A Taxonomy of Ambient Information
Systems: Four Patterns of Design
Zach Pousman, John Stasko
Information Interfaces Laboratory
Georgia Institute of Technology
What is an Ambient Information System?
Scope
Pinwheels
InfoCanvas
Informative Art
What is an Ambient Information System?
Sideshow
Ambient Orb
Digital Family Portrait
LiveWire
The Buzz
Notification Collage
Information
Percolator
Purpose
Settle on definitions of this class of
UbiComp systems
Provide a set of design dimensions and
design patterns that developers have used
Identify open areas in the design space
Identify “growth areas” in this design space
A definition
Ambient Information Systems
Representations in the environment (peripheral in space), Focus on
the tangible
Can move from the periphery to the focus of attention and
back again
Are aesthetically pleasing and environmentally appropriate
Display information that is important but not critical
A definition
Ambient Orb
Scrolling Ticker
Peripheral
Systems
Ambient
Systems
System Tray Icons (Mail)
Notification
Systems
Command and Control System
Previous Taxonomies
McCrickard et al
ToCHI ‘02
Design space split along 3 axes
interruption:
“we define interruption as an event prompting
transition and reallocation of attention focus from a
task to the notification.”
reaction:
“The second critical parameter we propose is the rapid and
accurate response to the stimuli provided by notification
systems, an effect which we refer to as reaction.”
(user responds to the stimulus)
comprehension:
“…often it is also ... vital to use notification systems with
the goal of remembering and making sense of the
information they convey at a later time.”
(sense-making of data with or without behavioral ‘reaction’)
Previous Taxonomies
McCrickard et al
ToCHI ‘02
Previous Taxonomies
McCrickard et al
ToCHI ‘02
Previous Taxonomies
Matthews et al
Berkeley TR ‘02
Design space has 3 main categories, each subdivided:
Notification:
A relative metric assigned to the
stream of information - its “importance”
Transition:
Explicitly supported switches between new
and old data (animations, fading, scrolling, etc)
Abstraction:
Mapping between information and display
Ignore
Change Blind
Make Aware
Interrupt
Demand Attention
Change Blind
Make Aware
Interrupt
Demand Attention
Degradation
Feature Abstraction
User Controlled Ambient Media. As another aspect of matching virtual information
with ambient media we investigated the concept of user controlled ambient media,
and developed end user tools for management of ambient links. To require users to
decide about usage of ambient media is quite in contrast to the philosophy that users
best not be aware at all of the technology that blends into their surroundings. Yet, to
have ambient media at one’s disposal to create highly individualized ambient spaces
seems very intriguing. Also given that the decision on choice of ambient media is far
from understood but clearly depends on users’
tasks and preferences (cf. Wisneski et al, 1998),
it may make sense to defer the decision from
design time to run time.
users easily aware of sounds over being
lly pleasing. Our evaluation results show that all
isplays enabled easy identification of sounds,
this requirement.
mportant that critical information displays
tely distract users. Since the criticality of
on shown on IC2Hear varies, users made it clear
wanted overt notifications for very critical sounds
distracting updates for less critical sounds. They
that the icon flash when the sound was “really
” and be more “visual[ly] quiet” for less important
We found that Single Icon and Spectrograph with
not adequately distinguish between sounds of
criticality. For this reason, users favored the
Display, which emphasized loud sounds (which
us were more important) by using larger bars.
group, alerting the user when new messages arrive. The Web tool summarizes headers, lists, and
hypertext links on a Web page, allowing the user to monitor news wires and hotlists. The weather tool
monitors the weather conditions and
forecast
for a and
given
city.
When users initially start Irwin, it tries to
Figure
3: Patterns
Table
Fountain
determine their informational interests by examining their email files and Web bookmarks. Since users
critical issue in giving the user control over
are not required to answer a listambient
ofAquestions
at startup, they are often more likely to use Irwin. The user
media is the effort required for creacan then configure both the information
monitored
and theofway
in which
tion, modification
and deletion
ambient
links. it is displayed after getting
used to the power and flexibilityWeofdescribe
Irwin. two tools that we developed for managing ambient links, both concerned
Design Space
with provision of easy-to-use mechanisms to link virtual information to ambient me-
The simultaneously,
first tool is for relating
page
access to ambient
presentations,
and the
To show both an overview and dia.
details
Irwinweb
uses
a confluent
zoom display.
A confluent
second
one extends
an email filter
for each
ambient
notification.
examples
uses
zoom provides multiple views of
a resource
simultaneously
with
view
having aBoth
different
level
of a
small table
fountain
ambient
medium.
granularity. With a confluent zoom,
the user
canassee
overview
and details with minimal effort, and the
context of the detailed views is Table
maintained
at all times by the coarser views. Since intermediate views
Fountain as Ambient Medium. A simple but nevertheless useful and pleasant
have proven useful in graphicalambient
tools such
browsers
medical
imagery
[11,
we that
expect
mediaasismap
a small
fountain and
placed
in our office.
