COMPUTING
Mobile URBAN Social Software: Facilitating Serendipity or Encouraging Homogeneity?
Mobile social systems can offer heterogeneous views of cities rather than encouraging users to socialize with people they already know and privileging consumption- and entertainment-based urban experiences.
ervasive computing researchers often represent the urban landscape as a space full of social opportunities and cultural landmarks that can provide users with enriching or educational experiences. The recent development of mobile social software draws on a similarly optimistic vision of the city. Mobile social software lets users connect to social networks through mobile phones and devices. These systems attempt to facilitate serendipitous social connections by providing additional contextfor example, by helping users broadcast their locations to a group of contacts in the hopes that faceJennifer Thom-Santelli to-face interaction will result. Cornell University, HCI Group, Mobile social software has and CEmCom largely targeted cities on the assumption that urban areas provide a sufficient density of people that users will encounter compatriots as they go about their everyday lives. The design of mobile social systems privileges a type of urban experience thats centered upon an idealized category of users. I draw from artsinspired practices to suggest design strategies that might help mobile social software systems widen their representations of the urban experience. At first glance, arts practices might seem far removed from the engineering of technology. There is, however, an established tradition of arts practices focused on creating new technologies.
10
PERVASIVE computing
P
These practices ask and answer critical questions about technology designs social and cultural implications. Design methods that draw from the arts can help uncover and alter the closed loop of relatively youthful and affluent users built into mobile social software, drawing on three themes: critical awareness of whats being designed into and out of these systems, use of interventions to both respond to articulated user needs and open space for new user practices, and encouragement of flexible interpretation to let appropriation emerge among unexpected stakeholders.
Configuring the ideal user
Critical analysis of technology design assumes that designers dont operate in an objective vacuum and that values play a role in how designers conceive the systems they create. Technological artifacts reflect and reinforce social structures. One way in which values shape a technologys uptake is in the configuration of an ideal user groupdefining who they are and how they will use the system. Steve Woolgar observed that when designers conduct usability tests with real users, they evaluate these users input according to how closely they correspond with the user the designer imagined.1 Science and technology scholars have
Published by the IEEE Computer Society I 1536-1268/07/$25.00 2007 IEEE
also demonstrated that users appropriate technologies in ways the designers didnt anticipate. However, theres a danger that designers will consciously or unconsciously constrain use by envisioning a limited set of actions overly guided by these imagined target users. The HCI field has formalized the creation of user types through the design method of personas. In this method, designers describe a type of idealized user on the basis of previously gathered user research such as interviews, questionnaires, or observation.2 Personas become a reference point as the design develops, letting the designers explicitly consider how the eventual user type might react to the systems functionality. Designers create these personas to help minimize their unconscious conception of users. But they often choose a target user type who resembles the designers or the types of people with whom theyre familiar. For mobile technologies, the target users will likely be well-educated professionals with disposable income, and the personas reflect that tendency. For example, eMotos3 designers used questionnaire data from prospective users to develop the persona of Sandra. Shes a young, technologically savvy woman whos socially active, comfortable with using her mobile phone to stay in touch, and expressive enough to want to use a mobile system to convey ambiguously affective messages. During user evaluations, eMotos designers developed scenarios such as conflict with racist doormen at a club or getting the right job. These scenarios were based on social situations that the subjects, young women like Sandra, might encounter. By making the scenarios so resonant with the target group that they almost guarantee a measure of success during testing, the designers constrain what an experience with the system should be like and to whom the system would be meaningful. eMoto is one instance of a current
JULYSEPTEMBER 2007
mobile social software system designed with a young, technologically adept user in mind, despite the uptake of the mobile phone across diverse user populations. Recent market research suggests, however, that teenagers and students have taken up mobile social software most avidly.4 Again, this information illustrates the tension between user agency and designer intent. Have younger users simply appropriated an available technology according to their needs and desires, or have designers made the technology appropriate only for younger users?
Seeking others just like me
The restrictive selection of certain types of personas, such as young singles, to describe users plays a role in limiting the systems potential appropriation. For example, Dodgeball (www.dodgeball. com) is designed to facilitate dates by letting you categorize people as crushes in your buddy lists. In such systems, users present carefully crafted public faces in the form of profiles in which they broadcast their romantic status and indicate their likes and dislikes. In the case of Dodgeball, you can infer fellow users
spective, such software presents a view of the city thats dominated by its nightlife, drawing attention to its bars, clubs, cafes, and restaurants (Dodgeballs venue reviews are an example). The citys diversity becomes less celebrated as users seek out homogenous experiences that the softwares design facilitates. The widely observed sociological phenomenon of homophilythe tendency to affiliate with people similar to you explains the desire to visit places that your social circle has vetted to encounter other people just like you. A study on adult friendships in an American city (Detroit) and a German city (Alteneustadt) showed that the bias to seek out same-status friends is highest among those with higher levels of education, in their twenties, and in occupations of high prestigea noticeably similar profile to the target mobile social software user.5 Mobile social software helps heighten this tendency by explicitly letting users select the people with whom they want to interact according to specific criteria. These selective aspects encourage homophily to the point that serendipitous interactions are only really possible with peo-
Serendipitous interactions are only really possible with people whose interests are probably similar to yours, at locations that others just like you have preapproved.
tastes through the places they have visited. By explicitly encouraging this form of connection, mobile social systems designers attempt to attract younger users who traverse the city to meet prospective romantic partners. If a system focuses on facilitating romantic encounters, the software will favor a particular type of urban location, especially when taken in combination with the users relative youth. From that perple whose interests are probably similar to yours, at locations that others just like you have preapproved.
