SAN FRANCISCO - For more than a year, a team of computer industry entrepreneurs has been using a roomful of computers to copy every World Wide Web page they can find.
Their massive digital library originally was viewed as a quixotic venture to assemble an Internet time capsule of the global computer network's early days.
But the electronic librarians appear to have found a more immediate - and potentially profitable - use for their 2 trillion-character pile of data: a new and precise way for ordinary computer users to navigate the sprawling Web.
To find what they're looking for on the Web, most computer users plug "keywords" into search engines - Web sites, with names such as Infoseek and Excite, that look for other Web sites containing that keyword and then generate a list of sites. That can produce a long list to wade through, and the search engines often miss relevant sites because the keyword is slightly off target.
But the Internet Archive's service - called Alexa, after the ancient library at Alexandria, Egypt - doesn't rely on a word search.
Instead, Alexa, a software add-on for Web browsers that can be downloaded from Alexa, uses several powerful computers to crunch through the archived data - a process called data mining - and look for patterns within the Web, ultimately providing a few recommended strands to travel.
From United Parcel Service's Web site, for instance, Alexa points the user to sites for Federal Express, the U.S. Postal Service and DHL Worldwide Express. Starting at the Web page of catalog clothing retailer Lands' End, sites for J. Crew, L.L. Bean and Victoria's Secret are suggested.
"Sometimes it's spooky how helpful it can be," said Brewster Kahle, the archive's founder and the president of Alexa Internet Inc., a San Francisco start-up company that intends to offer the new navigation service for free to Web users.
This is not the first time someone has claimed to have cracked the code for making the Web easily searchable. And it's not clear whether Alexa, which is being tested by about 10,000 people now, would function well if millions of people tried to use it. But in its early form it's getting generally upbeat reviews in the computer industry.
"Search engines can drown you in documents," said Jerry Michalski, managing editor of Release 1.0, a New York-based computer industry newsletter.
Asking the Infoseek engine to find sites with the keyword "shipping," for example, generates more than 291,000 citations, with sites for a British port and an exotic-bird store at the top of the list.
"Alexa is a big step forward," Michalski said. "It's like having a huge associative memory that's looked at most of the Internet."
© 2017. All Rights Reserved. | Contact Us