It has
four12],
pumps
can be
that the intermediate views willswitched
smoothonthe
from the
high-level
tothe
thepower
detailed
andtransition
off independently
(see
figure 3). Aoverview
PC controls
plugszoom
over
view.
the parallel interface board. The same interface can be used to integrate other devices
of critical information should be easy to learn.
e skeptical about using the Spectrograph because
ght it was too difficult to learn. Users thought
on and the History Display were easy to learn,
these two displays over Spectrograph with Icon.
that are controlled by switching on and off the power, for instance we have also used
splays of highly critical information should make
ns easy to discover. Single Icon and Spectrograph
n make breakdowns apparent since they are
y changing with new sounds. If unsure, users can
t the display by clapping their hands loudly and
the display change. The History Display,
only shows recognized sounds (phone, voices,
ning/closing, door knocking) which happen less
is could impede users from noticing a system
n. One possible solution would be to display
zed sounds on the History Display as well.
e
Mobile is a physical display of bus arrival
on (see Figure 4) [13]. It includes 6 physical
ach representing a bus. A token stays under a
rtain until its
White Curtain
minutes from
stop, at which
wers 25 inches.
h minute, the
ves up 1 inch,
that the bus is
closer to the
hen the token
the
white
the bus has
of our design
ns, the Bus
an opposite of
The
Bus
as a scope of 1
is meant to
one activity:
intensities, colors, and pulses and agreed with each other
ultrasonic humidifiers in other demonstrators. The media controller for the fountain is
upon private meanings to their interaction. One user used
CGI-program running on the control PC, and generating control information send to
a pulsing green and blue series to communicatea“good
luck”. A flashing red meant, “Good bye, I’m busy now.”
Figure 3: (Clockwise) A LumiTouch when off, a user working with
LumiTouch in passive mode, the feedback panel on top lights up when
sending a message, the display lights up when receiving a message, a
user interacting with LumiTouch.
Passive Communication
When a user is in front of her LumiTouch, the
corresponding LumiTouch will emit an ambient glow to
indicate her remote presence. This additional background
information helps people figure out if there is a recipient
on the other end, or also if it is a convenient time to
increase interaction levels (e.g. start active communication
mode or use an alternative, like a phone).
Active Communication
When a user picks up the picture frame and squeezes, the
feedback display area illuminates to show that the picture
frame has been squeezed. The display colors are
transmitted over the Internet to the corresponding remote
LumiTouch. The display varies depending on the squeeze
attributes (where, how hard, and how long the user
squeezed). While the recipient party can simply enjoy this
display, s/he has the option to pick up the frame and
squeeze back a response to the first user, and begin an
interactive exchange.
Interpersonal Language
A Bus Token
Figure 4: Bus Mobile.
The system was intended to allow users to develop an
abstract form of emotional language. People could
communicate in real-time by sending each other color
mixtures and light patterns. The combination of colors and
force allowed a grammar, while the duration of squeeze
provided syntax for creative interpersonal dialect between
two people.
PRELIMINARY OBSERVATIONS
Preliminary testing was done with users who were not
emotionally involved with each other. Figure 3 illustrates
some user interactions. More than one user noted that they
would find having a personal LumiTouch useful.
The generation of an abstracted language was observed in
many cases. Users created different combinations of light
In its confluent zoom, Irwin employs an auditory cue, an
icons view, a navigation bar, and several textual views. In
the icons view (the leftmost view), the state of each
An observer commented that they would like to give
resource is given by an icon. When the resource changes,
LumiTouch to a sibling who is currently in the hospital.
People who are unable to actively communicate for long
the appearance of the icon changes and an auditory cue is
periods of time (e.g. sick or elderly) might be able to use
the passive transmission of LumiTouch. Similarly users
played. If user selects an icon, the other views are updated
who lack the required dexterity or concentration for
to show information about the corresponding resource. A
pushing numerous buttons might appreciate this system
due to its small number of simple grasping inputs.
navigation bar (second from the left) shows a syntactic
Another observer noted that a potential problem might
encoding
of the resource contents and can be used to select the messages displayed in the textual views
arise from the asymmetry of participation. One person is in
active two
interaction
while theviews).
other person
passively
(the
rightmost
Users
can configure the orientation and placement of the icons with respect to
perceives ambient data. The transmitting person may be
the
other
views
and
canrespond.
evenHowever,
choose
anxious
and expect
that the
receiver
if to hide the other views until some event happens, e.g. an icon is
someone is so anxious that they feel the need to
clicked.
communicate immediately, they can just use the phone, or
other higher-bandwidth alternative.
CONCLUSION AND
FUTURE WORK
Auditory
cues
In closing, we have described our design of an emotional
communication device that can transition between passive
and active mode
seamlessly.
We havea also
outlinedto
some
Auditory
cues
provide
means
attract users to a change that has occurred. They may be the only way
preliminary observations of the usability of LumiTouch. It
the
user
knows
about
if the display is obscured or not in the user’s current line of sight.
appears
the dual
level of
attentionthe
this change
device supports
appeals to people. Eventually we hope LumiTouch will aid
in explorations of the development of interpersonal
emotional
language.sounds
We hope that
to gather
data by
We
selected
aremore
distinctive
but are not distracting and, most importantly, are short. Each is
adding support for the manipulation of real-time data to
less
than two
which we expect will be long enough to be noticed but not too long to
the LumiTouch.
Usersseconds
could recordin
andduration,
replay the touch
history, like an
a voice
mail message. Maybe
the history
of configuration, we tried to select sounds that were indicative of the
become
annoyance.