Encouraging the heterogeneous urban experience
Several practices can promote heterogeneity and help designers widen the scope of the social interactions that they seek to support. First, looking at the
PERVASIVE computing
11
URBAN COMPUTING
motivations of nonusers and resisters to a technological system might help designers consider populations other than youthful, affluent consumers. Second, designers can use prototypes as interventions to provoke a different type of user experienceone that makes explicit uncomfortable aspects of the urban experience as a way to spur critical reflection. Third, these interventions can be deployed creatively to a diverse audience instead of to the usual suspects. And finally, designing to support multiple levels of interpretation might encourage unexpected stakeholders to find great value in such systems. Moving beyond the typical user Designers can expand their conception of the configured mobile social software user and look beyond those already using the technology. Science and technology studies have questioned the boundaries of the definition of the user and argued that resisters to a technology often play a large part in how its acceptance is negotiated.6 Why, for example, do older mobile phone users avoid mobile social software systems when they could just as easily appropriate them to arrange their grandchildrens play dates? Again, system interfaces might constrain
early automobile, the company developed a Model T with rural farmers in mind (as opposed to the affluent urban users initial models targeted). The automobile was slowly negotiated into the fabric of rural life.7 Typical mobile social software might highlight venues for consumption of food, drink, and entertainment, but such systems largely fail to recognize employees first-hand experience. Instead, customer reviews define such venues, and any depiction of the staff portrays the users opinion of the service. Its easy to imagine developing a mobile social system for disenchanted waitstaff to keep track of badly behaved customers and inadequate tippers. Drawing from critical practice and tactical intervention By favoring a type of space that values entertainment, mobile social software offers an overly simplified vision of the urban experience. This overlooks the political consequences of viewing space as sites for consumption. Amanda Williams and Paul Dourish draw upon the work of critical geographers to make a case for seeing the city as made up of complex spaces linked by differing access to flows of resources, such as information and infrastructure.8
Its easy to imagine developing a mobile social system for disenchanted waitstaff to keep track of badly behaved customers and inadequate tippers.
use by targeting the ideal user in primarily defining activity partners as potential mates. Incorporating resisters objections into system design can create economic and political opportunities. After the Ford Motor Company acknowledged the existence and motivations of resisters to the
12
PERVASIVE computing
An example of the complicated relationship between resources, space, and social actors is gentrification. Artists move into low-rent areas for affordable live-work space, which begins transforming the neighborhood in a way that attracts higher-income residents. This squeezes out the original inhabitants,
who must move to the exurbs or out of the city entirely. Mobile social software exacerbates this conflict by specifically highlighting the places of consumption and entertainment, such as bars, restaurants, and cafes, that pop up in newly gentrified neighborhoods. As I mentioned earlier, arts-inspired design practices can help to widen the range of spaces that mobile social software depicts, especially in moving urban computing away from an optimistic perception of the city as full of positive social opportunities. Design noir, which Anthony Dunne and Fiona Raby proposed,9 focuses on designs for the complex and transgressive psychological experiences that traditional approaches usually omit. Satisfying darker human needs, such as deception and fear, design noir draws attention to the limitations of electronic systems that focus only on well-being and satisfaction. For example, the designers developed a series of seemingly benign household devices in their Placebo Project as a way to make explicit user fears about electromagnetic fields. This approach has recently been adapted within HCI as an urban computing resource to enable the exposure of the citys darker side. Exurban Noir focuses on the sprawl that forms at a citys edge. It aims to draw attention to urban areas that arent postcard-ready and that are populated by those without the economic resources to live in the urban center.10 Because these neighborhoods might not contain the consumerist venues that current mobile social software values, theyre rendered almost invisible, even as they grow in area and To population. meaningfully engage overlooked inhabitants and underserved city spaces, I propose that designers draw more deeply from the critical tradition of the arts and its use of tactical interventions. Tactical intervention is the practice of deploying a real-world design that subverts, challenges, and renegotiates accepted meanings and existing power structures.10 Such
www.computer.org/pervasive
Figure 1. The AIR device. Embedded sensors collect air quality data to find pollution hotspots in urban environments.
