﻿Swarm: Hyper Awareness, Micro Coordination, and Smart Convergence
through Mobile Group Text Messaging
Shelly Farnham
Microsoft Research
One Microsoft Way, Redmond WA 98052
shellyf@microsoft.com
Abstract
Text messaging has become an integral part of
mobile communication, with a significant impact on
cultural and social norms in many countries. Our
goal is to explore how group-based text messaging
enables continuous social awareness, group
coordination and smart convergence on social events.
We implemented a group text messaging system,
Swarm, and deployed it to several communication
groups for ten months. Through usage logs,
questionnaires, text analysis, and direct observation
we examined how people integrate group text
messaging into their day to day social lives. Swarm
was used primarily for lightweight distribution of
context information (availability, location, and event
status) in order to facilitate social convergence. We
discuss the lessons learned from our deployment, and
the implications for the design of mobile, group
communication systems.
1. Introduction
The ubiquity of mobile phones has created a
social-technical environment where anytime,
anywhere communication is a way of life for much of
the world’s population. Mobile phone subscriptions
have reached over 90% penetration in many countries
and in a few extreme cases 100% penetration [23].
Mobile text messaging, known as Short Message
Service (SMS), has played a vital role in the evolution
of interpersonal communication, impacting
relationships, politics, health care, and even religious
practices in much of Europe, Asia and Africa
[12,10,21,23]. Understanding how people use mobile
devices to communicate, and learning how to best
make interactions through mobile technology as
lightweight and unobtrusive as possible, is essential to
the ongoing evolution of mobile communication
systems.
Pedram Keyani
Carnegie Mellon University
5000 Forbes Ave, Pittsburgh, PA 15213
pkeyani@cs.cmu.edu
Our research focuses on group-based mobile
communication that enables continuous awareness
and coordination for members dispersed across
locations, referred to as hyper awareness and micro
coordination [17]. This paper describes an
experimental SMS system--Swarm--developed to
study the impact of group-based text messaging on
social interactions.
In another study [16], we experimentally studied
the use of Swarm for a social coordination gaming
task, and found that it served to help coordinate a
group of people almost as well as voice
communication. We expect however that a close
examination of Swarm’s deployment in a real world
social situation would provide a rich understanding of
how people would actually use group text messaging.
Lessons learned from the use of Swarm in a real
world social context can meaningfully inform the
design of future mobile communication systems.
As such we deployed Swarm to groups of socially
active friends to learn how in-time, in-place
coordination affects shared group activities. We
collected usage data, had participants complete
questionnaires, analyzed message content, and
directly observed the use of the technology in a real
world social context.
We found that Swarm provides a simple
communication system that has a significant impact
on people’s social experiences, allowing them to
explicitly share moment to moment contextual
information (availability, location, and status of social
situation) with groups of people without leaving their
current social context. We found that continuous
social awareness and coordination capabilities
strengthened group ties and increased the frequency
of positive social interactions.
2. Background
SMS is a lightweight text communication medium
for mobile phones, limited to 160 characters. Despite
.
the short message length of SMS and the
impoverished text input of most mobile phones, text
messaging is the dominant form of mobile
communication in much of Asia, Europe, and Africa,
and is gaining acceptance in the United States [ 23].
The low cost of SMS has made it a communication
medium of choice for young adults and it has become
an integral part of their social lives.
Studies of teens, primarily in Japan and the UK,
have shown how social practices and group identity
have become tightly coupled with mobile phones and
SMS [3, 24]. Not surprisingly, SMS is heavily used
for establishing and maintaining both friendships and
romantic relationships [3, 10, 11]. Mobile phones and
SMS foster a particularly strong sense of intimacy for
teenagers because it allows a continuous connection
to friends outside of parental control. Teenagers find
it very important to always be available to their social
network whether or not they are co-located [13].
Between friends and couples, SMS messages like
“smile :-)” are viewed as significant gifts [24].
Access to people through mobile communication
significantly connotes social affiliations. Teens are
often very aware of who is in each others’ address
books, and will remove people whom they do not like
from their friends’ contact list [13]. The
abbreviations teens use in their messages are often
used to signify what social groups they belong to [ 21].
