Storytelling with Digital Photographs

Marko Balabanović Lonny L. Chu Gregory J. Wolff

[email protected] [email protected] [email protected]
Ricoh Silicon Valley
2882 Sand Hill Road, Suite 115
Menlo Park, CA 94025 USA
+1 650 496 5700
ABSTRACT handing around a photo album while the photographer tells
Photographs play a central role in many types of informal us the story behind the pictures, or receiving a couple of
storytelling. This paper describes an easy-to-use device that photographs in the mail with a short note from a friend or
enables digital photos to be used in a manner similar to family member. In developing our prototype “StoryTrack”
print photos for sharing personal stories. A portable form device, we imagined two specific scenarios which guided
factor combined with a novel interface supports local the design and embody both the imposed constraints and
sharing like a conventional photo album as well as assumptions about user needs and priorities.
recording of stories that can be sent to distant friends and • Ben asks Fred about his recent trip to Alaska. Fred
relatives. User tests validate the design and reveal that fetches his box of pictures and pulls out the most recent
people alternate between “photo-driven” and “story-driven” batch of photos. He flips through the pictures one by
strategies when telling stories about their photos. one, and relates interesting anecdotes associated with
Keywords some of the photos. Occasionally Ben points at an
Digital storytelling, multimedia organization, digital image and asks about it. At one point Fred gets
photography, browsing. sidetracked and launches into a story about last year’s
trip to Canada. He looks back through his other photos
INTRODUCTION
to find a picture that illustrates the point.
One of the most common and enjoyable uses for
photographs is to share stories about experiences, travels, • Fred’s children Amy and Johnny want to send some
friends and family [2]. Almost everyone has experience pictures to Grandma. They pick out some pictures of
with this form of storytelling, which ranges from the themselves playing from the recent photos. While
exchange of personal reminiscences to family and cultural doing so they come across a funny picture of Dad in
histories. The World-Wide Web can facilitate the sharing of Alaska and add that in as well. In addition to creating a
such stories in digital form, and has inspired a movement brief message to go along with their photos, they spend
towards “digital storytelling.” For instance, Bubbe’s Back a lot of time recalling (and arguing) about what
Porch [3] is a collection of family stories expressed as Web happened in each picture.
pages containing text and photographs, captured during a These two scenarios illustrate the two basic modes of
series of workshops—a grandmother’s conversation as she interaction we address:
makes soup, a grandfather’s tale of how he met his wife. On
1. Sharing of stories locally, that is, with at least one other
a grander scale, national institutions such as the US Library
person present and viewing the same photos;
of Congress store oral histories from, for instance, migrant
farm workers in 1940s California [9]. 2. Sharing of stories remotely, that is, by sending
someone a set of photos along with some commentary.
The goal of our project is to support the sharing of digital
photographs and associated stories. The design objectives Existing tools do not naturally support these activities. As
were motivated by both formal research [2,7] and informal explained below, current form factors and user interfaces
observations and interviews regarding photo usage. This present barriers to sharing digital photographs. To
background research revealed the need for a means of overcome these limitations, we set out to build a prototype
relating to digital photos as one typically relates to print device that is easy to hold and pass around like a regular
photos. For example, all of us have had the experience of photo album, with an interface that requires no training to
select, display and comment on a sequence of photos.
Current Approaches
With printed photographs, people spontaneously generate
To appear in CHI 2000 Conference on Human Factors in stories as they view pictures and interact with one another
Computing Systems, April 2000, The Hague, The around the kitchen table, living room couch, or other
Netherlands. communal space. However, digital photographs are

