Community Designers



Vicki L. O'Day(1), 
Daniel G. Bobrow(1), 
Billie Hughes(2), 
Kimberly Bobrow(1), 
Vijay Saraswat(1), 
Jo Talazus(3), 
Jim Walters(2), 
Cynde Welbes(3)


(1)
Xerox Palo Alto Research Center	
3333 Coyote Hill Rd.	
Palo Alto, CA 94304	

(2)
Phoenix College	
1202 W. Thomas Rd.	
Phoenix, AZ 85013	

(3)
Longview Elementary School		
1209 E. Indian School Rd.
Phoenix, AZ 85012

Authors' email addresses: 
{oday, bobrow, kbobrow, saraswat}@parc.xerox.com 
{bhughes, walters}@pc2.pc.maricopa.edu 
{lvtalazu, lvcochra@pc2.pc.maricopa.edu

[Presented at the Participatory Design Conference, Boston,
November 1996.]

ABSTRACT

Pueblo is a cross-generational, network-supported learning 
community developed by its own members. This 
participatory design effort has been different from many 
work-oriented systems projects and has expanded our view 
of what participatory design entails in a network community. 
The technical foundation of Pueblo is a MUD, a text-based, 
multi-user virtual world, which has been integrated into 
classroom use in a K-6 elementary school. The design 
process has been decentralized and open-ended, reflecting 
the combined efforts of a diverse group of people: 
researchers in computer science and education, elementary 
school educators and students, senior citizens, college 
students, and friends and colleagues around the Internet. As 
the community has changed, the evolving participation, 
roles, goals, expertise, and personal and professional 
relationships have played an important part in the design 
experience. The history of this community has been marked 
by increasing social maturity, with transitions from 
questions of "what can we do" to "what should we do" to 
"how should we decide what we should do". 

Keywords

Learning community, MOO, virtual world, education

INTRODUCTION

Pueblo is a learning community situated on the Internet, centered
around a Phoenix K-6 elementary school. This community is built
in a MUD, a text-based virtual reality that allows one to move
around in and experience a virtual world, extend this world by
adding new objects and places, and interact with other people who
are connected at the same time. The community consists of local
and remote participants: teachers, students, researchers, family
members, college students, senior citizens, and Internet
participants. The purpose of this community is to support
collaborative learning for participants of all ages, from
kindergarten to seniors, through innovative on-line projects and
experiences. The learning context and the affordances of network
communities provide both resources and constraints that help to
shape the design process.

This paper is a reflection on the design and evolution of this
community by some of its designers, who are also participants and
"users." In Pueblo, participants are simultaneously designers of
a community and a community of designers. Distinctions between
users, designers, and developers are blurred. However, some of
these distinctions are still visible in the participation
structures of the project.  A core design team has taken
responsibility for trying to make this environment meet the
educational and organizational goals of the sponsors. The core
group follows participatory design principles and practices and
engages the larger community in developing goals and
implementations.  The core design team and the community itself
are different entities, each with its own activities and
politics, though the participants of the core group also consider
themselves to be members of the community.

As in a real-world community, the Pueblo network community is
constituted through the rules of governance, immigration, social
services, and other pieces of infrastructure that are developed
over time. It is also strongly influenced by the institutions
that support it and the collective activities, interactions,
identities, and histories of people who live there. As in a
real-world community, the development of Pueblo has been both
planned and unplanned. The purpose of this paper is to describe
the decentralized, grass-roots process of participatory design in
a school-centered network community.

The core group's composition reflects the institutional
arrangements that have provided funding and other resources to
Pueblo - it has about 20 members from the three organizations
that support the work. This group is focused on the Pueblo
project. When the external grant that funds the project comes to
an end, some of the institutional support may disappear, though
individuals from these institutions may choose to continue their
participation.

The community is a larger and more amorphous entity. It includes
Longview students, teenagers, "grays" (senior citizen
volunteers), college students, and people from across the
Internet who have found Pueblo and asked to be part of it. Some
community members have institutional affiliations that prompted
their participation, but many do not.



In the language of Gärtner and Wagner [6], the arena for 
participation for the core group is designing work and 
systems. Here, the work includes both the teaching practice 
of Pueblo teachers and the learning practice of the students 
and other community members, and the systems include the 
technical and social mechanisms that support the network 
community. A significant part of the expected learning in 
schools goes beyond curriculum content material to include 
learning how to learn, how to interact effectively with others, 
and how to develop one's own learning opportunities. The 
"work" of learning includes play, exploration, and reflection.

In Gartner and Wagner's terms, the arena for participation of 
the community is developing frameworks for action - not 
organizational frameworks, but frameworks that will sustain 
a cross-generational learning community. For example, the 
community is exploring ways to support senior citizens in 
becoming comfortable with new technology and learning to 
mentor children in an online setting.

Crucially, while the intentional design of work and systems 
is undertaken by the core team, no central authority oversees 
and directs community design. The community literally 
designs itself. Any community member can build and 
change the landscape in which the community operates, or 
offer services to others. In this way, the network community 
provides the "design by doing" methods recommended by 
Ehn [5] for participatory design efforts. 

This informal design process varies across different social 
groups in Pueblo. Participants get on-line at different times, 
talk with different people there, and spend their time in 
different work or play activities. The substance of the 
community emerges in this flux of daily interactions, only 
some of which are directed by processes that are understood 
and directed by the design group. In this sense, our 
community is both emergent and designed. As the designers 
of Habitat point out, centralized control is neither possible 
nor desirable in virtual worlds [8].