SMS is also used to facilitate ‘smart mob’ behavior
-- that is, the enhanced ability of groups of people to
coordinate and cooperate based on digital
communication [ 21]. Smart mobs quickly mobilize to
a collective action despite being geographically
distributed.
We expected that given the importance of mobile
phones and SMS for maintaining social ties, group
identity, and social coordination, a group-based text
messaging system would prove to be a powerful tool
for social groups. SMS in its basic form lacks any
features to support hyper awareness and micro
coordination amongst an entire group of friends.
However the simplicity of text-based message
systems offers flexibility in creating new services on
top of them. Several services are built on SMS,
including www.blah.com, www.sms.ac,
www.upoc.com and www.dodgeball.com; they offer
SMS-based services like match making, tarot reading,
and location broadcasting. While these services have
over 13 million users worldwide, little is know about
their effects on group identity, social coordination and
shared activities.
3. Social Awareness and Coordination
through Group-Based Mobile Text
Messaging
A group-based mobile text messaging system
allows individuals to broadcast a text message to an
entire group of people, to which any person in the
group may reply. Compared to voice communication,
which requires immediate attention, privacy, and
quiet surroundings, text communication allows people
to communicate with minimal interruption to their day
to day life stream [ 10, 11, 20]. See [16] for a more
thorough comparison of mobile text communication
with other forms of communication.
3.1. Hyper Awareness
The term hyper awareness refers to the social
awareness that individuals share with one another by
continually staying in touch across different locations
[16]. Text messaging enables hyper awareness
because it allows people to communicate from their
current social context without breaking norms that
prohibit voice calls. Individuals often text message
one another while in public areas or in transit, such as
in Japanese subways where mobile phone calls are
prohibited [13]. Hyper awareness strengthens social
ties and creates new opportunities for social exchange
[25].
Hyper awareness has been a goal of many
ubiquitous computing systems. For example ParcTab
[22] provided location information and in/out status
for its users, and Portholes [7] proposed the use of a
shared image database to support shared awareness
and build a sense of community.
3.2 Micro Coordination and Smart
Convergence
Micro coordination is the process by which people
coordinate the time, place and details of an event
from moment to moment [17]. This fluid form of
negotiation allows groups of friends to coordinate
more spontaneously and converge more often [8, 17,
18]. This type of distributed, moment-to-moment
coordination is a fundamental application of mobile
technologies.
As mentioned earlier, voice communication is not
necessarily the most efficient or appropriate way for a
group of people to coordinate its social activities. A
person may speak to only one person at a time, both
persons must be free to talk at the same time, and one
or both parties must have the means of capturing the
conversation details that might arise (i.e. addresses,
phone numbers, etc). Group text messaging allows a
person to read and write messages to and from an
entire group of people, in noisy environments (i.e.
meetings, bars and parties), without stopping his or
her current activities. The ability of a group of people
to opportunistically update each other about their
moment-to-moment context—where they are, and
what they are doing—may not only increase the
frequency of convergence, but the quality of
convergence points. Smart convergence [16] refers to
the ability to converge on the best location by using
up to date information about different events. For
example, at a conference, a group of coworkers might
go to separate presentations, send each other reports
about the quality of the presentations, and then
converge on the best one.
Providing relevant and timely contextual
information is key to successful micro coordination
and smart convergence. One approach is to create
systems that infer the user’s context and automatically
share it with others [1, 2, 14]. Current location,
potential destinations and interrupt ability are just a
few examples of context information that can be
collected [6]. Unfortunately, many of these contextsensing
capabilities rely on technology that is not
widely available on mobile phones and require users
to explicitly configure whom they share information
with in order to allay privacy concerns [2, 14]. A
more simple approach to collecting and sharing
context is to leave it to the user. The latter approach
may provide users with enough flexibility,
expressiveness and control over context sharing to
efficiently communicate and coordinate plans.
Work by Barkhuus and Dey [1] has investigated
active versus passive approaches to context
distribution. The awareness and coordination
activities enabled are very similar to the "passive
context-awareness" systems which were preferred by
users in that study. (Note that passive in this context
refers to the user explicitly distributing information
about his or her context, rather than having the mobile
phone do so automatically.) They further recognize
that "observational methods should be used to more
accurately determine how users handle highly
interactive applications”. The observation of our
Swarm system’s use provides a greater understanding
of the role context awareness plays for social
coordination when users directly control and
distribute their context information.
