Collaborative Passive Browsing using Graphical Images

Problem
The exponential growth of the World Wide Web has resulted in a flood of information that leaves users struggling to stay abreast of it all. One way to learn about new information is to rely on recommendations given by others to guide your browsing choices, but this extra effort often proves impractical and unreliable. As a result, many systems use information about user browsing patterns to generate these recommendations. As a simple example, some Web sites monitor the number of hits to pages and generate a "Top Ten list" of the most popular sites. In more complex systems, efforts are made to generate and match user profiles to the information in the Web pages so a user can see which sites people with similar interests have visited.

While there are a number of text-based systems for passive browsing, they typically ignore the rich graphical side of the Web. This project will look at ways in which embedded graphical images in Web pages can be used to guide browsing. Specifically, it will examine whether graphical images viewed by others can help in the selection of interesting Web sites. Consider that you and a group of your colleagues have similar interests. If you had access to a list of the Web pages they visit, you might be able to find pages that interest you as well. Viewing the images from these pages might provide a faster and less cognitively taxing way to browse the list.

Your task for this project is to design and implement a system that will collect the images viewed by a set of users and display them to the current user, who then can visit the corresponding site if an image appears interesting. Two important considerations in designing your system should be when and how to display the images. For example, your system could be a standard application that must be explicitly executed by the user, or a screen saver that is executed during idle times, or a desk accessory that is continually executing and displaying new images. Your choice for when to display the images will probably guide how the images are laid out on the screen and how the user controls the layout. You could display a single image at a time, or a collage of images, or a thumbnail of each image.

After implementing the system, it is important to observe how users interact with your system. Find a few volunteers who will use your system, and observe how they interact with it and what problems they have with it. In your writeup, discuss how you would improve the system based on the user reaction.

Background
The following projects will give you a flavor of the works that is going on in this area. You may want to search the Web for similar projects that interest you.

• The Grouplens Project at the University of New Mexico is an example of a system that uses recommendations of others to guide a user's browsing choices (in this case, Usenet news browsing).
• Different paradigms for displaying collected information (particularly graphical images) have been used by PointCast, Andruid Kerne's CollageMachine at CHI 97, and Jonathan Helfman's passive browser at HCIC 96.

Deliverables

• A pointer to the source code for your passive browser.
• A writeup (3-5 pages). Be sure to address the following items:
  • The execution paradigm used and why it was chosen.
  • The layout for the graphical images and the way in which a user accesses them.
  • The platform and language used.
  • User reactions to your system.

Evaluation
Evaluation is based on the implementation of the browser and the contents of the writeup.