Public Release: 

New Software Robot Uses UW Technology To Make On-Line Shopping A Breeze

University of Washington

Frustrated Internet shoppers who are unable or unwilling to wade through a flood of World Wide Web sites to find what they're looking for soon can call on a computer software robot named Jango to help with their on-line shopping. Using technology developed by University of Washington computer scientists, Jango automatically searches the web and quickly compiles a report listing vendors, prices and reviews for a desired product. It can even initiate an order. A Beta version of Jango has been released by Seattle-based Internet software developer Netbot, Inc.

At the heart of Jango is a novel web navigation technology developed by Oren Etzioni and Dan Weld, computer science and engineering professors at the UW and co- founders of Netbot. Etzioni and graduate student Erik Selberg originally designed the technology for an Internet search tool called Metacrawler that recently was voted the best search engine on the web by InfoWorld magazine. Metacrawler uses a "parallel pull" technology to simultaneously probe six web search engines like Alta Vista and Yahoo! and prioritize the results that most closely match the user's inquiry.

"Most of us don't have the time, know-how or desire to search all of the available web sites ourselves to gather the information we're looking for," Etzioni explains. "Our softbot (software robot) technology does most of the work for you by searching the overwhelming mass of information available on the Internet and presenting a manageable amount of high-quality information. It rewrites the book on searching the Internet."

In conventional web searches, users type in a web address or click on a link and their browser retrieves and displays a single web page. The UW's parallel pull approach uses artificial intelligence to search and retrieve information from dozens of web sites simultaneously while filtering out irrelevant information.

Working in concert with a user's web browser such as Microsoft Internet Explorer or Netscape Navigator, Jango employs the UW softbot technology to greatly reduce the time and effort required for Internet shopping. A user types in the name of a product -- or even some piece of information about a product, such as an author of a book -- and Jango automatically determines which stores and information sites are relevant. It consults those sites and prepares a report including detailed product information, comparative reviews, prices and manufacturers' specifications. The process usually takes less than a minute.

If shoppers find something they want to buy, Jango can automatically fill out the order form with information provided by users during the original registration process. For those concerned about sending credit card information over the Internet, Etzioni says such transactions are at least as safe as giving a credit card number to a restaurant waitress or a telephone operator. The credit card information is encrypted and guarded by a password on the user's computer and likely has the same protections at the vendor's site, he explains, so it is only vulnerable during the transmission.

"It's a perception issue more than anything else," Etzioni says. "And even if people decide they don't want to buy over the Internet, Jango can provide all kinds of helpful, pre-purchase information."

So far, Jango is set up to facilitate shopping for products in the following categories: books, cigars, computer software and hardware, movies, music and wine. Additional categories such as airline tickets, consumer electronics and sporting goods will be added weekly until the full spectrum of consumer goods available on the web can be accessed through Jango

"I can't think of a single thing -- from groceries to mortgages -- that you can't buy over the web," Etzioni says. "It's just a matter of whether users can quickly and easily access the information they need to make good, informed purchases. For that, there's really nothing out there like Jango."

###


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.