4. Swarm
Swarm is a group-based text messaging system
designed to enable hyper awareness, micro
coordination and smart convergence. In Swarm,
groups are the primary mechanism of social
communication. Swarm groups are similar to email
distribution lists where each member of the group can
send messages to the rest of group, enabling
conversations. Users interact with Swarm by sending
text messages to the Swarm server’s access number.
The server receives the users’ messages and
rebroadcasts to groups in accordance with the group
information maintained by the server database.
Swarm is designed around a simple command
language which works on any SMS-enabled mobile
phone without installation. The Swarm system parses
these messages for commands and performs group
management and communication. See [16] for a more
thorough description of the Swarm system.
Broadcasting a message is done by sending
“(<group>) <message>” to the Swarm server. The
first word is the name of a particular group followed
by the content of the message. The group members
then receive messages prefixed by the user name.
Mary to soccer: do you guys want to play during
lunch today?
Creating a group in Swarm is done by issuing the
command “create <group>”. At this point the user can
add people to this group by issuing the command “add
<name> <number> to <group>”. People are removed
through the command “remove <name> from
<group>”. Users can also remove a group by issuing
the command “remove <group>”.
5. Swarm Field Deployment
To explore how Swarm might be used to aid social
coordination in a real world setting we deployed it to
one of our own friendship groups and observed its use
over ten months. See [16] for a brief discussion of its
initial use. We were particularly interested in
understanding whether it could have a meaningful
impact on people’s social activities and their sense of
connectedness. Participants were recruited to join
Swarm because they were very socially active and
already used mobile phones to coordinate their social
activities. Participants were for the most part in their
thirties, college-educated, and professionally
employed (e.g. program manager, doctor, graphic
designer). Many had significant romantic
relationships but most were without children. Most of
them had SMS already, though we had to encourage a
couple to get SMS. Many of our users required some
instruction in how to use text messaging and T9.
We initially created a group called “party”, and
added eight people to the group who consented to
participate in our study. We continued to add people
to this primary group “party” over the course of the
deployment until there were twenty five total (the
maximum size of a Swarm group). Only four people
were ever removed from the group: two on request
and two because they never contributed messages or
appeared at any of the social events. Users were
instructed that the Swarm party group was to be used
to help them coordinate their social activities. The
participating researcher instigated use of the system
by posting messages about social events to the group.
We also allowed participants, interns, and other
interested parties to start their own Swarm groups,
though we did not widely advertise this feature due to
the technical limitations of using a SMS server that
could send and receive only up to 20 messages per
minute.
In order to study our users’ experiences of Swarm,
we a) solicited feedback through brief questionnaires
at the fourteenth week and the thirty-ninth week of
our deployment, b) analyzed usage logs for the
Swarm groups, c) analyzed the text communications
of the primary Swarm group “party” for the seven
users who consented to participate in our study at its
onset, and d) directly observed as a participant its use
in its real world social context.
It should be noted that our own participation in
hosting the primary “party” group allowed us to
directly observe people’s experience of Swarm,
providing a rich understanding of its use in a real
social environment. Because participants trusted the
researcher they were willing to participate in the study
and expose their social activities in a way they would
not have if under the observation of a stranger.
However, the fact that we deployed Swarm to a
known group of people seriously limits the scope of
the conclusions we can draw from the experience. In
particular the participating researcher’s enthusiasm
for the project and the nature of the social group –
extremely social urbanites – restricts our ability to
assess how a technology like Swarm would be
adopted by other types of social groups.
5.1 Basic Usage Statistics
In the first fifteen weeks of use the primary group
“party” was reasonably active, averaging 11.8
messages per week. Each member of the group sent
0.6 messages per week, 1 message a week when
excluding people who never sent any messages
(“lurkers”). Usage was sustained over these initial 15
weeks, suggesting that users found the system
valuable beyond the initial novelty effect (See Figure
1). Over the course of the next 26 weeks a total of 27
groups were created. See Figure 2. Out of that, 15
became active, having more than one post with more
than one poster over the course of the project. Of
those active groups, each had 6.8 people who sent a
total of 7.2 messages per week, about .9 each. The
pattern of growth for these groups is very similar to
that of our initial “party” group.