In our
initial
touch “conversations” and patterns can be documented.
resource;
for
example,
a
dog
bark
for
mail (dogs always bark at the mail carrier) and a splat for news
ACKNOWLEDGMENTS
(similar
toMedia
a newspaper
hitting
a front porch). Of course, the user can select from a list of sounds (we
We thank MIT
Lab for their support
on this project.
We also thank our loved ones for their inspiration.
provide
twenty) or create and use their own sounds.
REFERENCES
1. Tollmar, K., Junestrand, S., and Torgny, O., Virtually
Living Together.DIS’00 (Brooklyn NY, August 2000),
Icons
ACM Press, 83-91.
2. Strong, R., and W. Gaver. Feather, Scent, and Shaker:
Supporting Simple Intimacy. Proceedings of CSCW '96
(Boston MA, November 1996), ACM Press, 29-30.
3. Dodge, C. The Bed: A Medium for Intimate
Communication. Extended Abstracts of CHI’97, ACM
Press, 371-372.
4. Brave, S. and Dahley, A., inTouch: A Medium for
Haptic Interpersonal Communication, Extended
Abstracts of CHI ‘97, ACM Press, 363-364.
5. Fogg, B.J, Cutler, L.D., Arnold, P. and Eisbach, C.,
HandJive: a device for interpersonal haptic
entertainment, CHI ’98, ACM Press, 57-64.
7,((!&/%0!.)0",8%+!#!",$'(%!&%./0,I2%!4)+!4#3+,.#&,01!,$#
&/#&! #+%! (,>%(9! &)! 3%! +%'+%"%0&#&,*%! )4! ,04)+$#&,)0! 1,*%0
)0(9!&%;&2#(!4)+$5!!J/,"!,"!4)(()7%8!39!#!8,".2"",)0!)4!
%;&%0",3(%! )'&,$,F#&,)0! '+).%""! &/#&! 4)+$"! &/%! .)+%! )4!
H#08,0">9!"9"&%$5!!K%;&-!#!8,".2"",)0!)4!*,"2#(!'+)'%+
7,((! 3%! '+)*,8%8-! 4)(()7%8! 39! #! 8%&#,(%8! 8,".2"",)0! )4!
#"'%.&"! #.&2#((9! 2"%8! 39! &/%! "9"&%$-! #08! /)7! &/%9!
,$'(%$%0&%85! ! 6%! 7,((! &/%0! "/)7! ")$%! +%"2(&"! )4!
"9"&%$!#08!'+)*,8%!#!.)0.(2",)05!
!"#$%&'#&("!)*+,'",%'-*
Icons are used in interfaces because they provide a universal representation in a small amount of space.
In a WIMP environment they can both provide information and invite the user to click to obtain more
information.
!
!"#$%&'()'*+,-./&'01//,#&'2&3&%,4&5'!%1-6'
.#/0*(123456*7889*:*);<=*1>=;?2@A*"=BCD<BE*-4>?BF46GH*
!"#$%&$'&$$'()$'*+,'-./.(%#'#.01%12'
'341'54#41'6$1&.47&'43'%##'3./81$&9:!
"#$%!&%$'(#&%!)*%+!&,$%-!%#./!+%'+%"%0&,01!#!"$#((!02$3%+!
)4! $%""#1%"5! ! 6%! 7)2(8! &/%0! "()7(9! "/,4&! 3%&7%%0! &/%!
L,12+%! M! ,((2"&+#&%"! &/%! #+./,&%.&2+%! )4! &/%! H#08,0
"9"&%$5!!J9',.#(!2"%!3%1,0"!7,&/!&/%!#++,*#(!)4!,04)+$#&
&/#&!7%!$,1/&!7#0&!&)!3%!8,"'(#9%8!#$3,%0&(95!!G0!&/%!7
8%".+,3%8! /%+%-! 7%! /#*%! ./)"%0! %(%.&+)0,.! $%""#1%"! ?%
%$#,(! )+! 0%7"! 1+)2'! ')"&,01"D! #"! #! +%#89! #08! 4#$,
")2+.%! )4! ,04)+$#&,)0-! 32&! $#09! )&/%+! ")2+.%"! .)2(8
2"%8! ,0"&%#85! ! L+)$! &/%! ,0.)$,01! ,04)+$#&,)0-! 7%! %;&+
&%;&2#(!"2$$#+,%"!"2./!#"!$%""#1%!"23<%.&!(,0%"-!#"!7%(
#09! "2,&#3(%! ,$#1%"! &/#&! $,1/&! .)$%! 7,&/! &/%! ,04)+$#&
?4)+! &/%! %;#$'(%"! ,0! &/,"! '#'%+! 7%! /#*%! ./)"%0! &)! 7
)0(9! 4+)$! &%;&!&)!,((2"&+#&%!&/%!$)+%!./#((%01,01!'+)3(%
Our Taxonomy
Information Capacity
Notification Level
Representational Fidelity
Aesthetic Emphasis
Information Capacity
How much info the system can display
Quanta of information - “nuggets”
Tradeoff: Time vs. Space
Information Capacity
Low
Medium
High
Notification Level
How interrupting is the system?