Pueblo's hybrid nature as a network community growing out 
of pre-existing organizations adds complexity to design and 
decision-making processes. Members of the core group have 
in common a vision of educational change, but they are 
accountable to different professional standards and face 
different organizational expectations. The Longview 
principal is ultimately accountable to the district 
superintendent and school board for the education of 
Longview students. The Phoenix College faculty members 
are accountable to their dean for the use of college resources 
to meet the needs of people in their college's service area. 
Xerox PARC researchers are accountable to the funding 
agency that has made their involvement in Pueblo possible, 
as well as to Xerox management and the general research 
community. These institutions represent the third arena for 
participation described by Gärtner and Wagner, the 
framework for "industrial relations that define the norms 
for... work-related issues." [6]

Education as a Context for Design

The education focus makes Pueblo somewhat different from 
the geographically-centered "community networks" 
described by Schuler [11], which have a broad objective of 
bringing members into closer contact with one another. 
Pueblo is motivated by a particular agenda of educational 
change, with a constructivist, learner-centered focus. 

Pueblo has its roots in MariMUSE, a network community 
founded by education faculty at Phoenix College in 1993. 
Longview elementary school students and teachers were 
brought in a year later, focusing on the affordances of the 
MUD for improving literacy. In 1995, researchers from 
Xerox PARC joined the community and helped upgrade the 
infrastructure to a more technically-capable MUD. Pueblo, 
the new MUD, has more extensive integration in the 
curriculum, with a special focus on modeling to support 
science education. 

Sustainability of the Pueblo community is a significant 
concern of the Pueblo project. Sustainability is an important 
criterion of success for innovations in education. For the 
Pueblo community to be sustainable, it must serve the needs 
of its participants. Among participants' needs are the ability 
to take part in discussions of community issues, influence 
decision-making, and take independent actions that are 
perceived to add value to community life. As the community 
grows, it is important to ensure that the participation 
structures are effective for new members and groups. 

The community is participatory, but not egalitarian - as in 
real-world communities, different people play different 
roles. The skills and energy people bring and the 
accountability and responsibility people carry from the 
institutions that sponsor the community cannot be separated 
from the voices people have in many decision-making 
processes. As the community changes its composition, the 
core design group has found it necessary to distinguish its 
activities from the legitimate decision-making activities and 
forums of the broader community, sometimes creating new 
forums for discussion and action. These shifts have been 
useful from a project perspective, as well as a community 
perspective, since they have strengthened the project's 
understanding of what makes the community viable and 
sustainable.

At this point, the community has adapted to internal and 
external pressures and opportunities over three years. Each 
transition has been accompanied by increasing social 
maturity and self-awareness, though our understanding is 
always limited by inexperience with the community's 
current growth stage and the difficulty of integrating the 
separate views of individual participants into a coherent 
picture of where the community is now. Like the proverbial 
elephant, a network community is many things to many 
people. The design process we will describe includes events 
and changes the authors considered significant. All of the 
authors are members of the core design team.

Throughout most of the paper we use a single authorial 
voice, but we recognize that different members of the 
community would choose to tell different stories about its 
development. At times we will use attributed comments 
from one of the authors or one of the Longview students 
when we wish to highlight a unique, personal viewpoint.

At this time, support for the project is provided by 
Longview, Phoenix College, the Osborn School District, 
Xerox PARC, and an ARPA contract for research in MUD-
based learning environments. Project money has funded 
summer camps for Longview teachers and students, and it 
has provided teachers with some school-year release time.

CHARACTERISTICS OF A NETWORK 
COMMUNITY

Pueblo is based on a MOO, which is a kind of Internet- accessible
virtual world with its own geography, characters, and objects of
all kinds ([3, 4]; or see a general description of MUD-based
communities in [10]). People interact in the MOO by typing simple
text-based commands. Participants can move from one locale to
another, talk to other people with speech and gestures,
manipulate objects they encounter in each place, and extend the
world by creating and describing new places and objects. When
people first join the community, they usually begin by creating a
character (with name, appearance, gender, and other customized
attributes) and building a home for themselves. It is a policy of
our MOO that participants are not anonymous. Though the
characters people create may be fanciful, each person also
records information about his or her real-life identity, which is
available to all.

The Ambience of a Network Community

To give a sense of what it is like to be in a network community,
we offer a few students' reflections on the topic.  While we
worked on this paper, teachers asked some of their 10 and 11 year
old students what they thought a community was and whether Pueblo
was a community. Here are some of the students' written responses
(with minor spelling and punctuation corrections for
readability):

Dice: I think a community is a group of people that help each
other out and do things together and that think of new ideas to
improve themselves. To me what makes Pueblo a community is that
people get to make new friends when they page each other.  You
get to learn things that you never knew you knew. You get to
drive cars and program things and create things you could
possibly never have in the real world.

Nefertiti: A community is a group of people who live together in
a neighborhood. Pueblo is a community because people build homes
there and live there in Virtual Reality.

BabyT: A community is when a lot of nice people come together and
make fun games for us and we write to our friends. When people
like Jim and Hobbes work with the kids that are on. And us too,
we make houses that you can go in and look, and cars that you can
ride in with us... You can do a lot like talk to people that live
in New York and people that are older than you.

Coolio: A community is people that live in a spot all together
and help each other when needed... and they do things together
and have fun.

T-boz: People working and helping each other makes Pueblo a
community... it's fun and it's like our own little world.