Messages Per Week
30
25
20
15
10
Messages for "Party" for First 15 Weeks
5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Week
Figure 1. Message Activity for First 15 Weeks
Over the last 26 weeks of our deployment the
party group became even more active as it grew to
twenty five members, peaking at 69 messages in a
week, about 3.3 a week per non-lurker. These results
suggest that as people become accustomed to the new
form of communication and as their groups reach a
critical mass of membership, they will increasingly
use it.
As shown in Figure 3, Swarm usage for the party
group was substantially higher during the weekends.
This non-uniform pattern of usage reflects the role of
Swarm in helping this particular group coordinate
social activities.
Messages per Week
80
70
60
50
40
30
20
10
0
Messages per Day
Messages by Group for Last 26 Weeks
15 20 25 30 35 40
Figure 2. Message Activity for last 26 Weeks
6
5
4
3
2
1
0
We ek
"party"
Messages Sent by Day of Week
Mon Tues Wed Thu Fri Sat Sun
Day of the Week
Figure 3. Average number of messages sent per day
We performed an analysis of the semantic content
of the party group’s messages in order to better
understand the purpose of their Swarm messages.
That is, were people using Swarm messages to keep
in touch by engaging in social chat, or were they
using it to coordinate their face to face social
activities? We analyzed 100% of the conversational
text of seven users who consented to have their
conversations examined as a part of the study. Each
of 474 messages was coded for whether or not it
contained socially bonding chat or was intended for
social coordination.
A message was classified as intended for social
bonding if it served to increase awareness of each
other, expressed affection or humor, or shared
information without any apparent intention to
converge on a location or activity.
Hi from SF
We luv U 2 Mary! 1
how is the jagermeister hangover today
A message was classified as intended for social
coordination if the apparent desired outcome was to
converge with other people at a particular location.
mary n i going to barca. others?
An examination of the content of the text
messages shows that Swarm was used primarily for
social coordination (68% of all messages) and for
social bonding to a lesser extent (28%). See Figure 4.
However we did observe an increase in the
proportion of social bonding messages over the life
span of the group, going from 21% in the first three
months to 32% in the last three months, and 39% in
the last month. This suggests as the group reaches
critical mass and users become accustomed to each
other and to the system, they will engage in more
social chat.
We further broke down the social coordination
messages into the following categories: invitations,
commitments, reports, questions, directives. See
Figure 4.
Percent of Total Messages
80%
70%
60%
50%
40%
30%
20%
10%
0%
Goal of Swarm Messages
types of coordination
Total Coordination
Invitation
Report
Question
Directive
Commitment
Coordination Bonding
Figure 4. Percentage of Messages by Goal of Message
Invitations (50% 2 ). Most coordinating messages
were invitations in some form or another. A message
was classified as an invitation if it either explicitly
suggested getting together face to face, or implicitly
1 Note all names changes to preserve privacy.
2 Percent of total messages. These percentages do not
add up to the total percentage of coordination
messages (68%) because the subcategories are not
independent. For example, some messages would
contain both an invitation and a question.
did so by communicating one’s intended location. A
number of messages suggested an invitation without a
particular destination by simply communicating one’s
availability to socialize. These were classified as
invitations as well.
Come bid on art!
Going to coca
what up?
Reports (16%). The second most common
coordinating message was an event report. Messages
were classified as reports if they provided some
indication of the quality of a social event that would
impact the group members’ decisions to converge at
the location. Many of these messages had an
indication of the size of the crowd at the event, while
others were more descriptive, such as providing
information about the kind of music available.
Tost is packed
some avant garde music at the lo-fi. pretty mellow.
Questions (7%). People would sometimes post
coordinating questions to the Swarm group.
Messages were coded as coordinating questions if
they asked for information that would impact whether
members in a group might converge on a location.
which party better?
is there pinball at twilight?
Directive (7%). Messages were coded as
coordinating directives if they ordered a behavior or
arranged for materials that impacted an activity upon
convergence. Although we saw few such messages in
initial use, we did find that over time, people
increasingly used Swarm to send directives.
call when u git here i will get u in
Bring booze to the jungle
Commitments (4%). Messaged were coded as a
commitment messages if they were a response to an
invitation indicating the sender intended to join the
inviter. Few messages were commitment messages,
perhaps in part because they were sometimes
considered spam by the other group members.