User Request
Change Blind*
Make Aware*
Interrupt*
Demand Attention*
*Italicized categories are from Matthews et al (2002)
Notification Level
group, alerting the user when new messages arrive. The Web tool summarizes headers, lists, and
hypertext links on a Web page, allowing the user to monitor news wires and hotlists. The weather tool
monitors the weather conditions and forecast for a given city. When users initially start Irwin, it tries to
determine their informational interests by examining their email files and Web bookmarks. Since users
are not required to answer a list of questions at startup, they are often more likely to use Irwin. The user
can then configure both the information monitored and the way in which it is displayed after getting
used to the power and flexibility of Irwin.
User Request
To show both an overview and details simultaneously, Irwin uses a confluent zoom display. A confluent
zoom provides multiple views of a resource simultaneously with each view having a different level of
granularity. With a confluent zoom, the user can see overview and details with minimal effort, and the
context of the detailed views is maintained at all times by the coarser views. Since intermediate views
have proven useful in graphical tools such as map browsers and medical imagery [11, 12], we expect
that the intermediate views will smooth the transition from the high-level overview to the detailed zoom
view.
Make Aware
Interrupt
In its confluent zoom, Irwin employs an auditory cue, an
icons view, a navigation bar, and several textual views. In
the icons view (the leftmost view), the state of each
resource is given by an icon. When the resource changes,
the appearance of the icon changes and an auditory cue is
played. If user selects an icon, the other views are updated
to show information about the corresponding resource. A
navigation bar (second from the left) shows a syntactic
encoding of the resource contents and can be used to select the messages displayed in the textual views
(the two rightmost views). Users can configure the orientation and placement of the icons with respect to
the other views and can even choose to hide the other views until some event happens, e.g. an icon is
clicked.
Auditory cues
Auditory cues provide a means to attract users to a change that has occurred. They may be the only way
the user knows about the change if the display is obscured or not in the user’s current line of sight.
We selected sounds that are distinctive but are not distracting and, most importantly, are short. Each is
less than two seconds in duration, which we expect will be long enough to be noticed but not too long to
become an annoyance. In our initial configuration, we tried to select sounds that were indicative of the
resource; for example, a dog bark for mail (dogs always bark at the mail carrier) and a splat for news
(similar to a newspaper hitting a front porch). Of course, the user can select from a list of sounds (we
provide twenty) or create and use their own sounds.
Icons
Icons are used in interfaces because they provide a universal representation in a small amount of space.
In a WIMP environment they can both provide information and invite the user to click to obtain more
information.
Representational Fidelity
Is the mapping from information to visual
abstract or literal
What is “abstract”? - Language of
Semiotics
Indexical (maps, etc.)
Iconic (doodles, caricatures)
Iconic (metaphors)
Symbolic (language symbols)
Symbolic (abstract symbols)
Representational Fidelity
Is the mapping from information to visual
abstract or literal
What is “abstract”? - Language of
Semiotics
Indexical (maps, instruments, etc.)
Iconic (doodles, caricatures)
Iconic (metaphors)
Symbolic (language symbols)
Symbolic (abstract symbols)
Symbolic Symbolic
(abstract) (language) Metaphorical Iconic Indexical
Aesthetic Emphasis
Look at the designer’s intent, not results.
(Common) Tradeoff between Information Capacity and Aesthetics
Aesthetic Emphasis
Low
Medium
High
7,((!&/%0!.)0",8%+!#!",$'(%!&%./0,I2%!4)+
&/#&! #+%! (,>%(9! &)! 3%! +%'+%"%0&#&,*%! )4! ,0
)0(9!&%;&2#(!4)+$5!!J/,"!,"!4)(()7%8!39!
%;&%0",3(%! )'&,$,F#&,)0! '+).%""! &/#&! 4)+
H#08,0">9!"9"&%$5!!K%;&-!#!8,".2"",)0!
7,((! 3%! '+)*,8%8-! 4)(()7%8! 39! #! 8%&#,(%
#"'%.&"! #.&2#((9! 2"%8! 39! &/%! "9"&%$-!
,$'(%$%0&%85! ! 6%! 7,((! &/%0! "/)7! ")
"9"&%$!#08!'+)*,8%!#!.)0.(2",)05!
!"#$%&'#&("!)*+,'",%'-*
!
!"#$%&'()'*+,-./&'01//,#&'2&3&%,4&5'!%1-6'
.#/0*(123456*7889*:*);<=*1>=;?2@A*"=BCD<BE*-4>?BF46GH*
!"#$%&$'&$$'()$'*+,'-./.(%#'#.01%12'
'341'54#41'6$1&.47&'43'%##'3./81$&9:!
"#$%!&%$'(#&%!)*%+!&,$%-!%#./!+%'+%"%0&,01!#!"$#((!02$3%+!
)4! $%""#1%"5! ! 6%! 7)2(8! &/%0! "()7(9! "/,4&! 3%&7%%0! &/%!
.)((#1%"!&)!'+%"%0&!#!(#+1%+!3)89!)4!,04)+$#&,)05!!!
:+&,"&"! %;/,3,&! #! .#'#.,&9! 4)+! .+%#&,*,&9! #08! #%"&/%&,.!
<281$%0&! &/#&! ,"! '))+(9! 208%+"&))8! #08! 0)&! .2++%0&(9! 7%((!
"2,&%8! &)! #2&)$#&,)05! ! =#&/%+! &/#0! #&&%$'&! &)! 32,(8! +)32"&!
>0)7(%81%! )4! #%"&/%&,.! '+,0.,'(%"! ,0&)! &/%! "9"&%$-! )+! &)!