1
generally viewed on personal computers that do not authored stories then correspond to folders of bookmarks
facilitate shared interactions. People can only create digital that a user creates, representing interests that change over
stories with special-purpose software requiring computer time.
skills. (For our purposes, a story is an ordered set of photos The remainder of this paper discusses the application of the
with an accompanying audio narration.) Most digital photo story metaphor to digital photos. The following section
album software (e.g., [10]) explicitly distinguishes between describes our prototype device and explains how it attempts
authoring and viewing stories. Hypermedia composition to meet the stated requirements. Next, the results of
tools such as Isis [6] or MediaDesc [1] also support story informal usability tests are presented, demonstrating the
creation, but require the user to focus on structures such as differences between anticipated and actual use of the
temporal constraints and hyperlinks rather than just telling a device. The conclusion offers several suggestions for
story. In contrast, our design does not distinguish between developers of digital photo albums and story sharing
authoring and viewing modes thereby allowing people to devices.
view existing stories while simultaneously creating a new
one. DESIGN OF THE DEVICE
Our design can be broken down into two components: the
Our work also differs in some respects from research on
physical form of the device and users’ interactions with it.
image retrieval and organization [5]. Digital photos have an
We consider each in turn, and follow with a brief discussion
advantage over print photos in that users can search for and
of the prototype implementation.
retrieve them both by their content (e.g., features such as
color and texture) and by their metadata (e.g., user-supplied Form Factor
text annotations). Much previous work has focused on such Sharing photos in a natural setting requires a portable
searching as a key aspect of working with digital photos. device that can be used in different locations throughout
For example, the FotoFile system [7] provides a unified one’s home. The device should also be large enough to
interface for annotation and search, using categories such as show photos at a size similar to regular prints, viewable by
people, places and events that are commonly used for more than one person. Initial investigation with mocked-up
labeling photographs. With prints, however, people are interfaces led to a design for two-handed usage that allows
generally limited to very simple search strategies—scanning the device to be held easily, rested on one’s lap or a table,
through chronologically ordered images in an album or or shared with another person. Figure 1 shows our current
batch of photos. Our prototype derives great simplicity by prototype being shared by two people.
operating on very similar principles without any Note that a typical personal computer cannot easily be
sophisticated retrieval mechanisms. shared. One user typically controls both the mouse and
Stories as an Organizational Metaphor keyboard and has the best view of the monitor. In contrast,
Stories form an intriguing organizational metaphor, sharing photos requires that control pass easily from one
blending aspects of chronological orderings and user- user to another.
created groupings such as folders or directories in a file These design constraints led us to avoid using a touchscreen
system. There are two kinds of stories represented in the or pointing device since it is easier for two users to share a
StoryTrack: device when there is not a single locus of control.
1. Imported stories are the “rolls of film” with which a Furthermore, people point at pictures when talking about
user starts. In the case of scanned prints, they might
correspond to literal rolls of film. However, in the case
of digital photos, they correspond to a set of photos
downloaded from the camera in one session. Within an
imported story, photos are ordered by time of creation.
2. Authored stories are selections of photos that have
been grouped and ordered by the user.
The sets of imported and authored stories can themselves be
ordered according to the time of creation of the stories. At
any time, the story under construction has special status.
When complete, it joins the set of authored stories. A photo
appears exactly once in the set of imported stories, and can
appear in zero or more authored stories.
Objects from other domains could equally be organized in
this way. Imagine a Web browser that represents a user’s
browsing history as one or more imported stories,
consisting of chronologically ordered Web pages. The
Figure 1: The device in use

2
 Navigation area,
Play/Record showing thumbnails
indicator in three tracks
(currently
playing)
Current track/photo indicator

Current
narration
indicator Audio area, showing Central area, displaying
available narrations for the currently selected
current photo thumbnail in large form