Ill b there
A further examination of the content of the
messages indicates that an extremely important aspect
of mobile text communication is contextual
information indicating the location. 43% of all the
text messages had some form of location information
embedded in the message. Very few, however, had
explicit addresses. Location was generally indicated
by the name of a place or by the name of a person
indicating his or her house. In some cases, people
would express an invitation to converge simply by
posting a location.
zig zag we are all here
Marcus'
People also provided social context in a few of
their messages, indicating whom they were with (7%).
Interestingly, not that many messages had any
indication of time (4%), even those intended for social
coordination. A review of coordinating messages
suggests this is largely because most messages
assume the temporal context is now, and only indicate
time if it is not now.
swimming at madison park [implicitly now]
Swim 3:30
Although text messaging only allows 160
characters, we found that a lot could be
communicated with very few words for a group of
people with a shared social and environmental
context. Many people were able to converge on
social events with the simple exchange of a few
words.
5.2 Self-reported User Experience
In order to explore the more subjective experience of
Swarm’s impact on people’ social lives, we had users
complete two questionnaires. 8 users completed a
questionnaire after thirteen weeks of use, and another
9 users completed a questionnaire after 39 weeks of
use. See Tables 1 and 2. We found that people
reported liking Swarm, although it was not easy to use
at first as they adapted to text messaging. Most of our
users had very little exposure to text messaging prior
to the study, and found it difficult to learn. Despite
this difficulty, our users reported that Swarm had a
meaningful, positive impact on their lives. It made
them feel more connected to their friends, and helped
them coordinate their social activities. See Table 2.
Even after ten months of use, people reported that
Swarm had a meaningful impact on their social lives.
We summarize the lessons learned from usage
statistics, user feedback and our own observations in
the next section.
Question
Overall, I liked Swarm
Overall, I found Swarm confusing and/or difficult to use
Using Swarm made me feel more connected to my friends
Swarm helped me figure out where I wanted to go at times
I used Swarm in situations where it would have been inappropriate to
talk via cell phone
Swarm reduced the amount of time I spent coordinating social activities
through other forms of communication
I sent messages through Swarm
I read messages sent to Swarm
Table 1. Result of Swarm Survey at 14 Weeks
(1= strongly disagree and 4 = strongly agree)
Question Mean
To what extent do you like Swarm? 6.3
To what extend do you find Swarm confusing and/or difficult
to use? 2.3
Do you use Swarm to plan/coordinate your social activities? 5.0
Do you use Swarm to chat/keep up to date with your friends? 3.9
Do you feel you are spammed by Swarm too much? 3.6
Do you feel that Swarm disrupts or in other ways has a
negative impact on your life? 2.0
Do you feel that Swarm has a meaningful impact on your
social life? 5.9
Do you feel that Swarm has chagned the way you
planned/coordinated your social activities your friends? 5.1
Has Swarm helped you spend more time with your friends? 4.6
Has Swarm helped you make new friends? 2.8
Table 2. Result of Swarm Survey at 39 Weeks
(1= Not at all and 7 = Extremely So)
5.3. Lessons Learned from Deployment
Mean
3.7
2.2
3.5
3.8
3.4
3.2
2.3
4
The following section summarizes important lessons
learned from our deployment, with illustrative Swarm
text messages and comments from users solicited
through qualitative questions in our questionnaires.
Group text messaging provides a powerful tool for
social coordination. Mobile text communication
with a group of dispersed people provides a
significant advantage over other forms of group
communication such as mailing lists. Although
messages are constrained to be very short, people
have access to up to date information about the
activities of others and the nature of their social
contexts, enabling smart convergence with the
exchange of just a few words.
group SMS works way better, can socialize much
faster!
I like to see what's going on and who's doing what.