Several common themes run through these and other students'
responses: the impact of the physical world metaphors, especially
houses and neighborhoods; the continuity and persistence of the
environment, which leads students to talk about people "living"
in the on-line world; the idea of helping and being helped as a
pointer to community feeling; and the appeal of building, owning
things, and leaving a personal mark. These perceptions extend to
adults, including teachers and researchers. Many develop close
ties and working relationships, with a sense of Pueblo as a place
they belong.

The Dynamics of a Network Community

As people build and interact in Pueblo, they are conscious of
where they are. It is not a neutral collaboration medium, but a
place for a group of people who share common interests and values
related to education.

There is a dynamic interaction between individuals and the
community, under the influence of the educational theme.  Bandura
describes the reciprocal relationships between individual actions
and social norms [1]. He points out that people are influenced in
choosing their own actions by the anticipated and actual
reactions of other people in the community. Through these
collective actions and reactions in the social milieu, people
"create and activate" the community itself ([1], p. 344). In
Pueblo, the interpretations different individuals bring to the
educational theme sets expectations that help to develop the
voice and norms of the community.

To work toward the educational vision, the core design team
designs tools and places to support learning activities.
Examples include a brainstorming room, a reflecting pool into
which reflections can be dropped, journals, teacher utilities,
and writing tools. These are the intentional designs of the team.

However, the intentional designs interact with the interests of
community members. Individuals across the community often take
playful license in construction, such as building a lake in the
City Park. These can turn out to be interesting to others, who
sometimes adapt them in unexpected ways. (For example, a
fourth-grader added fish and fishing poles to the lake.) These
"design accidents" (from the perspective of the formal project)
are valuable sources of ideas, functionality, and enjoyment for
the community. Project designers did not consciously decide on
cotton candy or beachfront property, but interactions with
children led to the creation of many objects that combined play
and learning in compelling and engaging ways. Designing personal
cars and pets has been an especially popular activity.

"Accidental" designs have often led to new intentional
designs. An example is the trivia game, which began as a
game-show like activity in which a quiz master could collect a
group of people together in a virtual room to try their skill at
trivia questions. The quiz master determines the "correctness" of
answers to the questions, which may be unexpectedly
open-ended. The children quickly appropriated this game for
themselves, designing trivia sheets on topics of interest to
them, doing library research to find stumping questions, and
alternating between the roles of quiz director and audience. This
impromptu activity gave adults a better understanding of the
motivating effects of children being able to follow their own
agendas in the MOO.

As part of the dynamics of a network community, it is important
to note that the community intertwines the personal and
professional aspects of the lives of its participants. Ehn claims
that participatory design requires a "shared form of life", which
the network community offers to an extreme degree. The nature of
the interactions in the virtual environment has led to more
intense commitments and involvements of community members than
many of us have been accustomed to in other
projects. Relationships have an intimacy, depth, and warmth
uncommon in design projects, especially those projects in which
cross-cultural differences raise communication barriers. Design
team members seem more willing to listen to divergent points of
view when they live and communicate together on a regular basis,
over an extended period of time, informally as well as formally.

The strength of the affective engagement has negative as well as
positive aspects. Intense commitments and special relationships
are accompanied by more vulnerability and risk. Cherny recounts
some of the upheavals and emotional struggles that have taken
place in larger and older network communities in [2].

TRANSITIONS IN COMMUNITY DEVELOPMENT

In this section, we will describe transitions in the first three
years of the community's development. Some changes were
intentional; the design team has planned and implemented
significant shifts in population, priorities, and on-line
activities. Other changes came about unintentionally as people
worked in the environment together.

Overall, the project and community have developed an increased
self-awareness. The focus on the question of "what can we do" has
expanded to include the questions of "what should we do" and "how
should we decide what we should do." Gärtner and Wagner note in
other participatory design projects this same kind of shift
between the design of work and systems, the design of
organizational frameworks for action, and consideration of the
political and organizational context [6].

The Pueblo project has developed several pilot projects
integrating the MOO into classroom use, and it has begun to
create policies, workable practices, and technical support for
bringing new people in, getting and giving help, resolving
conflicts, making decisions, property rights, privacy, and
etiquette, all of which are necessary to maintain a viable
community.

1. Opening the MariMUSE Frontier

Jim Walters, Phoenix College: Once upon a long, long time ago,
three faculty from Phoenix College paid a few visits to
MicroMUSE, an early on-line community. They ran into Moulton, who
talked about informal science education and began telling tales
of users who were interacting in a community that existed in this
MUD... Moulton and someone called shkoo offered to help load the
software on a UNIX machine that we had managed to salvage from an
old demo system we had been given, and that was the beginning of
MariMUSE...

The Pueblo community is built on the grounds settled by MariMUSE
pioneers Billie Hughes, Jim Walters, and Greg Swan. They were
interested in finding effective methods and settings for
educational change and saw promise in the MicroMUSE model
[7]. MariMUSE was centered around the Maricopa community colleges
in the Phoenix area, but it drew participation from across the
Internet. The founders recruited active builders who would
develop interesting places, objects, and ambience that would
attract others.

MariMUSE gave teachers a different place to work with new
approaches - a place that was free of the years of conditioning
teachers and students have to overcome to do things differently
in a traditional classroom. Phoenix College courses in education
and computer science used MariMUSE for non-traditional
projects. A Cambridge theologian offered a long-distance course
in the New Testament.

Though MariMUSE was successful in encouraging exploration, it had
a frontier feeling in some other, less desirable ways. One
problem was that it was not clear who was in charge. One of the
central characteristics of these communities is that they enable
individuals to act independently and communicate freely across
boundaries, including those that represent traditional lines of
authority.  Unfortunately, some of the young students who had
power and influence in the community based on their technical
expertise did not have the social maturity to match. They used
their technical powers to play tricks and reduce the capabilities
of others. One turning point was the development of a robot that
roamed around and asked insistent interview questions of other
(human) characters.  When the robot's creator refused to moderate
its behavior, a few annoyed participants locked the robot away,
changing the robot's software in the process. The robot was never
the same afterward, and neither were the relationships between
some of the people in MariMUSE.