interventions provide a mechanism to expose social, political, and cultural questions that an engineering-based practice of system design might not explore. Im not necessarily suggesting that designers should co-opt tactical intervention as an institutionalized method, similar to how HCI has taken up personas. However, posing difficult questions, such as whom a specific system might ignore, through an early intervention in the design process might provide a better picture of nonusers and resisters as well as uncover potential political ramifications. For example, Homeless Vehicle focuses on those who reside on city streets and draws attention to a constituency thats often overlooked or purposefully ignored.11 This work transforms the shopping cart that the urban homeless have appropriated into a mobile structure with a horizontal tube that provides space for sleeping. Its placed on top of a wheeled storage container for belongings and recyclable and scrap materials that are redeemed for cash. Homeless Vehicle is deliberately imposing in scale, rocketshaped and almost resembling a weapon, to make this marginalized group visible to the citys other inhabitants. The artist considered the specific type of mobility that the homeless experience; many of their movements are in response to the actions of the police, quite different from the relative freedom of movement that other urban dwellers experience. Homeless Vehicle highlights the vast inequities among the citys populations by addressing the pressing shelter and storage needs of those who reside on the streets. It also illustrates how designers of urban computing systems, such as mobile social software, assume that urban mobility is largely active and discretionary. Leveraging homophily for diversity Earlier, I proposed that mobile social
JULYSEPTEMBER 2007
software designers should expand their conception of a typical user beyond the configured target group. To this end, I describe an alternative strategy for reaching different urban residents that designers could use to find participants in a tactical intervention. To bring these hidden users to the foreground, I draw inspiration from a sociological research methodology called snowball sampling, in which participants in an intervention are referred by a starting group of users. Again, this technique relies on the assumption of homophily, so each new referral would resemble the referrer in some desired way. The difference is that the initial group, in this case, wouldnt be the typical mobile social software user. Snowball sampling has been successful in studies of truly marginalized urban populations, such as active drug injectors,12 but it can be adapted for gathering design insights through tactical intervention. The AIR (Areas Immediate Reading) project used a snowball-sampling approach to introduce a system of portable air quality monitoring devices that visualizes urban areas pollutant levels (see figure 1). To ensure a diverse seed group, the designers encouraged participation by holding open office hours at the project headquarters in downtown New York City, setting up distribution tables
on city streets, and encouraging people to sign up to participate through the project Web site (www.pm-air.net). More importantly, they asked each user to keep the AIR device for only 24 hours and then to pass it on to someone that they knew. By decentralizing the distribution mechanism, AIRs designers attempted to spread the intervention out to as many users and spaces as possible. This created a more expansive portrait of the city while providing a way for urban residents to reflect on their surroundings. In working from a position of relative economic and social privilege, theres a danger, especially in arts-inspired systems, that a tactical intervention could become an inside joke that only a privileged few understand. The experimenter should approach the framing of the interaction carefully, especially with those less familiar with technology, to lessen the possibility of condescension or intimidation. In the case of AIR, when more experienced users passed the mobile monitoring device on, the experimenters asked them to show new users how it worked and how easy it was to operate. By operating in the context of an existing social relationship, this procedure might lessen the fear of the allknowing experimenter.
PERVASIVE computing
13
URBAN COMPUTING
Figure 2. A Yellowarrow image. This system lets users physically mark finely grained urban sites as meaningful using yellow arrow-shaped stickers. Passersby can text message the number on the sticker to receive a message from the user who marked the spot.
SMS number and receive a description of why the user decided to leave the sticker (see figure 2). In effect, it encourages a microlevel representation of the city that lets urban residents contribute their expertise of locations that arent just sites of consumption. Designing for multiple interpretations Phoebe Sengers and Bill Gaver13 approach system design with enough flexibility to encourage multiple levels of interpretation and to let multiple meanings emerge as users appropriate the technology. Supporting diverse interpretations through design can encourage unexpected stakeholders to take up mobile social software in ways the designers didnt originally envision. Here I discuss an example that draws from one suggested design strategy: supporting usability by not constraining use. Such a system might provide clear direction as to how to control and operate it, but its ultimate meanings and purpose are open to user interpretation. TXTmob is a text-message-broadcasting system that was used to communicate and coordinate during decentralized protest activities at the 2004 US Democratic and Republican conventions.14 Users could create their own groups and specify them as public or private and moderated or unmoderated. Because TXTmob was so structurally flexible, a complex variety of stakeholders took it up. Although originally designed mainly for activists, TXTmob was purposefully open enough that other usersand even resisters to activism, such as law enforcement agencies could use it to monitor protests or even to figure out where traffic would
14
PERVASIVE computing
develop. TXTmobs reach also extended beyond the cities where the activism was occurring. Another group of users, geographically distant supporters who couldnt attend and participate in the planned protests, expressed solidarity by sending supportive messages to the unmoderated public groups. When these alternative groups began to appropriate the system, its designers didnt constrain the interpretations that emerged. In fact, they viewed them as a measure of a successful intervention. Sengers and Gaver also suggest that its just as important for such systems to lift restrictions on the type of urban location thats worthy of review. Excluding locations that arent represented on a postal map might make it impossible for users to comment on places integral to their urban experience by using mobile social software. By letting users finely control what they believe are meaningful...