I feel more a part of/involved w/my community
it is excellent for partying. Multiple people at
multiple parties can give reports on where to go…
Group text messaging provides a strong sense of
connection – hyper awareness -- to which users
became very attached: People reported appreciating
the light-weight awareness of group activities and the
sense of interpersonal connection they got from the
system.
keeps me in touch! I see my friends more and
know what options there are
At one point two people who never sent messages or
participated in the party group’s social activities were
removed from the group. They became upset and
insisted they be added back, saying they liked that
feeling of knowing what people were up to even
though they were not that actively involved with the
group. Several people reported feeling lost during
brief periods when Swarm was down due to technical
reasons.
when it crashes, my social life reverts back to cell
a real pain in the [*]…
People particularly liked the ability to coordinate
with a group of people without leaving their social
context. People reported finding that Swarm helped
them find the best social event, and coordinate with
their friends, because they could monitor each other’s
activities and send text messages while still in their
social context.
I really liked being at a party, standing in a group
of people talking - my phone would signal a new
swarm message and rather than having to leave
the circle to answer the phone and give the caller
my undivided attention, I was able to maintain
interaction with those around me while I checked
the message, got their feedback and replied [16]
Group text messaging easily integrates into
people’s social lives. Swarm was often used in a
social context, so an increasingly large number of
people became aware of the party group. People
started asking to be added to Swarm. Two people got
SMS on their phones so they could be added to
Swarm, and one person got a cell phone. Most new
people were added to Swarm while in a social
context, such as at a party. Because Swarm became
integrated into their social lives, users were concerned
about it stopping. Since the Swarm group was
discontinued, many of its users migrated their groups
to a similar technology: www.upoc.com.
Group text messaging enables “fishing” for
company. People commonly broadcast that they
were available for socializing even without any
intended location or activity in mind. Their goal was
primarily to find company, not to achieve a particular
destination. By broadcasting their availability to the
group, they could cast a fairly wide net that would
solicit several responses from which they could then
choose how to spend the next several hours of their
lives.
Group text messaging enabled “scouting”. We
observed that members of the group would often be
aware of who was intending to go to any particular
event, and then decide their evening based on the
reports produced by these “scouts”. We even
observed on a few occasions people deliberating
parting company to scout out different events and
then sending each other reports. It became a sort of
game, where the person who found the best event won
the most social points.
Groups quickly developed conversational norms.
As discussed in [16], communication conventions
started developing in Swarm almost immediately. For
example, people would send “up?” as a short hand for
“what’s up?”. People would plan to “Swarm” in their
spoken language, saying “will you swarm me later?”
or “Let’s swarm later”. Within the party group we
noticed certain kinds of messages were discouraged
by other members of the group. If the message was
not immediately relevant or in some way important or
unusual, users felt like they were being spammed.
Aside from social coordination, the most acceptable
messages included notes from people when they were
out of town “Hi from SF!” or when a particularly
notable event occurred, e.g., “baby!” when one
member’s baby was born.
Group text messaging scenarios tended to be
oriented toward micro coordination. It is clear that
as intended the original Swarm group “party” was
oriented toward social coordination. However new
groups were created at our users’ own instigation, and
what we know about these groups can provide some
insight into the scenarios that a group text messaging
system will support. From our users’ descriptions and
our own observations, we saw the following types of
new groups form. These groups were also
predominantly geared toward coordination.
Project groups. We saw two active groups form
that were project based. The Swarm messages were
used to coordinate last minute location and time
changes and last minute collections of materials.
Activity groups. Several groups formed that were
specific to activities that were largely affected by last
minute changes in weather. One such river rafting
group was originally formed in the party group, but
became so active they started their own group called
FLOAT.
tube tube tube tube tube!
Getting rafts now. Call if you're coming
"I'm stuck @ work, too. wah"
Goin tubing who wants a ride? Got VW van.
Got enough rafts for 10 people
Anyone have paddles?
[*] a paddle. Think float.
Actually rafts for 8. Berry Mary ann honey mark me.
kim? make your reservations now
We r on the river now! Woo hoo!
@ home reading ny times. woo hoo! wish u were all
here.
Out of town groups. Several groups formed
specifically to enable coordination for out of town
events, such as going to a wedding together or a
camping trip. Similarly the one work group we
observed tended to be active only when members of
its team were attending a conference. These groups
appeared to die once the event was over.
Best buddy groups. Some of our users created
smaller groups of close friends who liked to stay in
continuous touch with each other. These Swarm
groups appeared to sustain the same private,
personalized language and jokes the members had as
a set of friends. Of all the groups we observed these
small social groups engaged in the most social chat,
perhaps because people felt their chat messages were
not spamming a larger group of less known people at
2 to 10 cents each.