Hughes and Walters decided to make it clear that an educational
vision was the foundation of the community, and the frontier town
started to have a sense of being ruled by principles. The
founders began to provide leadership in what should be done in
MariMUSE, as distinct from what could be done. This was a
necessary step in making MariMUSE a hospitable place for
educators and students.

2. Centering on Longview Elementary School

The partnership with Longview developed because its principal's
vision for community education matched that of Hughes and
Walters.

Jo Talazus, Longview principal: I am always searching for the
ways and means to create a community school - reaching out to the
local and worldwide communities. I firmly believe that the
greater the number of successful adult relationships a child
establishes, the greater the likelihood of success for that child
in the future...Therefore, our school has business tutors, police
mentors, classroom grandparents, Hispanic attorney mentors,
Junior Achievement volunteers, Phoenix College partners, and
others. All but Junior Achievement provide one-on-one contact
between adult and child. It is the village concept.

When Hughes and Talazus met at a local education think tank
meeting, they realized that MariMUSE and Longview might be a
perfect match. Longview had a population of limited-English
speaking students from poor economic backgrounds; MariMUSE had a
world of conversations and descriptions in a motivating medium
for language use.

With the partnership came new issues. Safety was a crucial
concern of the professional educators in the environment.
Children not only needed to be safe, they needed to feel safe and
taken care of, just as they would in classrooms at school.  The
high-level vision of a learning community, as expressed in an
introductory letter to newcomers, now stressed helpfulness,
non-violence, and modeling effective ways of resolving
conflicts. Throughout the life of Pueblo, the vision of a
learning community has guided design and implementation
decisions.

The new immigrants from Longview included some very excited
students and teachers. Teachers and their students began as
novices and fellow learners together, a situation very different
from the traditional classroom. They took on roles as active
participants and designers of learning activities
on-line. Teachers found that creating personal objects such as
homes in the MOO was highly motivating for students; students
enjoyed writing and revising descriptions of their creations.

Cynde Welbes, Longview teacher: Knowing how MUSE made me feel,
watching it affect the kids that I taught, made me a very staunch
MUSE supporter.  It fired me up inside with possibilities, and I
loved to share what it made me feel, and what I had seen it do
for others... I was also very personally meshed in this medium. I
was able to get on from home, and I started to spend time in the
evening online, talking to friends, making new friends, creating.

Welbes' comments are a reminder that a network community is above
all a social world, even when its use is related to professional
or educational objectives. Most of us in the project have
developed friendships online. Being in Pueblo sometimes seems
like strolling down a busy street in a small town; there is a
feeling of walking along, chatting with people in shops or street
corners along the way. People find that they sometimes hold
personal and professional conversations with the same people at
the same time, sometimes using different communication mechanisms
to distinguish conversational themes. Because of this mix,
personal styles and agendas, as well as professional goals, have
an impact on the development of the community.

The inclusion of Longview was a positive move for
MariMUSE. Centering on a school with a particularly needy student
population gave the network community added energy and focus. It
raised the stakes for making MariMUSE a success.

3. MariMUSE to Pueblo

Vijay Saraswat, Xerox PARC: Jim, Billie, and I met at the first
MUDshop organized by Kirstie Bellman (ARPA) and her colleagues in
December 1994. At that time we at PARC had begun to be fascinated
with the possibilities of MUD-based network learning
communities. MariMUSE was beginning to reach some of its inherent
technological limits, and Jim and Billie were looking to
collaborate with "technology" partners.  It was obvious to us
that a partnership would be mutually beneficial.

A major shift occurred with the move to Pueblo in 1995. The PARC
researchers were strong new voices in the community. The
underlying server changed from MUSE to the more sophisticated MOO
server. Though the general characteristics of different servers
are the same, each one has a different set of commands for users
to learn. Another significant difference was that the existing
objects in MariMUSE were not moved to Pueblo, so participants
could start with an empty world.

There was an exciting new-world feeling in Pueblo for PARC people
and some others, with the same broad participation and playful
experimentation in building and creating that early MariMUSE
participants had experienced.  For members of MariMUSE, Pueblo
represented a loss of familiarity. While both MariMUSE and Pueblo
existed, students alternatively visited both. They went to Pueblo
to do projects together, but they went to MariMUSE to tinker with
old creations and talk with old friends. When Walters and Hughes
decided it was necessary to focus all of their energies on
Pueblo, they closed MariMUSE. For many, Walters said, it was
"like their dog died." Though by the time of this writing Pueblo
has expanded to include most of the Longview students who had
been in MariMUSE, many of the scattered Internet participants did
not make the move.

Pueblo did not open its doors wide, as MariMUSE had.  PARC
participation gave Pueblo a group of people who could provide
infrastructure support, and from the beginning Pueblo was seen as
a second-generation effort that would be more deliberate and
intentional in setting expectations.  Pueblo looked for members
who would contribute to the community's vision of
inter-generational learning, instead of recruiting from across
the Internet anyone who was interested in building in a MUD, as
MariMUSE had. Even at Longview, classes were brought onto Pueblo
in small batches as teachers and students were ready. Over time,
Pueblo has grown to include Phoenix area senior citizens,
Hispanic and Native American college student groups, and adults
on the Internet who are interested in education.

