AudioFile overlays imperceptible tones on standard audio tracks to embed digital information that can be decoded by standard mobile devices. AudioFile lets users explore their media more deeply by granting them access to a new channel of communication. The project creates sound that is simultaneously meaningful to humans and machines. Movie tracks can be annotated with actor details, songs can be annotated with artist information, or public announcements can be infused with targeted, meaningful data.

People form dynamic groups focused on topics that emerge serendipitously during everyday life. They can be long-lived or of short duration. Examples include people interested in buying the same product, those with similar expertise, those in the same location, or any collection of such attributes. We call this the Human Discovery Protocol (HDP). Similar to how computers follow well-established protocols like DNS in order to find other computers that carry desired information, HDP presents an open protocol for people to announce bits of information about themselves, and have them aggregated and returned back in the form of a group of people that match against the user’s specified criteria. We are experimenting with a web-based implementation (brin.gy) that allows users to join and communicate with groups of people based on their location, profile information, and items they may want to buy or sell.

Collaborating and media creation are difficult tasks, both for people and for network architectures. CoCam is a self-organizing network for real-time camera image collaboration. Like with all camera apps, just point and shoot; CoCam then automatically joins other media creators into a network of collaborators. Network discovery, creation, grouping, joining, and leaving is done automatically in the background, letting users focus on participation in an event. We use local P2P middleware and a 3G negotiation service to create these networks for real-time media sharing. CoCam also provides multiple views that make the media experience more exciting–such as appearing to be in multiple places at the same time. The media is immediately distributed and replicated in multiple peers; thus if a camera phone is confiscated or lost, other users have copies of the images.

Electric Price Tags are a realization of a mobile system that is linked to technology in physical space. The underlying theme is that being mobile can mean far more than focusing on a portable device—it can be the use of that device to unlock data and technology embedded in the environment. In its current version, users can reconfigure the price tags on a store shelf to display a desired metric (e.g., price, unit price, or calories). While this information is present on the boxes of the items for sale, comparisons would require individual analysis of each box. The visualization provided by Electric Price Tags allows users to view and filter information in physical space in ways that was previously possible only online.

Have you ever been in the middle of a conversation and needed to share your location with the other party? Geo.gy is a location shortener service. It allows you to easily share your location with your peers by encoding it in a short URL which we call a "geolink." It is platform-independent, and based on HTML5, so you can use any device with a modern browser to generate a geolink, simply by visiting the project's page. There are no user accounts, so geolinks remain anonymous. You can use Geo.gy to add location context to a post, SMS, anything you want decorated with location context.

Line of Sound shows how data can be used to deliver sound information only in the direction in which one looks. The demonstration is done using two 55-inch screens which transmit both human and machine relevant information. Each screen is used to show a video that flashes a single bit indicator, which transmits to a camera mounted on headphones. This is used to distinguish between the two screens, and to correlate an audio track to the video track.

Point & Shoot Data explores the use of visible light as a wireless communication medium for mobile devices. A snap-on case allows users to send messages to other mobile devices based on directionality and proximity. No email address, phone number, or account login is needed, just point and shoot your messages! The project enables infrastructure-free, scalable, proximity-based communication between two mobile devices.

Human beings have long shaped the physical environment to reflect designs of form and function. As an instrument of control, the human hand remains the most fundamental interface for affecting the material world. In the wake of the digital revolution, this is changing, bringing us to reexamine tangible interfaces. What if we could now dynamically reshape, redesign, and restructure our environment using the functional nature of digital tools? To address this, we present Recompose, a framework allowing direct and gestural manipulation of our physical environment. Recompose complements the highly precise, yet concentrated affordance of direct manipulation with a set of gestures, allowing functional manipulation of an actuated surface.

SonicLink is a fully decentralized, proximal communication framework for personal devices to seamlessly discover, connect, and interact with arbitrary public installations (such as digital billboards). It establishes connections based on audio proximity when you are not on the same network: being near in physical space does not mean you are near in "network space." SonicLink uses near-ultrasonic acoustic signals that permit devices and installations to discover each other. It also exploits peer-to-peer proximal wireless networking techniques for establishing a high-bandwidth, low-latency link between the device and the installation. Possible uses include “borrowing” a large-screen TV and a camera in the public space for personal video conferencing, presenting personal notifications to a public display, and taking over neon lights for visualizing music on a phone.

T+1 is an application that creates an iterative structure to help groups organize their interests and schedules. Users receive instructions and send their personal information through mobile devices at discretized time steps, orchestrated by a unique, adaptive scheduling engine. At each time-step, T+1 takes several relevant factors of human interactions (such as participants' interests, opinions, locations, and partner matching schedules), and then computes and optimizes the structure and format of a group interactions for the next interval. T+1 facilitates consensus formation, better group dynamics, and more engaging user experiences by using a clearly visible and comprehensible process. We are planning to deploy the platform for both academic and political discussions, and will analyze how user opinions and interests evolve in time to understand its efficacy.