Social groups are very dynamic, warranting
features for light weight group formation. Perhaps
the most challenging aspect of Swarm use was
reflecting fluctuating social relationships through a
simple text messaging command system. The groups
tended to grow in membership as people were added
from the extended social circles. However over time
a few awkward situations were encountered as
couples broke up, friends got into arguments, and so
forth. We observed our users create newer, smaller
subgroups to resolve these socially awkward
situations rather than try to remove people from the
larger group.
Group text messaging was integrated with other
forms of communication. Most of the conversations
we observed in Swarm occurred in the context of
many other conversations. For example many of the
social events discussed in the party group had already
been mentioned in group mailing lists or through
evites (evite.com) earlier that day or that week. By
the time the users were referencing these events in
their text messages they already had a shared
understanding for the reference points. We also
observed that people were often exchanging one to
one text messages even as they received group
messages, and that text messages that broadcast an
invitation would often be answered with a phone call
to the person sending out the invitation. The group
text messages served as a form of gateway to the
more one-on-one communications that required more
investment.
People need to be able to turn off their groups.
Several people thought they were getting too many
messages or that the cost became too expensive. A
few people reported they were disturbed from their
sleep by messages sent at four in the morning. While
some people liked to receive messages when they
were out of town, others reported finding it
distracting. In response to these issues, we added a
command for users to toggle their groups off.
6. Conclusions
We found that in a real world setting, people actively
used Swarm to coordinate their activities, and
converge on the best social events. While they
reported spending more time with one another and
feeling more connected, they rarely used Swarm to
chat. Rather Swarm was used primarily for
lightweight distribution of status and context
information (availability, event status, location and
destination) among friends.
Our findings show that lightweight, mobile group
text communication can significantly affect the
feeling of connection shared amongst a group of
friends, and create new opportunities to share face-toface
social activities. Additionally, our examination
of message content has shed some light on the status
and context information people choose to share when
engaged in group awareness and coordination
activities. The distribution of information about
current location, destination, quality of an event, and
the participants’ availability was key to facilitating
hyper awareness, micro coordination, and smart
convergence.
7. Future Work
The ubiquity and simple nature of SMS makes it an
ideal platform for developing social communication
tools that can impact many people. Unfortunately, one
of the pitfalls of SMS-based command languages is
the difficulty users have in remembering infrequently
used commands. The difficulty remembering
commands revealed itself through malformed
commands, misdirected messages, and people only
using the most basic functionality. To overcome this
limitation, we are planning to create a graphical,
menu-based client application to allow users to
manage their groups. This will allow us to optimize
the lengthy and less memorable commands, letting the
users exploit the full power of the system. It will also
enable us to integrate photo messaging, even further
affording the exchange of contextual information with
minimal effort. However, by building on top of SMS,
users with phones that do not support the advanced
interface features can still participate, easing the path
for deployment and adoption.
We found that users frequently provided context
information like location or destination, suggesting
that automatic location information could prove
valuable when accurate location tracking becomes
integrated with mobile phones [4, 5, 9]. We would not
argue that mobile communication systems should
always provide automated location information,
because we saw that the act of explicitly sending
location information communicates both where the
user is and her desire to socialize, often with just one
word. In addition, many of the potential automatic
methods [1, 6, 15] raise serious privacy concerns.
However, it might be valuable for a user to be able to
broadcast location information with a simple touch of
one key that expresses to an entire group of people “I
am here now, and I want to play”.
8. References
[1] Barkhuus, L., Dey, A.K. Is context-aware computing
taking control away from the user? Three levels of
interactivity examined. In Proc. of the 5th International
Symposium on Ubiquitous Computing, 2003, 149-156
[2] Barkhuus, L., Dey, A.K. Location-based services for
mobile telephony: a study of users' privacy concerns. In
Proc. of the 9th IFIP TC13 International Conference on
Human-Computer Interaction. INTERACT 2003
[3] Berg, S., Tayler, A. S., Harper, R. Mobile Phones for the
Next Generation: Device Design for Teenagers. In Proc.