Centering on Curriculum

Teachers had integrated a few curriculum projects in MariMUSE,
but they had focused primarily on language and literacy. Since
the network community was an environment that made students at
all levels of academic achievement eager to read and write, the
principal and teachers encouraged its use whether or not online
conversations and activities were thematically tied to
curriculum. With the transition to Pueblo and the externally
funded project, the goals shifted to place a greater emphasis on
curriculum- related work.

During a successful summer camp, teachers and other design team
members experimented with cross-grade projects, portfolios, and
other ideas that would bring a curriculum focus to Pueblo. At the
end of camp, brainstorming about fall curriculum projects was
ambitious. Many ideas were proposed, though none of them was
developed in sufficient detail to create a timeline.

In the press for the start of the school year for teachers and
project meetings for researchers, curriculum projects took a back
seat. After several months, the core team decided that something
needed to be done. The "important" had taken second place to the
"urgent" for too long.

Several of the researchers organized a two-day curriculum retreat
at Xerox PARC. The school principal, teachers, technology aides,
and Phoenix College faculty traveled to the research
setting. This was an important symbolic step; it reinforced the
idea that we were design partners.

Each teacher formed the nucleus of a small design group of 4 to 5
people. Each group identified a specific learning objective and
created one or two detailed scenarios of a concrete, small-scale
Pueblo activity that could help students meet the learning
objective. As a large group, we explored the implications of each
scenario for assessment, student self-monitoring, new software
programming, and the logistics of computer equipment use and
classroom activity flow.

Most of the small groups stayed together during the next months,
though this had not been planned. The technical people who had
helped to create the scenarios became committed to helping them
become real. The research computer scientists had to juggle these
short-term implementation efforts with other design and
development activities associated with their longer-term research
agenda (providing support for constraint programming and model-
building in the MOO). This was not easy; developing robust and
easy-to-use implementations of simple tools is a very different
kind of activity than designing new architectures and languages
in a research setting. Each of the scenarios was at least partly
implemented, facilitated by continued communication between the
teachers and their programming helpers.

The curriculum projects teachers designed included a
collaborative counting activity for first graders; an in-MOO
writing process with peer critiquing for third graders; building
solar system models for fifth graders; and building walk-through
human body systems for sixth graders. Using what they had learned
from MariMUSE, teachers designed activities that were inherently
social and that gave students the opportunity to create and
personalize.

The results were mixed, from the teachers' perspectives.  When
the school year ended, the core group met to reflect on the
year's experiences in Pueblo and to contrast them with earlier
experiences in MariMUSE. MariMUSE had not been supported through
any formal project or external grant, and the teachers noticed
the difference. They felt that MariMUSE had been play, and Pueblo
was work. One teacher said:

MariMUSE was less structured, more fun for everybody.... It was
much more free - we weren't trying to do things
instructionally. It was not as goal-driven or money-driven. Now
we have to have a product - this thing has to work because it has
to be proven to someone else.

Another teacher put it this way:

If you have a specific goal, it's easier to feel failure.  If
you're after learning, you look back and see what happened and
feel success... We have overfocused on what part of the
curriculum [is in Pueblo], but it's really about learning. We
took a more global view about facilitating learning before.

One of these teachers (the first quoted) had designed in Pueblo
an activity that was modeled closely on one she had done earlier
in MariMUSE. Though the Pueblo and MariMUSE activities were very
similar, this teacher's understanding of her own accountability
was different in Pueblo and MariMUSE. As an active member of the
Pueblo core design team, she had developed a plan that fit the
project's goal of curriculum integration. As a teacher and
community member, she had enacted the plan in her classroom and
in Pueblo, where she was accountable to herself and the school
administration as a creative and independent teacher. When there
were challenges with managing classroom activity flow and delays
in implementing the curriculum project, this teacher and others
were uncomfortably aware that people outside their setting were
invested in the results.

Since their earliest experiences with MUDs, teachers and
researchers in the project have believed that one important
factor in students' motivation to participate in MUD environments
is having the freedom to choose what to do.  After the discussion
of differences between Pueblo and MariMUSE, teachers decided that
having choices was important for their participation too. As one
teacher said, "teachers need their own freedom and independence,
just like the kids."

Rather than design specific curriculum projects for the coming
year as a group, teachers individually reflected on what would
constitute successful implementation of Pueblo in their
classrooms. Some of the goals set by teachers included "students
seeing ways to create curriculum projects," "teachers and
students having active roles in projects," "students looking at
their own work and making assessments based on a class-created
rubric," "students actively problem-solving," and "Pueblo should
be an added tool to use, not feel like one more thing that has to
be done."  Teachers seem to have appropriated the interpretation
and use of Pueblo in a new way; they are focused on learning
objectives and finding a fit within the classroom context, rather
than particular outlines of student activities. This
appropriation is an important step toward long-term
sustainability.

For the community to continue beyond the period of the project,
people need to have their own reasons for involvement. But
because the network community is an extension of the school and
community college, the character of involvement is not completely
open to individual interpretation. Community members have freedom
to choose what to do within the constraints and expectations of
the educational setting and the sponsoring
institutions. Participants are still accountable for educational
value and effective use of resources.

4. Mediation

New community members offer new talents, but they can also bring
new stresses because of different languages and different
goals. Williams has pointed out that translation is a crucial
function in participatory design teams [13, 14]. We have also
found this to be true in the design and use of Pueblo.