Figure 2: Screen layout

them. Using the same gestures to control the device might ordered according to their time of creation; the display
be confusing and produce unexpected behavior. Instead, all again visually distinguishes separate stories using
input controls are mounted on the edges of the device. As colored backgrounds.
seen in Figure 1, this enables control to pass more fluidly 3. The bottom track represents a story in progress: the
between two users. “working set” of photos selected during the current
Interaction Design session. A photo will appear in the bottom track if it
The imagined usage scenarios presented quite a design has been added to the working set by pressing either of
challenge. At any point, a user may switch from recording the + (add) or record buttons, as detailed below.
an anecdote to viewing another set of images, may hand the The permanent display of all three tracks enables an
device off to somebody else, or may simply browse through essentially modeless interface where a user can
the latest photos. To accommodate such usage, we simultaneously view stories, see their original unedited set
developed a novel interface based on the story metaphor. of photos, and see the story they are creating. The display
Figure 2 shows the layout of the display. It is divided into also provides helpful context for viewing the current photo.
three main regions. The large central area of the screen In a typical interaction, a user comes across a photo that is
displays the currently highlighted thumbnail. The audio interesting and adds it to the working set, possibly also
area shows available audio narrations for the highlighted recording a related voice annotation. At the end of the
photo. The navigation area at the top of the screen is a session, the bottom track is grouped into a single story that
graphical representation for browsing and navigating is then appended to the middle track. We now explain the
through photos. Each of the three horizontal tracks of photo interaction in detail.
thumbnails serves a distinct purpose:
Navigation
1. The top track shows the imported stories: all existing Figure 3 labels the control buttons. The top two buttons on
photos currently stored in the device, ordered each side, easily accessible to the thumbs, scroll the current
chronologically1. Photos from digital cameras are track either to the left or the right. A bright yellow vertical
ordered according to when they were taken, while line (labeled in Figure 2) indicates the current track and the
images scanned from print photos are ordered by center thumbnail of this track. This selected photo is always
scanning time. Alternating background colors also displayed in the large central area. When one of the
distinguish separate stories (corresponding to different scroll buttons is pressed, the current track shifts to the left
rolls of film or download batches). or right. As a different thumbnail moves under the
2. The middle track shows the authored stories: photos indicator, the corresponding photo is displayed.
that have been grouped into stories by the user. Each Variable-speed scrolling allows the user to quickly traverse
story appears as a sequence of thumbnails. Stories are the photos on a track. At slow speeds, the display appears
as shown in Figure 2. Faster scrolling speeds are enabled by
1
Since meaningful timestamps for the photos used in the rendering only low-quality versions of the thumbnails
user tests were not available, the prototype as illustrated (quickly accessible from a separate index), and by not
does not display them. rendering the central or audio areas at all. An earlier

3
 Microphone
Scroll left Scroll right

Expand/collapse view Select track

of current track
Play
Add photo to
current story

Stop
Remove photo from
current story Record
Save current story

Figure 3: Prototype device showing functions of buttons

prototype used a center-weighted dial to control scrolling Pressing the record button begins recording of an audio clip
speed. The left and right buttons are better suited to the way associated with the currently selected photo in the working
the device is held, but cannot control scrolling as easily. At set. If the photo is not already in the working set, it is
present, a single “click” of a scroll button moves the track appended before recording begins (as though the + button
by exactly one thumbnail. Holding down a scroll button for were pushed first). If the selected photo is already on the
a longer time causes scrolling at increasing speed. working track, the new recording overwrites any previous
There are two remaining navigation buttons. The track recording associated with it in the story under construction.
selection button moves the indicator between the three The recording can be stopped with the stop button. If the
different tracks. As illustrated in Figure 4, the user scrolls to a new photo while recording, the recording
expand/collapse button controls which of two different for the first photo terminates and a new audio clip is started
views of a track is shown: that is associated with the new photo. Our aim is to make
recording a story as similar as possible to viewing a story.
• The expanded view, shown by default, where every However, to prevent accidental erasures, scrolling
photo is shown in thumbnail form. backwards automatically halts the recording.
• The collapsed view, where each story is represented We hypothesized that some users would compose a story by
by a single image, allowing faster navigation. The first selecting a working set of photos using the +/- buttons
thumbnail of the first photo in a story is used to and then annotating each photo in order, a “select then
represent the story. narrate” strategy. We believed other users would
Story Playback continuously record annotations while navigating and
The cluster of buttons at the bottom left controls the selecting photos. This “select while narrating” strategy is
creation and playback of stories. The play button starts also supported: when recording is active, each new photo
playback from the currently selected thumbnail. If an audio that a user views for longer than a short time interval is
narration exists, it is played through the built-in automatically added to the working set along with any
loudspeaker. Subsequently, or after a short pause if there is narration.
no audio narration, the currently selected track Hyperlinks
automatically scrolls forward to the next photo (we assume The same photo may appear in many stories, with a
left-to-right as a default viewing and storytelling direction) corresponding audio narration for each. Whenever a photo
and continues playing until the stop button is pushed or the is displayed, all of its associated narrations are displayed in
end of story is reached. If the user navigates to a different the audio area. Figure 2 shows two available narrations for
photo while playing, playback will continue from that point. the current photo. Each is marked with the time of
Story Creation recording and the name of the recording user. The length of
Pressing the + (add) button appends a copy of the currently the wavy lines is proportional to the duration of the audio.
selected thumbnail onto the working set (the bottom track). The narration associated with the currently selected story is
The – (remove) button, conversely, removes the selected listed first and is played by default when the play button is
thumbnail from the working set. pressed. Pressing the play button multiple times in quick
succession selects one of the alternate audio clips and