.+%#&%!#!4)+$!)4!#+&,4,.,#(!#+&,"&!?#!*%+9!./#((%01,01!&#">-!"%%!
4)+! %;#$'(%! @A-! BCD-! 7%! /#*%! ./)"%0! ,0"&%#8! #0! #''+)#./!
&/#&!#(()7"!&/%!#+&,"&!&)!%;'+%""!#%"&/%&,.!.)0.%'&"!,0!*,"2#(!
4)+$!E!#0!!"#$%"$&'($")*+!$"5!!:%"&/%&,.!&%$'(#&%"!#(()7!#0!
#+&,"&! &)! %;'+%""! )*%+#((! .)$')",&,)0#(! 4)+$-! 8)$,0#0&!
.)()+"-! #08! )&/%+! #%"&/%&,.! '+)'%+&,%"! &/#&! &/%! 4,0#(! .)((#1%!
,"! &)! 4)(()7-! #()01! 7,&/! 4,;%8! %(%$%0&"!)4!,&"!.)0&%0&5!!6%!
&/%0!#&&%$'&!&)!+%2"%!&/#&!%;'+%"",)0!39!.#"&,01!&/%!.)((#1%!
1%0%+#&,)0! '+).%""! ,0&)! #0! )'&,$,F#&,)0! '+)3(%$! E! )0%! )4!
$#&./,01! #08! $#0,'2(#&,)0! )4! &/%! *,"2#(! '+)'%+&,%"! )4!
')&%0&,#(!.)((#1%!%(%$%0&"!,0!)+8%+!&)!$,$,.!&/%!'+)'%+&,%"!
)4! *#+,)2"! '#+&"! )4! &/%! &%$'(#&%5! ! G0! &/,"! 7#9-! 7%! 2"%! &/%!
.)$'2&%+! 4)+! #! +%'%&,&,*%! #(1)+,&/$,.! &#">! &/#&! ,&! ,"! 7%((!
"2,&%8! 4)+! ,0! )+8%+! &)! (%*%+#1%! #08! #$'(,49! &/%! >0)7(%81%!
#08!">,((!)4!#!/2$#0!#+&,"&5!!!
G0! &/%! 0%;&! "%.&,)0! 7%! 7,((! '+%"%0&! #0! )*%+*,%7! )4! &/%!
.)$')0%0&"!#08!3#",.!#.&,)0!)4!&/%!H#08,"0"9!"9"&%$5!!6%!
L,12+%! M! ,((2"&+#&%"! &/%! #+./,&%.&2+%!
"9"&%$5!!J9',.#(!2"%!3%1,0"!7,&/!&/%!#++
&/#&!7%!$,1/&!7#0&!&)!3%!8,"'(#9%8!#$3
8%".+,3%8! /%+%-! 7%! /#*%! ./)"%0! %(%.&+)
%$#,(! )+! 0%7"! 1+)2'! ')"&,01"D! #"! #!
")2+.%! )4! ,04)+$#&,)0-! 32&! $#09! )&/%+
2"%8! ,0"&%#85! ! L+)$! &/%! ,0.)$,01! ,04)+
&%;&2#(!"2$$#+,%"!"2./!#"!$%""#1%!"23<
#09! "2,&#3(%! ,$#1%"! &/#&! $,1/&! .)$%! 7
?4)+! &/%! %;#$'(%"! ,0! &/,"! '#'%+! 7%! /#
)0(9! 4+)$! &%;&!&)!,((2"&+#&%!&/%!$)+%!./#
6/%0!,$#1%"!#+%!0)&!'+)*,8%8-!7%!2"%!
7,&/! &/%! G$#1%N)0<2+%! "23"9"&%$! ?8%"
"%.&,)0D! &)! 4,08! ,$#1%"! 7,&/,0! #! (#+
.)((%.&,)0! &/#&! #+%! (,>%(9! &)! +%4(%.&! &/%! "
&/%! $%""#1%5! ! J/%! H#08,0">9! "9"&%$! &/
,$#1%"! )0! &/%! 3#","! )4! #0! #%"&/%&,.! &%
.)((#1%5!!!
:"!8,".2""%8!,0!&/%!"%.&,)0!#4&%+!0%;&-!#0
,"!8%4,0%8!2",01!#!.)((%.&,)0!)4!(#9%+%8!,
#0! #+&,"&5! ! =%1,)0"! .#0! 8,+%.&(9!
.)$')0%0&"! &)! &/%! 4,0#(! +%"2(&-! .#0! "
$#&./,01! .+,&%+,#-! #08O)+! .#0! "'%.,49! ,$
&)!3%!'%+4)+$%85!!!
J/%! H#08,0">9! "9"&%$! 2"%"! &/%! (#9
.+,&%+,#! )4! &/%! &%$'(#&%! ?4)+! %;#$'(%
$#&./,01!#1#,0"&!&/%!&%$'(#&%!,$#1%D!#"!
)'&,$,F#&,)0! '+)3(%$5! ! G0! '#+&,.2(#+-! %
"9"&%$!.)$%"!7,&/!#!".)+,01!420.&,)0!7
7%((!&/%!.+,&%+,#!/)(8"!7,&/,0!#!1,*%0!.)$
.)$')",&,)05! ! =)21/(9! "'%#>,01-! &/%! "
"%+,%"!)4!.#08,8#&%!#"",10$%0&"!)4!,$#1
(#9)2&! 7,&/,0! &/)"%! +%1,)0"5! ! P#"%8! )0
!