Mediating style differences

In the move to Pueblo, the three organizations committed to work
as equal partners to develop curriculum projects integrated in
the MOO. The diversity of the groups could have led to divergence
in many aspects of the project, but the most noticeable
difference has been in working style. As Schwab, Hart-Landsberg,
and Reder have described [12], teaching professionals have highly
constrained schedules that can make collaboration challenging. In
our project, we noticed differences in people's styles as they
related to time: how we paced activities, how we scheduled time,
what uses of time we considered productive and what we considered
a waste of time.

Hughes and Walters began to play new, central roles as mediators
between teachers at Longview and researchers at PARC, drawing
upon their background as education researchers to translate in
each direction.

Teachers wanted concrete tools and ideas they could use easily
and quickly in the classroom; they have little free time for
formal planning or curriculum design and are adept at
improvisation. Researchers often work at a slower pace, on a more
abstract level, with a longer-term view; they try to understand
possibilities, processes, reasons for decisions, and
implementation effects. Researchers in our team feel a need to
design elegant solutions, observe and reflect on activities in
the community, and write. Teachers feel a need to be ready to
manage lessons for their students each day, cover the many
essential parts of the curriculum well, and respond to district
requirements for student assessment. On the other hand, sudden
short-term deadlines for writing new code or producing a brief
position paper for a conference did not bother the research
groups, but these deadlines (of which there were several) were a
real hardship for the teachers, whose schedules were both full
and inflexible.

Some subtler cultural differences were our different attitudes
about productive uses of time and tolerance for ambiguity.  As
part of our design process, the core design team has held monthly
all-day meetings in Phoenix since the funded project
began. Initially, these meetings were taken up entirely by
discussions on issues of common concern - brainstorming about new
tools, planning new activities, thinking through the logistics of
bringing in new participants. Researchers seemed to feel
productive even if a whole day was spent discussing complex
issues in detail and making no firm decisions at all. Teachers
found this use of meeting time frustrating - one concrete plan,
however small, would be preferable to a whole set of ideas that
might never be implemented.

Hughes and Walters monitored teacher responses and then named the
problem. Once it had been explicitly raised, the trust, good
will, and friendships that had been nurtured by our participation
together in the network community helped us to adjust our styles
with good humor. The team discussed the problems and worked out
solutions. After the curriculum retreat, teachers began to focus
on specific projects with dedicated research partners and
concrete deadlines. To improve meetings, the day was split into
large group discussions, small curriculum group discussions, and
individual planning and classroom observation, which each group
has found helpful.

Mediation in online mentoring

In diverse communities, people have different skills and may not
always recognize or value the unique things they know.  In
Pueblo, PARC researchers and others on the Internet were
interested in the process of mentoring, so they asked the
teachers and school principal how to be a good coach, how to
approach a child online, how to handle a conflict between two
students, and similar questions. But these skills were so natural
to the professional educators that it was hard for some of them
to see what the researchers were asking or to advise them on
mentoring.

This situation continued for several months as a low-level issue,
until Anne Mourning (Ladybug), a New York enrichment teacher who
was also an active participant of both MariMUSE and Pueblo, wrote
an informal guide to online mentoring. This guide was concrete
and basic, and it was full of examples from Pueblo. It talked
about how to engage students in conversation and how to give them
help without giving them answers. This guide has been valuable
for new mentors, though as with any new skill, mentors need
practice in real situations to become adept at coaching.

Ladybug was able to play an important mediation and translation
role because of her unique perspective as both a teacher and a
remote outsider to Longview. When the school principal and
teachers understood what the mentoring technique questions had
really been about, some were surprised to learn that the
researchers had needed such basic tutoring in how to interact
with students.

Though the participants of our network community subscribe to the
educational vision and local culture, what each wants from their
participation has a different emphasis, depending on background,
professional agenda, interaction style, and so on. As the
community grows, mediators help to bridge many different gaps.

Missed mediation

Billie Hughes, Phoenix College: I tried to help facilitate
development between teachers and PARC researchers, but without
much success. What I think happened is that the researchers took
responsibility for development and Longview teachers and
principal came to grips with what it was they had to do to keep
their commitments to the project... the virtual community enabled
and empowered the teachers and the researchers to work
together. Yet the instructional expertise that Jim and I have is
lost. Often, because time is so tight, communications occur
solely between researchers and teachers, which can mean that
developments may need significant revision to work for other
audiences... Yet the virtual community and the strong
relationships and loyalties that developed among all team members
meant that this changing of roles did not create a serious
problem for the community or the individuals in it. The values of
the community and the shared vision meant that people did not
become obsolete, even if their roles changed.

As designers of a learning community, we want the community to be
strong and self-sufficient, with many open communication paths
across boundaries of all kinds. Many in the community of
designers feel they would like to be involved in a large number
of different activities. Being an intermediary and translator is
an emotionally rewarding role, and in this example there were
also unique professional skills missing when the intermediary was
out of the loop. In Hughes' reflection on this lost opportunity,
she considered what her changing role meant to her sense of
belonging in the community. She was no longer a gatekeeper for
all of the interactions between teachers and researchers, but she
was by no means "obsolete."

There is an issue here that has not yet been resolved in Pueblo -
how to make sure that appropriate mediators are present in
activities where they are needed, without inhibiting the
non-hierarchical, open social groupings that form so easily in
the network community.

5. Growth and changing participation

As the community grows, individuals have taken on different roles
and responsibilities. How people do this and what they are able
and willing to do in this environment plays a crucial role in how
the community grows, what kinds of growing pains it experiences,
and how these are addressed.

