Q&A@T

A common feature of a modern society with all sorts of information technological developments is, we are more connected than never before. We can easily privatize public spaces using our cellphone, iPod, and all kinds of personal mobile technologies. We can easily connect to people who share the same interests or political views. We can be anywhere because we have all sorts of technologies that transport us virtually and immediately.

But we are also more detached and alienated like never before. We are being transported to places and connected to different people all the time. We live in Boston, but whether it is Boston or not, doesn't really matter anymore. Places become meaningless. We do not spend time learning about the city and its people anymore.

This project is an experiment on how we can change that through introducing a commmunication channel to the public. We understand people are already bombarded with information, so the channel has to be reall lightweight, in terms of the cost of time and effort of participating.

We focus on the Boston subway system: T, in this thesis. It is not uncommon to realize how a subway system connects people living in different parts of a city. We would like to take advantage of the fact that T gathers people of different cultures, occupations, and ages, and turn the experience of waiting for a T and riding on a T train into a time of appreciation the reality and complexity of the city. We would like to provide the people a way to express and be informed with local volatile issues, matters of minor importance, observations made on site, collected view, voices and moods of the people around us.

tagNet

It is already a common practice to document every aspect of life with digital media format such as images and video clips. tagNet is an experiment on how in the future we will be sharing and browsing those personal memories, and how the public will be able to explore the huge media space.

We challenge the existing ways of organizing media contents by directories, lists, and hierarchies. We see contents on one flat surface, and each of it is tagged with keywords that the owner believes best describe the content. Those keywords are then merged together to form a huge network where each word is a node and a link between two nodes represents their semantic relationship, defined by millions of people who use the service.

tagNet is such a network we built using tags from a commercial online album service called flickr.com. It is a semantic network that both the vocabulary and how much words relate to each other are defined collectively by all the users on flickr. We built two applications using tagNet: (1)don't search, explore and (2) contextual browsing.

(1)"don't search, explore" is an application designed for graphic designers or people who would like to explore the whole visual space of online images. One can type in words and the interface will locate you to a portion of the whole image space to get you started to explore it. The location of images and concept words around it give people a sense of where they are.

(2)"Contextual browsing" is a browsing interface for people to browse their own media contents by setting up the "context." One can type in words and will be returned with images, while the relative distances (locations) of images will change depending on the context. For example, by typing in geographical words like "Boston" and "New York," images tagged with MIT will be close to ones tagged with Fenway. But if one types in "baseball," images tagged with "Shea Stadium" in New York will be close to those tagged with Fenway in this case.