!&'.-/$"0+%!"#$%&'()$"
*$+%,-
!
!
!
4-5(+'3
! !
!"#$%&'()$"!
!&'.-2
!&'.-2
!&'.-2
*-3-,()$"9
:'")8+3'()$"9
;<=3',-&-"(
!
!
Parallel Coords Plot
Four Design Patterns
Analysis of trends in the data
Look at each group, find the coherent
wholes - these are design patterns
Using the patterns
Symbolic Sculptural
Display
of the web site are in comparison, which may well stimulate competition among
t groups to gain popularity for their web pages. Our small study was not suited
ve significant results on such effects but we feel confident to propose comparaeb site awareness as a group motivation tool, for example for groups editing
onic journals or other highly dynamic web sites.
User Controlled Ambient Links
Design Pattern Members
Controlled Ambient Media. As another aspect of matching virtual information
mbient media we investigated the concept of user controlled ambient media,
eveloped end user tools for management of ambient links. To require users to
about usage of ambient media is quite in contrast to the philosophy that users
ot be aware at all of the technology that blends into their surroundings. Yet, to
mbient media at one’s disposal to create highly individualized ambient spaces
very intriguing. Also given that the decision on choice of ambient media is far
understood but clearly depends on users’
and preferences (cf. Wisneski et al, 1998),
y make sense to defer the decision from
time to run time.
ure 3: Patterns and Table Fountain
ritical issue in giving the user control over
nt media is the effort required for creamodification and deletion of ambient links.
scribe two tools that we developed for managing ambient links, both concerned
rovision of easy-to-use mechanisms to link virtual information to ambient mehe first tool is for relating web page access to ambient presentations, and the
d one extends an email filter for ambient notification. Both examples uses a
able fountain as ambient medium.
Fountain as Ambient Medium. A simple but nevertheless useful and pleasant
nt media is a small fountain placed in our office. It has four pumps that can be
ed on and off independently (see figure 3). A PC controls the power plugs over
rallel interface board. The same interface can be used to integrate other devices
e controlled by switching on and off the power, for instance we have also used
onic humidifiers in other demonstrators. The media controller for the fountain is
-program running on the control PC, and generating control information send to
intensities, colors, and pulses and agreed with each other
upon private meanings to their interaction. One user used
a pulsing green and blue series to communicate “good
luck”. A flashing red meant, “Good bye, I’m busy now.”
An observer commented that they would like to give
LumiTouch to a sibling who is currently in the hospital.
People who are unable to actively communicate for long
periods of time (e.g. sick or elderly) might be able to use
the passive transmission of LumiTouch. Similarly users
who lack the required dexterity or concentration for
pushing numerous buttons might appreciate this system
due to its small number of simple grasping inputs.
Figure 3: (Clockwise) A LumiTouch when off, a user working with
LumiTouch in passive mode, the feedback panel on top lights up when
sending a message, the display lights up when receiving a message, a
user interacting with LumiTouch.
Passive Communication
When a user is in front of her LumiTouch, the
corresponding LumiTouch will emit an ambient glow to
indicate her remote presence. This additional background
information helps people figure out if there is a recipient
on the other end, or also if it is a convenient time to
increase interaction levels (e.g. start active communication
mode or use an alternative, like a phone).
Active Communication
When a user picks up the picture frame and squeezes, the
feedback display area illuminates to show that the picture
frame has been squeezed. The display colors are
transmitted over the Internet to the corresponding remote
LumiTouch. The display varies depending on the squeeze
attributes (where, how hard, and how long the user
squeezed). While the recipient party can simply enjoy this
display, s/he has the option to pick up the frame and
squeeze back a response to the first user, and begin an
interactive exchange.
Interpersonal Language
The system was intended to allow users to develop an
abstract form of emotional language. People could
communicate in real-time by sending each other color
mixtures and light patterns. The combination of colors and
force allowed a grammar, while the duration of squeeze
provided syntax for creative interpersonal dialect between
two people.
PRELIMINARY OBSERVATIONS
Preliminary testing was done with users who were not
emotionally involved with each other. Figure 3 illustrates
some user interactions. More than one user noted that they
would find having a personal LumiTouch useful.
The generation of an abstracted language was observed in
many cases. Users created different combinations of light
Another observer noted that a potential problem might
arise from the asymmetry of participation. One person is in
active interaction while the other person passively
perceives ambient data. The transmitting person may be
anxious and expect that the receiver respond. However, if
someone is so anxious that they feel the need to
communicate immediately, they can just use the phone, or
other higher-bandwidth alternative.
CONCLUSION AND FUTURE WORK
In closing, we have described our design of an emotional
communication device that can transition between passive
and active mode seamlessly. We have also outlined some
preliminary observations of the usability of LumiTouch. It
appears the dual level of attention this device supports
appeals to people. Eventually we hope LumiTouch will aid
in explorations of the development of interpersonal
emotional language. We hope to gather more data by
adding support for the manipulation of real-time data to
the LumiTouch. Users could record and replay the touch
history, like a voice mail message. Maybe the history of
touch “conversations” and patterns can be documented.
ACKNOWLEDGMENTS
We thank MIT Media Lab for their support on this project.