Different levels of participation

Cynde Welbes, teacher: So I made sure that I made it to every
meeting, added in my thoughts, listened to others thoughts, and
after the meetings, continued to talk to those who stayed online
afterwards. I felt like a sponge, and I just couldn't get enough
information. All of this, I thought, was a positive thing. But
soon I started hearing things like, "You don't have to be online
as much as Cynde is to be involved." Suddenly my being on- line a
lot was a negative thing...because I had taken part in online
discussions and therefore had information that others didn't
have...

Though there are many effective styles of participation, you have
to show your presence and participate in live conversations to be
part of a network community. "Face time" has been an issue for
most members of the community at one time or another; it has not
just been a problem for classroom teachers with constrained
schedules. (The school principal has been an exception; she is
rarely in Pueblo but retains a strong voice in the design team.)
Even when design team members were active in email or doing other
activities for the benefit of the community, they began to feel
left out if they were absent for a period of time from
Pueblo. Those who spent time online together simply knew more
about what was going on. They were present and could be consulted
when on-the-spot decisions had to be made or when interesting
design discussions began spontaneously.

Some of the tensions around missed opportunities for
participation were eased through technical means. Optional
recorders were added to public communication channels. (A channel
is a mechanism for talking with a group of people who are not in
the same room together in the MOO.) Certain channels were
routinely recorded, including the "core" channel for members of
the core design team and channels dedicated to discussion topics
such as pedagogy and programming. With recorders, core channel
discussions left a reliable trail. The core channel became a
synchronous communication tool for people who were present and an
asynchronous tool for people who were absent. Even with this
tool, however, interacting with others online is still very
important for maintaining relationships in the community.

Governance

As the community grew, it became apparent that it needed more
wide-ranging discussion of certain public issues, such as who
could become a member of Pueblo or when someone should be granted
additional quota (a built-in building allowance). There were no
advertised policies in these areas, and the rationale for
individual decisions was not always clear to those who stood
outside the decision process. The administrators of the community
were interested in developing policies and guidelines, but they
were also sensitive to the implied questioning of past decisions,
which they had made for the benefit of the community.

Through a series of email discussions and face-to-face meetings,
new immigration and guest visitation policies were specified. New
technical commands were built into the system that would allow
people other than technical wizards to do appropriate
administrative tasks, such as invite guests, adjust quota, and
change students' passwords. A "teacher utilities" package had
always been planned, but its priority was raised during this
process. These sensitive discussions were held partly in email,
but the core team made the final policy decisions in a
face-to-face meeting. A more detailed account of the co-evolution
of some of Pueblo's social policies and technical mechanisms is
given in [9].

Recognition and status

Recently, governance structures and the function of the core
design group have been challenged by a group of pre-teens and
teens in Pueblo, prompting new attention to the question of
status in the community. This group formed a "secret club," an
exclusive group with a clubhouse, special communication channel,
and rules for voting on membership applications. The secret club
wasn't really such a secret, and signs of its existence
occasionally spilled out into public view. One member left in a
dispute. This prompted discussion among the core team - what
action, if any, should be taken? The teachers and school
principal had a united response. Since they would discourage this
kind of club if it emerged on the playground or in the classroom,
they would also discourage it in Pueblo, since they saw Pueblo as
an extension of these other places.

In discussion with the club members themselves, the heart of the
issue seemed to be status, authority, and responsibility - who
had each and why. Though the core design team had thought through
these same questions as it considered immigration policies and
teacher utilities, the wider community had not been involved in
that discussion. When the teenagers defended their secret club,
they drew comparisons to the core team itself and the staff of
admins (wizards), as an exclusive group of people with special
powers which regularly held closed discussions on its own
communication channel. The teenagers asked why they could not be
admins or core members too, so they could "actually count as
someone on the MOO," as one said.

In the core group itself, there are clear areas of responsibility
and corresponding status based on the organizational structures
of the project and the participating institutions - there are
principal investigators for the grant, a school principal, and a
department head, for example. In the network community, status is
not clear and there are few visible markers of difference in
responsibility or expertise.  The teenagers were anxious to
become wizards probably because this is the most visible status
indicator in the MOO.

The timing of the secret club affair was perfectly aligned with
the formation of a new documentation project, which did provide
more opportunities for power, responsibility, and
recognition. This team is open to anyone who will take seriously
the responsibility of creating high-quality help text, and many
of the secret club members have joined. The team has its own
meeting place and communication channel, some of the external
signs of being a group. It also has real power and responsibility
(both technical and social), since the new help text created by
team members replaces the standard help text. This is the first
extension of MOO-caretaking responsibility to non-core members,
and though the group is still new, it appears to be very
successful so far.

The community will not be sustainable if students do not choose
to build and participate in the development of the
community. Their creativity, engagement, and independence of
action are crucial. And they will not want to be there unless
they see how they have an impact and are part of
decision-making. In this community design project, there is a
strong mutual need between the community and the project.  As a
school-centered community, the community needs organizational and
technical support from the institutions around it. But without
the children, teachers, grays, and other adults who are not on
the core team, the community would be a very dull place.

Keystone players

One special community member, Kimberly Bobrow (Hobbes), became a
unifying member of the team.

Kimberly Bobrow: Cynde Welbes and I feel so strongly about the
idea of welcoming members to the community that we created an
online Welcome Wagon, complete with brownies in a foil covered
plate. We keep a close eye on who logs in, and we are both on
regularly enough to know instantly when someone new shows
up. When that happens, we both hightail it to the wagon, step on
the virtual gas, and zip off to the Visitor's Center to spread
some of that Pueblo feeling.