ACM SIGCHI 2003, 433-440
[4] Colbert, M., A Diary Study of Rendezvousing:
Implications for Position-Aware Computing and
Communication for the General Public. In Proc. GROUP
2001, 15-23
[5] Colbert, M., A Diary Study of Rendezvousing: Group
Size, Connectivity and Time Pressure. In Proc. Mobile HCI
2002, 21-35
[6] Dey, A., K., Abowd, G., D. Towards a better
understanding of context and context-awareness. Proc. of
the CHI 2000 Workshop on The What, Who, Where, When,
and How of Context-Awareness.
[7] Dourish, P., Bly, S. 1992 Portholes: Supporting
Awareness in a Distributed WorkGroup. In Proc ACM
Conference on Human Factors in Computing Systems, 92
[8] Farnham, S., Kelly, S., U., Portnoy, W., Schwartz, J.,
L.K. Wallop: Designing Social Software for Co-Located
Social Networks. In Proc. HICSS 2004
[9] Fithian, R., Iachello, G., Moghazy, J., Pousman, Z.,
Stasko, J., The design and evaluation of a mobile locationaware
handheld event planner. In Proc. Mobile HCI 2002,
145-160.
[10] Grinter, R.E., Eldridge, M., y do tngrs luv 2 txt msg?.
In Proc. Seventh European Conference on Computer-
Supported Cooperative Work ECSCW 2001, 219-238.
[11] Grinter, R.E., Margery, E., Wan2tlk?: Everyday Text
Messaging. In Proc. ACM SIGCHI 2003, 441-448.
[12] Griswold, G. W., Shanahan, P., Brown, W. S., Boyer,
R., ActiveCampus – Experiments in Community-Oriented
Ubiquitous Computing. Technical Report CS2003-0750,
Computer Science and Engineering, UC San Diego, June
2003.
[13] Ito, M., Mobile Phones, Japanese Youth, and the Replacement
of Social Contact. Proc. Ann. Mtg. Society for
Social Studies of Science, 2001.
[14] Jiang, X., Landay, J. "Modeling Privacy Control in
Context-aware Systems using Decentralized Information
Spaces," IEEE Pervasive Computing, Vol. 1, No. 3 July-
September 2002.
[15] Kandogan, E., Dieberger, A., Dorohonceanu, B.,
Farrell S., Harrison, B., Moran, T. P., Smith B. GroupSense:
Omnipresent Awareness Information. Interactive Poster
CSCS 2002.
[16] Keyani, P., and Farnham, S. Swarm: Text messaging
Designed to Enhance Social Coordination. In Harper, R.,
Palen L., and Taylor A. (eds.) The Inside Text: Social,
Cultural, and Design Perspectives on SMS. Springer: The
Netherlands,
[17] Ling, R., B. Yttri, Nobody Sits at Home and Waits for
the Telephone to Ring: Micro and Hyper-Coordination
Through the Use of the Mobile Phone, 1999.
[18] Nardi, B.A., Whittaker, S., & Bradner, E. Interaction
and Outeraction: Instant Messaging in Action. In Proc.
ACM CSCW 2000 79-88.
[19] Nonnecke, B., Preece, J. Why Lurkers Lurk. AMCIS
Conference, Boston, June 2001.
[20] Palen, L. Mobile Telephony in a Connected Life.
Communications of the ACM. Vol 45, No.3, March 2002.
[21] Rheingold, H. Smart Mobs: The Next Revolution.
Perseus Publishing, Cambridge, MA, 2002.
[22] Schilit, B., N., Adams, N., Gold, R., Tso, M., Want, R.
The ParcTabl Mobile Computing System. In Proc. Fourth
Workshop on Workstation Operating Systems, 1993.
[23] Srivastava, L. Social and Human Considerations For a
More Mobile World. Published in ITU/MIC conference
workshop on shaping the future of mobile information
society, 2004
[24] Taylor, S. A., Harper, R. Age-old Practices in the
“New World”: A Study of Gift-Giving Between Mobile
Phone Users. In Proc. ACM SIGCHI 2002, 439-446.
[25] Woodruff, A., Aoki, P.M. Media Affordances of a
Mobile Push-To-Talk Communication Service. Computing
Research Repository (CoRR) Technical Report
http://arxiv.org/abs/cs.HC/0309001, 2003.