We also thank our loved ones for their inspiration.
REFERENCES
1. Tollmar, K., Junestrand, S., and Torgny, O., Virtually
Living Together.DIS’00 (Brooklyn NY, August 2000),
ACM Press, 83-91.
2. Strong, R., and W. Gaver. Feather, Scent, and Shaker:
Supporting Simple Intimacy. Proceedings of CSCW '96
(Boston MA, November 1996), ACM Press, 29-30.
3. Dodge, C. The Bed: A Medium for Intimate
Communication. Extended Abstracts of CHI’97, ACM
Press, 371-372.
4. Brave, S. and Dahley, A., inTouch: A Medium for
Haptic Interpersonal Communication, Extended
Abstracts of CHI ‘97, ACM Press, 363-364.
5. Fogg, B.J, Cutler, L.D., Arnold, P. and Eisbach, C.,
HandJive: a device for interpersonal haptic
entertainment, CHI ’98, ACM Press, 57-64.
Multiple Information
Consolidator
Design Pattern Members
7,((!&/%0!.)0",8%+!#!",$'(%!&%./0,I2%!4)+!4#3+,.#&,01!,$#1%"!
&/#&! #+%! (,>%(9! &)! 3%! +%'+%"%0&#&,*%! )4! ,04)+$#&,)0! 1,*%0! ,0!
)0(9!&%;&2#(!4)+$5!!J/,"!,"!4)(()7%8!39!#!8,".2"",)0!)4!&/%!
%;&%0",3(%! )'&,$,F#&,)0! '+).%""! &/#&! 4)+$"! &/%! .)+%! )4! &/%!
H#08,0">9!"9"&%$5!!K%;&-!#!8,".2"",)0!)4!*,"2#(!'+)'%+&,%"!
7,((! 3%! '+)*,8%8-! 4)(()7%8! 39! #! 8%&#,(%8! 8,".2"",)0! )4! &/%!
#"'%.&"! #.&2#((9! 2"%8! 39! &/%! "9"&%$-! #08! /)7! &/%9! #+%!
,$'(%$%0&%85! ! 6%! 7,((! &/%0! "/)7! ")$%! +%"2(&"! )4! &/%!
"9"&%$!#08!'+)*,8%!#!.)0.(2",)05!
!"#$%&'#&("!)*+,'",%'-*
!
!"#$%&'()'*+,-./&'01//,#&'2&3&%,4&5'!%1-6'
.#/0*(123456*7889*:*);<=*1>=;?2@A*"=BCD<BE*-4>?BF46GH*
!"#$%&$'&$$'()$'*+,'-./.(%#'#.01%12'
'341'54#41'6$1&.47&'43'%##'3./81$&9:!
"#$%!&%$'(#&%!)*%+!&,$%-!%#./!+%'+%"%0&,01!#!"$#((!02$3%+!
)4! $%""#1%"5! ! 6%! 7)2(8! &/%0! "()7(9! "/,4&! 3%&7%%0! &/%!
.)((#1%"!&)!'+%"%0&!#!(#+1%+!3)89!)4!,04)+$#&,)05!!!
:+&,"&"! %;/,3,&! #! .#'#.,&9! 4)+! .+%#&,*,&9! #08! #%"&/%&,.!
<281$%0&! &/#&! ,"! '))+(9! 208%+"&))8! #08! 0)&! .2++%0&(9! 7%((!
"2,&%8! &)! #2&)$#&,)05! ! =#&/%+! &/#0! #&&%$'&! &)! 32,(8! +)32"&!
>0)7(%81%! )4! #%"&/%&,.! '+,0.,'(%"! ,0&)! &/%! "9"&%$-! )+! &)!
.+%#&%!#!4)+$!)4!#+&,4,.,#(!#+&,"&!?#!*%+9!./#((%01,01!&#">-!"%%!
4)+! %;#$'(%! @A-! BCD-! 7%! /#*%! ./)"%0! ,0"&%#8! #0! #''+)#./!
&/#&!#(()7"!&/%!#+&,"&!&)!%;'+%""!#%"&/%&,.!.)0.%'&"!,0!*,"2#(!
4)+$!E!#0!!"#$%"$&'($")*+!$"5!!:%"&/%&,.!&%$'(#&%"!#(()7!#0!
#+&,"&! &)! %;'+%""! )*%+#((! .)$')",&,)0#(! 4)+$-! 8)$,0#0&!
.)()+"-! #08! )&/%+! #%"&/%&,.! '+)'%+&,%"! &/#&! &/%! 4,0#(! .)((#1%!
L,12+%! M! ,((2"&+#&%"! &/%! #+./,&%.&2+%! )4! &/%! H#08,0">9!
"9"&%$5!!J9',.#(!2"%!3%1,0"!7,&/!&/%!#++,*#(!)4!,04)+$#&,)0!
&/#&!7%!$,1/&!7#0&!&)!3%!8,"'(#9%8!#$3,%0&(95!!G0!&/%!7)+>!
8%".+,3%8! /%+%-! 7%! /#*%! ./)"%0! %(%.&+)0,.! $%""#1%"! ?%515-!
%$#,(! )+! 0%7"! 1+)2'! ')"&,01"D! #"! #! +%#89! #08! 4#$,(,#+!
")2+.%! )4! ,04)+$#&,)0-! 32&! $#09! )&/%+! ")2+.%"! .)2(8! 3%!
