Contextual Web History

A web browser history that improves the visibility of the history feature, and makes use of the visual appearance of the web page as well as the user's working context to help people find previously visited web pages.

Full paper accepted at ACM CHI 2009 and awarded IBM's Innovation that Matters Award in 2008.
Co-project lead & Academic researcher
Carnegie Mellon University
Contextual Web History, browser research project from Carnegie Mellon University
Self-started independent research project. Co-project lead & Academic researcher with Jing Jin / Carnegie Mellon University.
Problem space
Even though people browse the web in a non-linear fashion, the web browser histories at the time were organized by recency. As a result, people had trouble finding some of the websites they have been to before. We were motivated to help users find and return to webpages they've seen before.

I worked with another student and our faculty advisor to conduct user studies, design mockups, and implement live prototypes.
User studies: Recall & Recognition
Contextual Web History: Recall and recognition
The purpose of this user study was to understand what aspects of web pages were memorable to users. User study participants were emailed a list of websites several days prior to the study.

For the Recall part of the study, participants were asked to describe or draw about a subset of previously visited web pages. During the process, they were asked about specific characteristics of each page they visited. Participants were also prompted to describe the web pages so someone can identify and recognize the page based on their descriptions.

For the Recognition part of the study, participants were asked to identify the thumbnails of the web pages they visited among eight to twelve thumbnails and explain how they chose those thumbnails. We wanted to evaluate visual thumbnails would aid in revisiting web pages.

From the study, we were able to find a list of more memorable and computable cues about a web page. The participants cited large text or images, the page's layout, and the colors as clues in the correct thumbnail.
Designs & Prototypes
We iterated through several designs that allowed a user to specify a query with time of day, general layout, category of topics, and any tags previously added, and search terms. However, participants were daunted and confused by the abundance of options. There is also the technical difficulty of determining the layout and category of web pages, so we simplified the design to searching and showing thumbnails for our software prototype.

The prototype was built as a Firefox extension, a popular web browser with add-on capabilities, and was working and functional with real data.

Key features included
  • Search & Browse of web history integrated into sidebar
  • Easy and visible access from toolbar search input
  • Thumbnails provided for each visited website
  • Full-text search
  • Contextual Web History results embedded into search engine results
We combined several relatively straightforward ideas for helping people find the history feature, browse and search through history, see relevant information about past pages, and quickly find previously visited web pages. We developed and evaluated two prototypes of CWH, and found through user studies that CWH is effective for helping people find web pages again.

For details on our methods, user studies, and the prototypes, check out our paper published by ACM CHI.
Contextual Web History: Using Visual and Contextual Cues to Improve Web Browser History / Full paper // ACM SIGCHI 2009
Contextual Web History: Screenshot of the prototype
  • IBM "Innovation that Matters Award": May 2008
  • "Thought" Prize for Excellence in Research Presentation: May 2007
  • Research Grant (SURG): January - December 2007