which I seamlessly embedded a high-res LCD display. This resulting
display imitates the structure of the human eye with its low-res periphery
and hi-res fovea.
Since then, I have focused mostly on helping
users interact with different types of displays. I have worked on a series
of techniques to extend the mouse to work with very large displaysurfaces.
The projects I enjoy working on most are techniques for reaching distant
display content on wall-size touch screens. While previous work suggested
extending the users' reach, we propose the opposite approach, i.e.,
bringing content to the users instead. Our user studies show that our
approach is not only faster, but also more accurate and less tiresome,
because our approach allows users to interact with content within their
natural arms reach. If you are curious, you can try out prototypes of
"drag-and-pop" and "tablecloth" on my homepage
My second main research area is small
screens. Currently I am focusing on visualization techniques that help
users view large documents, such as maps or web pages on the small
screens. A recent project is "summary thumbnails". We display
web pages as miniature versions that are just wide enough to fit the width
of the screen. Unlike techniques proposed by other researchers, our
thumbnails are readable because we delete text fragments instead of
shrinking fonts into unreadability. What makes me passionate about phones
is the huge yet still growing impact they have on people worldwide. We
start seeing more and more tasks that we used to perform on PCs now
covered by phones, things like checking email, driving directions, or
shopping. I believe that this trend will continue and I am wondering how
far it will go. In the not so distant future, we will look back and we
will find it surprising that we used to log into a PC to do this. New
visualization and interaction techniques will make a tremendous
Doantam: How do you select a
particular project to work on? What's the design process you take? How
long do you usually work on a project? How many projects do you work on at
a time? At what point do you consider a project finished?
Patrick: I always have a large number of projects I
work on at the same time. Somewhere is the range of, say, three to eight.
Then, when I am getting closer to a deadline, such as a paper deadline, I
narrow this down to 1-3 projects that I can actually get done in time. The
others go back into the pool. I may pick them up later--or not.
The core motivation behind my approach is that it helps me let go of a
bad project idea. Not all ideas can be "good", and sometimes an
idea that looked promising in the beginning turns out not to be novel, not
applicable, or simply "not that great". By simultaneously
pursuing multiple ideas, letting go of a bad idea is much less painful. As
a result, I paint myself into a corner less often then I used to do, when
I put all my eggs in one basket.
Having a rich pool of ideas to
choose from is crucial for my approach. To make sure I don't forget ideas
I always record them right away, typically using the voice recorder in my
phone. Sometimes I take voicenotes late at night and it is quite
challenging to decipher my mumblings the next morning. While many of my
ideas are in my field of interaction and visualization techniques, I keep
a log of ideas for pretty much everything--from information filtering to
what I think might be an easier to use umbrella.
What are the
other challenges in interaction design (other than what you are currently
Patrick:The space of interaction design and
interaction is full of exciting challenges. While the underlying hardware
of our systems still advances at a dramatic rate, the capabilities of
users do not.
Enhancing this "bottleneck" is therefore as
important as never before. Cell phones and other small screen devices
offer a host of interesting challenges way beyond the visualization
questions I am working on today.
To just pick one example, I would
love to see multimodal interface researchers join forces with interaction
designers to design eyes-free interfaces. Such a system might allow me to
query my calendar while I am on the phone. Phones in particular also have
the potential of reaching a much wider audience worldwide than PCs ever
have or will. Rural computing will be a particularly rewarding field to
In the wider space of computer science/design I feel that
technology around online communities had brought up questions we have not
been able to answer so far. The copyright question around file sharing, is
an example that could benefit from our attention. Ten years ago, engineers
talked about the "information highway" and all the technical
challenges that would need to be solved in order to deliver video to the
Today, the technical part is more than solved by file
sharing systems. Even the question about micro payments, a big question
still a couple of years ago, has made a huge step forward with the broad
acceptance of online payment systems, such as PayPal. But downloading
video content, it turned out, is foremost not a technical challenge--it is
a legal question. Today, a lot of work is done on enforcing copyright
using digital rights management software. Can we go beyond that--can we
come up with a system that creates value from the existing infrastructure
rather than limiting it?
Copyright 2005 Ambidextrous