4
 Figure 4: Expanded and collapsed views of two stories

playback “jumps” to the corresponding story, providing a Stories and metadata about photos are stored on the tablet’s
kind of automatic hyperlinking between stories. hard drive in Extensible Markup Language (XML)2. This
Saving Stories
allows for easy translation to Hypertext Markup Language
The final group of buttons, at the bottom right of Figure 3, (HTML) or Synchronized Multimedia Integration Language
controls story operations. The save button “saves” the (SMIL) so stories can be shared with others who do not
current working story (on the bottom track) by moving it to have a StoryTrack. The SMIL format is especially
the end of the middle track. Note that the current state of appropriate as it allows the synchronization of audio with a
the device is also saved to disk at that time, matching users’ series of images, exactly matching the structure of our
expectations about “save” and persistent storage. stories.

The user would also have the option of electronically Digital photos can be loaded onto the device from digital
sending a completed story to another user for viewing on a cameras by inserting a flash memory card, or may be
similar device or on a regular PC via a media player downloaded through a wireless network connection.
application or standard Web browser. However, the OBSERVATIONS OF USAGE
implementation of this feature was not part of the user To evaluate the efficacy of this design for sharing digital
testing reported here. photos, we conducted an informal usage test [4,8]. Subjects
Design Summary were presented with the device in a natural setting and
This interaction design achieves our primary goals for encouraged to think aloud while using it. Observation of the
modeless sharing of digital photos and stories. A user can user interaction helped identify features of the interface that
seamlessly switch between browsing, viewing, authoring, did or did not work well. Moreover, it provided valuable
and showing. In addition, by removing the complexities of insight into the ways that people use photos to tell stories.
dragging, menu selection and other artifacts of windows- Usage Study Description
based interfaces, we arrived at something that we hoped We observed nine sessions of subjects. In seven of these
would be intuitive and simple to use in conjunction with our sessions, we observed pairs of subjects consisting of one
hardware controls. primary (P) and one secondary user (S). The remaining two
Implementation sessions involved a single primary user. For each session,
The prototype described here is based on a portable tablet the StoryTrack was preloaded with 2–3 recent sets of
computer (a Fujitsu Stylistic 2300) augmented with several photos provided by P. We intended that P treat the device
specially built control buttons. Although fairly heavy at as its owner. Eight subjects provided prints that we
3.9lbs (1.8kg), we expect future prototypes to be lighter as scanned, one subject provided photos taken with a digital
hardware technology advances. The screen viewing angle camera. The total number of photos provided by each
permits up to two users seated side by side, and the machine subject ranged from 45 to 234, with half of the subjects
provides sufficient battery life and computational resources falling in the 45-50 range (the user with a digital camera
for testing our design. Photos displayed in the central area provided the 234 photos). Note that on initial introduction
are 4.5x3” (11.4x7.6cm), about 60% of the size of standard to the device the first track included all of P’s photos; the
4x6” photo prints. second and third tracks were blank.
The software is implemented using Java 2 on Windows 98, Each session was divided into three stages: initial
with additional native libraries for audio I/O. The control exploration, sharing, and feedback.
buttons are simple pushbutton switches connected to the • In the initial exploratory stage, P was given a brief
tablet’s serial port via a BASIC Stamp 2 microcontroller. introduction to the device without explanation of the
Relative to other input devices, these buttons are very easy interface. P then played with the device for 10–15
to use, robust, and provide sufficient functionality for the minutes without any further input from the
desired interaction. For sound recording, a microphone is experimenters. At the conclusion of this stage, a written
attached to the external body of the tablet, while the built-in instruction sheet was provided detailing the functions
speakers are used for playback. of each of the control buttons as well as an overview of