2"%8! ,0"&%#85! ! L+)$! &/%! ,0.)$,01! ,04)+$#&,)0-! 7%! %;&+#.&!
&%;&2#(!"2$$#+,%"!"2./!#"!$%""#1%!"23<%.&!(,0%"-!#"!7%((!#"!
#09! "2,&#3(%! ,$#1%"! &/#&! $,1/&! .)$%! 7,&/! &/%! ,04)+$#&,)0!
?4)+! &/%! %;#$'(%"! ,0! &/,"! '#'%+! 7%! /#*%! ./)"%0! &)! 7)+>!
)0(9! 4+)$! &%;&!&)!,((2"&+#&%!&/%!$)+%!./#((%01,01!'+)3(%$D5!!
6/%0!,$#1%"!#+%!0)&!'+)*,8%8-!7%!2"%!&%;&2#(!,04)+$#&,)0!
7,&/! &/%! G$#1%N)0<2+%! "23"9"&%$! ?8%".+,3%8! ,0! &/%! 0%;&!
"%.&,)0D! &)! 4,08! ,$#1%"! 7,&/,0! #! (#+1%! ,08%;%8! ,$#1%!
.)((%.&,)0! &/#&! #+%! (,>%(9! &)! +%4(%.&! &/%! "%$#0&,.! .)0&%0&! )4!
&/%! $%""#1%5! ! J/%! H#08,0">9! "9"&%$! &/%0! .)$')"%"! &/%"%!
,$#1%"! )0! &/%! 3#","! )4! #0! #%"&/%&,.! &%$'(#&%! &)! .+%#&%! #!
.)((#1%5!!!
:"!8,".2""%8!,0!&/%!"%.&,)0!#4&%+!0%;&-!#0!#%"&/%&,.!&%$'(#&%!
,"!8%4,0%8!2",01!#!.)((%.&,)0!)4!(#9%+%8!,"-&./#!"2''(,%8!39!
#0! #+&,"&5! ! =%1,)0"! .#0! 8,+%.&(9! .)0&+,32&%! ,$#1%!
.)$')0%0&"! &)! &/%! 4,0#(! +%"2(&-! .#0! "'%.,49! (#9)2&! #08!
Information
Monitor Display
Design Pattern Members
High Throughput
Textual Display
group, alerting the user when new messages arrive. The Web tool summarizes headers, lists, and
hypertext links on a Web page, allowing the user to monitor news wires and hotlists. The weather tool
monitors the weather conditions and forecast for a given city. When users initially start Irwin, it tries to
determine their informational interests by examining their email files and Web bookmarks. Since users
are not required to answer a list of questions at startup, they are often more likely to use Irwin. The user
can then configure both the information monitored and the way in which it is displayed after getting
used to the power and flexibility of Irwin.
To show both an overview and details simultaneously, Irwin uses a confluent zoom display. A confluent
zoom provides multiple views of a resource simultaneously with each view having a different level of
granularity. With a confluent zoom, the user can see overview and details with minimal effort, and the
context of the detailed views is maintained at all times by the coarser views. Since intermediate views
have proven useful in graphical tools such as map browsers and medical imagery [11, 12], we expect
that the intermediate views will smooth the transition from the high-level overview to the detailed zoom
view.
Design Pattern Members
In its confluent zoom, Irwin employs an auditory cue, an
icons view, a navigation bar, and several textual views. In
the icons view (the leftmost view), the state of each
resource is given by an icon. When the resource changes,
the appearance of the icon changes and an auditory cue is
played. If user selects an icon, the other views are updated
to show information about the corresponding resource. A
navigation bar (second from the left) shows a syntactic
encoding of the resource contents and can be used to select the messages displayed in the textual views
(the two rightmost views). Users can configure the orientation and placement of the icons with respect to
the other views and can even choose to hide the other views until some event happens, e.g. an icon is
clicked.
Auditory cues
Auditory cues provide a means to attract users to a change that has occurred. They may be the only way
the user knows about the change if the display is obscured or not in the user’s current line of sight.
We selected sounds that are distinctive but are not distracting and, most importantly, are short. Each is
less than two seconds in duration, which we expect will be long enough to be noticed but not too long to
become an annoyance. In our initial configuration, we tried to select sounds that were indicative of the
resource; for example, a dog bark for mail (dogs always bark at the mail carrier) and a splat for news
(similar to a newspaper hitting a front porch). Of course, the user can select from a list of sounds (we
provide twenty) or create and use their own sounds.
Icons
Icons are used in interfaces because they provide a universal representation in a small amount of space.
In a WIMP environment they can both provide information and invite the user to click to obtain more
information.
Further Discussion
Using the patterns
Growth areas in Ambient Information
Systems
Information Capacity - getting away from nuggets toward convolution?
Notification Level - Ambient Intelligence agenda
Aesthetic Emphasis - True “ambient art” systems
Future Work
Building/analyzing more systems
Extending the taxonomy dimensions
Social versus personal
User-configurability
How does the experience change habits?
Thanks! Questions?
Zach Pousman, John Stasko
Information Interfaces Laboratory
Georgia Institute of Technology
Acknowledgments: Funding via NSF 11S-0118685. Discussions with Chris
Plaue (and the II Group generally), Niklas Elmquist, and Tara Matthews.