Finding Pages from Browser History
A new tool aims to make a Web browser’s history more useful.
Web browsers remember the sites that they have visited in the past, but few people seem to use this information. Jing Jin, a graduate student at Carnegie Mellon University, has developed a new browser-history tool, which she and her colleagues developed after studying how people use their browser history. They demonstrated the prototype in a presentation this week at the Computer-Human Interaction (CHI 2009) Conference, in Boston.
“Despite the fact that bookmarks and Web histories are built into every browser, we found that people tend to either use search engines or reconstruct the path by which they got to a page in the first place,” says Jason Hong, an assistant professor at Carnegie Mellon, who was involved in the research. “Most people either found Web history too hard to use or didn’t even know that it existed.”
The researchers tested users’ ability to recall Web pages and found that URLs and textual descriptions (by which most browsers organize their history) weren’t as easy to remember as colors or images collected from the Web pages themselves. So the researchers’ tool–currently a plug-in for the Firefox browser–lets users browse images of websites that they have visited in the past, or type in search queries that find previously visited pages.
The researchers also used the new history tool to improve Web search, by adding thumbnails from browser history at the top of Google search results. The thumbnails were selected according to the search terms that the user entered into the search engine.
In testing, the researchers discovered that people could find the page they were looking for within about a minute on average using the prototype add-on, compared with an average of three minutes using the standard browser history. The user tests also showed that people were able to actually find a given old page more often with the prototype.
Eytan Adar, a PhD student at the University of Washington, is also researching ways to redesign the browser’s history. He says that the current design doesn’t match the way that people revisit pages. In collaboration with scientists from Microsoft Research, Adar found that people tend to visit Web pages in certain patterns. He says that it might be possible to split the history so that it treats these pages differently according to the user’s patterns. For example, those pages that people revisit over and over in a short period could appear in a temporary “working queue,” Adar says. The pages that they access regularly could be displayed in a more persistent list, while the pages that they revisit rarely might be accessed through a search function.
However, some experts argue that the browser history is less broken than these researchers suggest. Larry Constantine, a usability expert and professor at the University of Madeira, in Portugal, notes that some browsers already make sophisticated and hidden use of history information. For example, the Firefox 3 browser is good at guessing URLs from a user’s history based on keywords entered in the address bar. Whether the browser history is useful, he says, “depends a lot on which version of the browser people are using and on their personal habits.”
Jared Spool, founding principal of User Interface Engineering, a consulting firm based in North Andover, MA, says that people tend to revisit pages by retracing the way that they found them in the first place. “It’s sort of a cliché, but we are creatures of habit,” he says.
If the first path to the information worked well and quickly, it’s not natural to seek out a second path, Spool says. He’s not sure that most users would change their behavior even if the history were better designed, but he sees potential for using it for specific purposes. For example, he says, authors might want to use an improved browser-history tool to revisit research resources, or a company’s employees might use such a tool to help them navigate a poorly organized intranet, which might contain material that’s harder to search than is content on the wider Internet.
Spool says that the ideas behind the prototype history tool are likely to filter into consumer products in a very different form. “What we’re seeing here is the first piece of the pollination process,” he says.
Indeed, the Carnegie Mellon researchers point out that browsers are already starting to explore alternate ways to use data from people’s browsing habits. For example, Google’s Chrome browser features a “speed-dial” page when a user opens a new tab that shows thumbnails of frequently visited websites. Carnegie Mellon’s Hong notes that a redesign of the browser’s history could be particularly helpful for less Web-savvy users, who might have trouble figuring out the steps of the path that they originally took to a piece of information.
Keep up with the latest in intelligent machines at EmTech Digital.
The Countdown has begun.
March 25-26, 2019
San Francisco, CA