2
XML, HTML and SMIL are all standards of the World-
Wide Web Consortium, available via http://www.w3.org

5
 the interface. At this point subjects were welcome to show photos to a local audience by holding the tablet in
ask any questions regarding the basic functionality of both hands while pressing the buttons. However, the weight
the device, but were not provided answers regarding of the tablet and size of the BASIC Stamp assembly (a box
usage strategies. For example, we would explain that attached to the underside of the tablet) made it difficult for
the + button added an image to the current working subjects to lift and maneuver the device comfortably during
track, but would not explain how to use the + button usage. All of the primary subjects dealt with these problems
along with the recording capabilities to create audio- by resting the tablet on their laps and rotating it as needed
annotated stories. Once P felt comfortable with the to show the other subject the photos.
device, S was brought into the room and seated By holding the tablet in this position, the subjects
adjacently. If desired, the device was reset, erasing any maintained easy access to the button controls. Many
stories created during the first stage. subjects mentioned that the buttons were conveniently
• In the second stage, lasting 20–30 minutes, P was placed so that it was easy to simultaneously hold and
assigned two tasks. First, to share photos locally with control the device. In several of the multiple-subject
S. Second, to create a story that could be sent sessions, one manipulated the controls along the right edge
electronically to any specified recipient (although note of the device while the other manipulated the controls along
that for these tests no stories were actually sent). In the the left edge. Occasionally, the device was handed over to
two single-subject sessions, the local sharing task was the other subject. Subjects frequently used one hand to
skipped. point to images on the screen. This is similar to a way of
browsing through print photo albums—one person holds
• The third stage consisted of a series of follow-up
one side of the album, another person holds the other side
questions directed at both subjects, broken down into
and both of them can point to photos.
three discussion topics. First, comments on the
interface (e.g., describe usage difficulties, confusing Subjects had no difficulty operating the control buttons
aspects of the interface, overall likes and dislikes about although very few subjects discovered and used the fast
StoryTrack). Second, comparisons of the StoryTrack scroll capability. Rather than hold down the scroll button,
experience with the users’ current handling of print or subjects would press it multiple times quickly to advance
digital photos (e.g., methods of organization, methods through a track. Since many subjects asked for a fast
of sharing). Finally, feedback regarding the storytelling method of advancing through the tracks (especially
aspects of StoryTrack. returning to the beginning of a track) we suspect that an
alternative input device, such as a thumb wheel or pressure-
Two of the tests were conducted in subjects’ homes, the
sensitive pad, may be more effective at facilitating variable-
remainder in an office library. With comfortable seats and
speed scrolling. The buttons used in this study give a
no office furniture, the library somewhat resembles a home
satisfying tactile click when pressed, and so create the
environment. All sessions were videotaped.
expectation of a discrete rather than a continuous action.
Usage Study Results
Software Interaction
Overall we were pleased to find that subjects could operate
As already explained, subjects had few difficulties with
the prototype with little or no instruction. Eight of the nine
basic navigation and the saving of stories. The
primary users were able to create sequences of photos and
experimenters did explain the save function to one subject
"save" them onto the second track during the initial
and the operation of the recording function to two subjects.
exploratory stage. At the end of the first stage subjects
All other subjects learned to operate the basic functions of
asked few questions and spent very little time, less than 30
the device during the exploratory phase without
seconds on average, reviewing the instruction sheet.
experimenter help. The expand/collapse feature (Figure 4)
Primary users had no difficulty demonstrating the device to
occasionally caused confusion, due in part to the alternating
the secondary users without experimenter help. Everyone
background colors. Upon first use, some subjects thought
intuitively understood the use of the three tracks although
most of their photos had disappeared. However, the
they did not necessarily label the groupings as “stories.”
meaning usually became clear after pressing the button
Subjects saved 2–8 stories each. Some of these were
again (to expand the photos) and several subjects achieved
unintentional, occurring during the exploration stage, and
faster navigation using the expand/collapse function.
some did not include audio recording. Most of the stories
ranged in length from 3–7 photos. We examined these The record function also caused confusion for several
saved stories along with unsaved browsing behavior in our subjects. A number were surprised to hear their own voices
analysis. while experimenting with the play button after having
inadvertently recorded some narration. Subjects uniformly
Physical Interaction and Form Factor
favored the “select then narrate” strategy, preferring to add
The primary complaints about the physical nature of the
photos using the + button and explicitly start and stop audio
device referred to its weight and bulkiness. Our prediction
recording. In general, few subjects used audio recording
during the design of StoryTrack was that subjects would
except when creating a story to be sent.

