Concepting
Mobile phones have traditionally been used to connect remote people. With Bluetooth, W-LAN and other short range radio technologies, mobile phones could also be used to connect proximite or co-located people. Mobile devices could exchange data with or without the users' immediate awareness (Figure 1). In an encounter between spatially proximate people, how can information in digital realm support and augment existing social behavior, practices and experiences taking place in real space?

Figure 1.

DigiDress application allowed mobile users to create a page on their phone with text and imagery describing themselves, their interests, dreams, things they are proud of, favorite jokes or any other content. Pages were then viewable by other users within Bluetooth range (typically 10-20m), without the page owner's explicit consent. In this way, DigiDress users could take a 'peak view' at others without revealing their identity, similar to peaking at non-aqcuinated people in public spaces.

Design principles of DigiDress included:

• Must support rich identity expression, including text, pictures and graphics
• Creating and editing identity expression should be simple and fast
• Should support one-way communication ('lurking') as well as mutual dialogue
• Should support encounters between aqcuinated as well as non-aqcuainted users
• Should support planned and unplanned social encounters
• Should support usage in public spaces, e.g. streets and bars, as well as semi-public spaces (e.g. meeting rooms and class rooms)
• Balance concerns about privacy and need for identity expression

Prototyping: DigiDress features
DigiDress prototype was implemented on Series 60 platform, being able to install and run on four commercially available phone models in fall of 2003 ( Nokia 7650, Nokia 3650, N-Gage game deck, Nokia 6600).

Creating identity expression. Emphasizing identity expression, the editor was at the core of DigiDress. Creating 'a DigiDress' on the mobile phone should be a media-rich, yet easy process, with templates assisting the creativity of the user.

Figure 2. Creating and editing a DigiDress.

The editor comprised 4 tabs (Figure 2A). Tabs 1 and 3 allowed users to add free text and one image. Tabs 2 and 4 contained fields, each of which could be filled with either text or image (B). We provided suggestions for headings, in order to stimulate the imagination of the users. User could ignore the headings, change them or create new ones. While saving a DigiDress, user was asked to choose one of 12 graphical frames (C), which would cover the margins (D). If no content was added to a tab or a field, that tab or field would not show up in the viewer mode (D). Since all but one of the supported phone models had camera, creating and adding images was an intuitive feature. At any time, users could come back and modify their DigiDress.

Lookaround. This feature (Figure 3) scanned the environment for other DigiDress users and Bluetooth devices, and presented them in a list. The identification of users was not based on phone numbers, but on the unique Bluetooth device identifier (MAC address). If the identified device was running DigiDress application, the Tab 1 image and the first 45 characters of text were automatically fetched and listed (Figure 3E). From the list, user could choose to open the full DigiDress (F). By scrolling sideways in the viewer (F), the full contents of the DigiDress could be browsed. From the viewer, user could also save the DigiDress locally, send Bluetooth messages, add public comments or view the comments this user had received from other users (G). If the device did not have DigiDress application running, only the Bluetooth name of the device was listed.

Figure 3. Lookaround, result list and viewing another user's DigiDress.

Lookaround and viewing DigiDresses were conducted without prompting the DigiDress owner for authorization. As long as the DigiDress application of any given user was running in the foreground or in the background, other users within Bluetooth range could fetch that particular DigiDress. Thus, the DigiDress owner did not have to actively use the application (or the phone) at the time of someone else downloading his/her DigiDress. Neither were owners made aware that another user was downloading the DigiDress pages. The only trace of someone's viewing the pages were provided in My Popularity feature, which listed the number of 'views' the DigiDress had received as well as the timestamp of those. This rather 'public' design of the system aimed to be in line with the design principles presented above, mimicking existing social practices in public spaces.

Figure 4. Comments, Bluetooth messaging, and DigiDress watch.

Bluetooth messages. Bluetooth messages were simple and private text messages between two users (Figure 4J). Recipients, who were notified with a sound signal, could reply to an incoming message but not store them.
Comments. Comments provided a more public form of communication. Received comments were automatically stored in a list (Figure 4H and I). This list was available not only to the page owner, but could be accessed by any visitor to a DigiDress (Figure 3G). In this way comments worked as a proximity 'guestbook'. Page owner could delete any comments, and also add own comments.
DigiDress watch. The DigiDress watch feature (Figure 4K), allowed users to set a time interval for automatic Lookarounds. If a DigiDress device was found nearby, a sound notification would be delivered.
Leaving the DigiDress application could be done in two ways. By 'shutting down', the application was closed and the pages were not made available to others anymore. In this mode, the device would be treated as any other non-DigiDress device when scanned for. 'Exit' on the other hand, merely pushed the application into the background. In this mode, DigiDress was still viewable by others and the user could still receive comments and Bluetooth messages.
Phone-to-phone distribution. DigiDress application also allowed users to distribute the software to non-DigiDress users via Bluetooth or infrared. This was thought to facilitate the uptake of the application which was critical to the success of DigiDress.

DigiDress user study
DigiDress was provided to Nokia employees for user trial. The software was made available and users with compatible phones were invited to download and try it out. The DigiDress prototype was equipped with a logging functionality that enabled us to collect very detailed information about what features were used and how much. During the study we collected 46 DigiDresses which were later subjected to analysis. We also interviewed 10 of the most active DigiDress users.

During the trial period (89 days) 618 users installed DigiDress on their phones. The average use span was 25 days. The identity expressions created were both serious and playful, revealing and non-revealing. Factors influencing the identity expression included strategies for personal impression management, privacy concerns, and social feedback. The application was used with both acquainted and unacquainted people, and viewing the identity expression of people nearby was one major motivation for continued use. Direct communication features such as Bluetooth messages were not commonly adopted. In several instances, DigiDress acted as a facilitator for 'real' social interaction between previously unacquainted users. Privacy concerns and their alleviations, as well as use barriers, were identified.

Read more

Persson, P., Blom, J. & Jung, Y. (2005) DigiDress: A Field Trial of an Expressive Social Proximity Application, in Beigl et al (Eds.): UbiComp 2005, LNCS 3660, pp. 195-212, 2005.

Presentation slides, Ubicomp 2005 (ppt).