Hobbes has used the strength of her personality and her expertise
to turn Pueblo into a place where children create worlds of their
own. She has provided easy and popular ways for children to make
custom cars, adopt pets, and create and consume vast amounts of
virtual food. Her constant online presence allows her to stay in
touch with what is going on all around Pueblo; one of the roles
she has adopted is to make connections between people with common
interests. Her background as a teacher, technical wizard, and
all-round confidante have made her one of the central mediators
in the community. We suspect that every network community needs
such "keystone" people who migrate easily between roles and
social groups.

Evolving niches

Soon after Pueblo opened, different kinds of expertise began to
emerge in the network community. One of the teachers (Welbes)
worked on developing technical skills in the MOO and acquired a
"wizard" character, which gave her the ability to do on-site MOO
administration at Longview. This suited her own aptitude and
changing goals, since she had earlier decided that she was
interested in a technical career. Another teacher (Cyndy Olson)
who was particularly good at articulating teachers' viewpoints
was named a member of a small cross-organization executive
committee that was formed to handle broad project and community
issues.

These new roles were a visible recognition of particular kinds of
expertise. The changes were sensitive because they seemed to
place value on some kinds of expertise more than others. Each of
the seven teachers in the core team had unique strengths, but
only some were being openly recognized. The Longview principal
addressed this by explicitly announcing to the core design team
each teachers natural leadership areas (writing, science,
cooperative learning, and so on). This move eased tensions, and
by exposing teachers' special strengths to non-Longview team
members it also lay groundwork for new conversations and
partnerships.

In Pueblo, the core team consciously sought to recognize and
nurture the many talents needed for the project to succeed.  One
person became a vocal champion of educational assessment and
outcomes, another of usability. The special interests and talents
of each person emerged, often informally through the cumulative
effect of interactions in different social groupings, rather than
through people being appointed into particular roles.

CONCLUSION

In many work-oriented design projects, designers work closely
with users for a time to develop a significant new
technology-based system. Participatory design in/of our network
community has been distinctive in several ways:

· Boundaries between users, designers, and implementors are
fuzzy. Everyone in Pueblo is both a builder and a community
member. People's roles and responsibilities evolve over time.

· Social interactions and relationships are important aspects of
community life. Some social concerns (such as immigration and
property rights) became explicit elements to be addressed in the
design process. An ongoing issue is finding ways of recognizing,
leveraging, and making visible the different kinds of value that
community members bring.

· Rather than a large design with many contributors, people work
toward a set of independent goals in the context of a larger
vision of a learning community.  Design is decentralized and
open-ended. Participants learn from the emergent properties of
the medium and its evolving use, and they design new activities
as their understanding of current experience grows.

· Projects in the community require collaboration among people
who have different backgrounds, professions, agendas, styles, and
interaction patterns. The community benefits from mediators and
translators to cross these many boundaries.

Major transitions in Pueblo's evolution have been marked by
cyclical periods of action and reflection, as the community looks
outward, inward, and outward again. The community increases in
complexity, becomes aware of the effects of complexity (such as
changing roles or social tensions), and makes adjustments. The
reflection itself is an important part of the process and is a
key to how a community of designers intentionally affects the
design of the community.

ACKNOWLEDGMENTS

We thank Bill Lightfoot, Paula Melton, and Cyndy Olson for asking
their students to think and write about Pueblo, and we thank the
students who wrote thoughtful and entertaining responses. We are
grateful to everyone in the Pueblo community for making it such
an enjoyable place to work and play. Our thanks to Markus
Fromherz for useful feedback on earlier drafts of this
paper. This work has been supported by ARPA (Contract
N66001-95-C-8610).

REFERENCES



1. Bandura, A. The self system in reciprocal determinism.
American Psychologist, April 1978, 344-358.

2. Cherny, L. The MUD Register: Conversational Modes of Action in
a Text-Based Virtual Reality. Ph.D. Disser- tation, Stanford
University, December 1995.

3. Curtis, P. Mudding: social phenomena in text-based vir- tual
realities. Intertrek 3(3), 26-34.

4. Curtis, P. and Nichols, D. MUDs grow up: social virtual
reality in the real world. Presented at the Third Interna- tional
Conference on Cyberspace (Austin Texas, 1993).

5. Ehn, P. Scandinavian design: on participation and skill.  In
D. Schuler and A. Namioka, eds., Participatory Design. Lawrence
Erlbaum, 1993, 41-78.

6. Gärtner, J. and Wagner, I. Systems as intermediaries:
political frameworks of design and participation. In Proc. PDC
`94 (Chapel Hill NC, 1994), CPSR, 37-46.

7. Kort, B. MicroMUSE. 

8. Morningstar, C. and Farmer, F.R. The lessons of Lucas- film's
habitat. In M. Benedikt, ed., Cyberspace: First Steps. MIT Press,
1994, 273-302.

9. O'Day, V.L, Bobrow, D.G., and Shirley, M. The social-
technical design circle. To appear in Proc. CSCW `96.

10. Rheingold, H. The Virtual Community: Homesteading on the
Electronic Frontier. Addison-Wesley, 1993.

11. Schuler, D. New Community Networks: Wired for
Change. Addison-Wesley, 1996.

12. Schwab, R.G., Hart-Landsberg, S., Reder, S. Collabora- tion
and constraint: middle school teaching teams. In Proc. CSCW `92
(Toronto, 1992), ACM Press, 241-248.

13. Williams, M.G. Enabling schoolteachers to participate in the
design of educational software. In Proc. PDC `94, (Chapel Hill
NC, 1994), CPSR, 153-158.

14. Williams, M.G. and Begg, V. Translation between soft- ware
developers and users. Communications of the ACM, 36(6), 102-103.