6
Several subjects requested additional functions for editing Local sharing Remote sharing
stories. Users wanted to insert photos into the middle of Photo- Use mainly top track Use mainly top track
stories, and to bring saved stories back down to the bottom driven
track for editing. Scroll forward and Scroll forward and
comment on each add interesting photos
Sharing and Storytelling photo to working set
During the two-person tests many stories about photos were
told. As observed in [2], it is socially inappropriate to be Lots of conversation Record annotation
silent while showing your photos to someone. There but little or no audio after all photos
appeared to be two different styles of storytelling: recording selected
1. Photo-driven—the subject explains every photo in Working set used to
turn, the story prompted by the existing sequence of reorder photos
pictures. Narration often comprises a sequence of chronologically
sentences of the form “This is my wife,” “This is my Story- Usually prompted by Search for
parents at home,” etc. This corresponds to the well- driven a particular photo particular/related
documented use of picture-taking to preserve memory Search for other photos and add to
and aid recall [2]. photos to illustrate working set
2. Story-driven—the subject has a particular story in mind anecdote Narrate story after
(e.g., “Zachary’s first camping trip”), then gathers the Related photos added selecting all photos
appropriate photos and recounts the story. to working set
Rather than sticking to one or another of these styles,
subjects would segue from one to the other. A familiar Table 1: Observed interaction types
photo would remind them of a particular story (shifting
seemed to enjoy using the device and did not ask for or
from photo-driven to story-driven), or in the midst of telling
mention the lack of search or other retrieval tools.
a story an unexpected photo would come up (shifting from
story-driven to photo-driven). For example, one subject was Audio clearly plays a big role in sharing photos. All
creating a story about a camping trip until she came across subjects talked a great deal while showing photos to their
a Thanksgiving photo. This photo received a brief partners in the study. Subjects did not record these
commentary before she moved on to creating a new story conversations, nor did they record audio annotations for
about a musical performance. their own use. However, they did record narrations for
sequences of photos to be sent to a friend or family member
Table 1 summarizes the characteristics of each strategy. All
(“Hi Laura, here are some photos…”). It seems quite likely
subjects started with a photo-driven style when showing
that the use of audio may change with experience as users
photos to a local audience and were more likely to use
become accustomed to multimodal albums. Indeed, our
story-driven strategies when assembling photos to send to a
youngest tester (age 6) had a very different way of
remote audience. Note that there was often not a clear
composing a story. Favorite pictures were added multiple
distinction between photo driven and story driven usage.
times, and the voice annotation consisted of sound effects
People normally expect photos to be in chronological order
such as “Splash!” and “Neee-haw!”
and often explain the sequence of events as they go through
each photo. At least one subject created a story for local Comparison to Current Behaviors
sharing that was simply a re-ordering of the photos in Subjects’ strategies for managing their current photo
correct chronological order. collections influenced their use and perception of
StoryTrack. All subjects reported using one or more of
When sending photos to a remote friend or family member,
three different organizational tools:
subjects first selected a set of photos, using a combination
of photo-driven and story-driven browsing, then recorded • The “shoebox:” A disorganized container for all sets of
an accompanying audio narration. photos, possibly in approximate chronological order.
These observations shed some light on the relative merits of Apart from actual shoeboxes, closets and desk drawers
browsing and search. If subjects cannot think of a story to were commonly cited containers.
tell (or a photo to search for) until prompted by another • The album: A carefully selected and ordered set of
photo, then browsing has to precede search. Furthermore, photos, presented in an album.
since subjects often move between photo-driven and story- • The Web site: Subjects with digital cameras or
driven styles, it is important to support both without context scanners select a small number of photos at regular
switching in the interface. Whether these styles were the intervals to post on a personal Web site, in order to
preferred strategies or due to characteristics of the share them with friends and relatives.
prototype remains to be determined. However, subjects

7
Approximately one third of the subjects use the shoebox • Browsing linear structures allows easily understood
exclusively, one third use a combination of the shoebox and controls and interaction, providing sufficient
the album, and one third use a combination of the album for functionality for at least hundreds of images;
print photos and the Web site for digital photos.
• Users constantly mix creating, viewing and telling
In many cases, the albums and Web sites include short text stories. Modeless interfaces that simultaneously
annotations describing the photos. This is the conventional support these activities should be preferred.
way to record stories with personal photos. It may also help
ACKNOWLEDGMENTS
explain why subjects showed a preference for “select then
We thank all of the testers, and also Zachary Phillips (age 6
narrate” over “select while narrating,” since the process of
months), whose likeness is featured in all of the figures.
annotating a print album or Web site typically occurs after
the photos have been selected. REFERENCES
Different advantages were cited for the StoryTrack 1. Caloini, A., Taguchi, D., Yanoo, K. and Tanaka, E.
depending on the subject’s current organizational methods: Script-free scenario authoring in MediaDesc, in
Proceedings of ACM Multimedia '98 (Bristol, UK,
• “Shoebox” subjects liked the idea that all of their 1998), ACM Press, 273-278.
photos would be easily browsable and all in one place;
2. Chalfen, R. Snapshot versions of life. Bowling Green
• Album creators liked the fact that a photo could be in State University Press, Bowling Green OH, 1987.
more than one story;
3. Don, A. Bubbe’s Back Porch. Available at
• Subjects with digital photos and Web sites liked the http://www.bubbe.com
fact that now they could share these pictures without
4. Gomoll, K. Some techniques for observing users. In
having to sit around the computer screen, which was
Laurel, B., editor, The Art of Human-Computer
not seen as a sociable activity.
Interface Design, Addison-Wesley, Reading MA,
CONCLUSIONS 1990, 85-90.
The StoryTrack device demonstrates that digital photos can 5. Idris, F. and Panchanathan, S. Review of image and
be used to support some of the same kinds of story sharing video indexing techniques. Journal of Visual
that people enjoy with print photos. It also provides a Communication and Image Representation 8, June
convenient way of recording stories and sending them to 1997, 146–166.
family and friends, much more easily than is possible with
conventional albums or tools. 6. Kim, M.Y. Creative multimedia for children: Isis story
builder, in Proceedings of CHI '95 (Denver CO, May
The novel "three track" interface enables a very clean 1995), ACM Press, 37-38.
design that was easy to use for all of our test subjects
(including the 6-year-old). In less than 15 minutes of using 7. Kuchinsky, A., Pering, C., Creech, M.L., Freeze, D.,
the device, people very naturally mixed browsing, Serra, B. and Gwizdka, J. FotoFile: A consumer
composition, and annotation of photos while seamlessly multimedia organization and retrieval system, in
switching between "photo driven" and "story driven" Proceedings of CHI '99 (Pittsburgh PA, May 1999),
strategies. It remains to be seen whether these simple ACM Press, 496-503
navigational tools and story metaphors will suffice for very 8. Nielsen, J. Guerilla HCI: Using discount usability
large collections of photos. Looking ahead, we are curious engineering to penetrate the intimidation barrier. In
whether access to this kind of device would alter the Bias, R.G. and Mayhew, D.J., editors, Cost-Justifying
quantity or types of photos people take. Usability, Academic Press, Boston MA, 1994, 245-272
The lessons learned from the design and testing of the 9. Todd, C.L., and Sonkin, R. Voices from the dust bowl.
StoryTrack prototype are summed up with the following Available at http://memory.loc.gov/ammem/afctshtml/
observations for developers of digital photo albums and tshome.html
story sharing devices: 10. Yagawa, Y., Iwai, N., Yanagi, K., Kojima, K. and
• A shareable device with shareable controls facilitates Matsumoto, K. The Digital Album: A Personal File-
spontaneous interpersonal interaction; tainment System. In Proceedings of MULTIMEDIA ’96
(Hiroshima, Japan, June 1996), IEEE Computer
Society Press, 433-439.