The Problem with the Internet

Page and Brin were always together. On campus, they became known as LarryandSergey. The banter between the two knew no limits. They loved challenging and debating each and anyone else they could suck into a good argument. They talked endlessly about computers, philosophy, and whatever else popped into their minds. Once they argued loudly about whether it was possible to construct a building-size display out of lima beans. Under Page's desk, they built a computer rack out of Legos. The others in their office found it virtually impossible to get any work done without tuning them out.

Rajeev Motwani, Sergey's advisor, watched the intellectual relationship grow between Brin and Page. "They were both brilliant, some of the smartest people I have ever met," Motwani said. "But they were brilliant in different ways." Brin was practical, a problem solver, an engineer. If something worked, it worked. Page, on the other hand, was a deep thinker. He wanted to know why things worked.

Brin had been working closely with Motwani on "data mining," finding ways to extract information from large mountains of data. Data mining can be used by retailers, for example, to see what combinations of items customers purchase in stores so they can arrange products better. Brin and Motwani experimented with applying the same techniques to the new, disorganized Internet.

In the mid-1990s, the Web was a virtual Wild West-unregulated, uninhibited, and unruly. Millions of people logged on and began using email, but serious researchers grew frustrated amid the clutter of Web sites. Early efforts at "search"-programs to help users find information on the Internet-fell short.

Jerry Yang and David Filo, another pair of Stanford graduate students, tried a new approach to search. Rather than relying on technology alone, they employed a team of editors who selected Web sites for an alphabetized directory. They called their company Yahoo! Although their approach simplified finding valuable information, it could not keep up with the fast growth of the Web.

Brin and Motwani tried Yahoo! and other directories and search engines, but nothing got the job done. Instead, a simple search would yield hundreds or thousands of results in no apparent order. It took them hours to sift through the pages to find whatever they were seeking. Brin and Motwani became convinced that there had to be a better way to search the Internet.

At the same time, Page began hunting around the Web using a new search engine called AltaVista. While it returned somewhat better and faster results than the other search engines, Page noticed something else entirely. In addition to a list of Web sites, AltaVista's search results included seemingly useless information about something called "links." Computer users seeing a link-a highlighted word or phrase-could click on that link if they wanted to learn more, and they would instantly be taken to another Web page. Instead of focusing on AltaVista's main search results, Page wanted to analyze links and see how they might be used further.

But to test any of his theories, Page would need a big database. Always ambitious, he quickly did some calculations and then told his startled Stanford advisor that he was going to download the entire World Wide Web onto his desktop computer. On its face, Page's idea seemed absurd. He even declared that downloading the Web could be done fairly easily and quickly. While others scoffed, Page was deadly serious, and on a mission to capture nothing less for his research.

Larry's Big Idea

As 1996 wore on, Page and Brin teamed up to download and analyze Web links. It took longer to get the data than Page had envisioned, but he desperately wanted to see it through. His drive to discover the importance of these links also attracted the attention of Brin's advisor, Motwani, since it held out the promise of improving Web research. Brin was drawn to the project by the chance to work with Page and by his own interest in mining giant amounts of random data.

From his research, Page developed a theory: counting the number of links pointing to a Web site was a way of ranking that site's popularity. While popularity and quality don't always go hand in hand, he and Brin both had grown up in homes that valued scholarly research published in academic journals with citations (references to others' work). Links, in a sense, reminded Page of citations. Scientists would cite the published papers their work drew upon, and these citations were a helpful way of tracking credit and influence in the academic and research communities. A large number of citations in scientific literature, he said, "means your work was important, because other people thought it was worth mentioning."

The same could be said for Web sites, Page concluded. Taking things a step further, he had a breakthrough: All links were not equal. Some mattered more than others. He would give greater weight to incoming links from important sites. How would he decide what sites were more important? The sites with the most links pointing to them, quite simply, were more important than sites with fewer links. In other words, if the popular Yahoo! homepage linked to an Internet site, that site instantly became more important. Playing off his own last name, Page began calling his link-rating system "PageRank."

Brin and Page believed they had found the path toward a PhD thesis. By early 1997, Page had developed a primitive search engine that he named "BackRub" because it dealt with the incoming-or "back"-links to Web pages. Page, Brin, and Motwani all contributed ideas to the evolving project. Without intending to, the trio had devised a ranking system for the Internet, and in the process had solved one of the core problems of searching for information on the Web.

Brin, Page, and Motwani put together a prototype of their search engine for use at Stanford, combining traditional search engine technology with PageRank. While other search engines relied on matching words in "queries"what a user types into the search box-with words on Web pages, PageRank had an extra dimension: it put search results in a logical order. For the first time, there was a way to do an Internet search and find useful answers swiftly.

Becoming Google

In the fall of 1997, Brin and Page decided that the BackRub search engine needed a new name. Page asked his office-mate Sean Anderson for help coming up with a catchy name that hadn't already been taken. For days, Page rejected every one of Anderson's ideas. Finally Anderson suggested "Googolplex," saying, "You are trying to come up with a company that searches and indexes and allows people to organize vast amounts of data. Googolplex is a huge number." (A googol is a 1 followed by a hundred Os; a googolplex is a 1 followed by a googol Os.) Page liked that, and suggested shortening it to "Googol." Misspelling it, he registered the domain name Google.com. By the time the pair realized the spelling mistake, it was too late to change it.

Lacking the funds to hire a designer and the artistic talent to create something elegant, Brin kept the Google homepage simple. From the start, Google's clean, pristine look attracted computer users. Its primary colors and white background stood in marked contrast to the growing number of busy-looking Internet pages with flashy ads and crowded graphics and type. Because it didn't feel as though Google was trying to sell anything, people readily adopted the search engine as their own. Around the Stanford campus, Google's popularity grew by word of mouth. It quickly became the favorite search engine of Stanford professors and students.

As the database and user base grew, Brin and Page needed more computers. Short of cash, they saved money by buying parts, building their own machines, and scrounging around for unclaimed computers. After cramming as many computers as they could into their office, they turned Page's dorm room into a data center. They learned how very much they could accomplish by assembling and stringing together inexpensive PCs, a method Google still uses on a larger scale today.

The Next Step

In March 1998, Page and Brin talked to Paul Flaherty, a Stanford PhD and an architect of AltaVista, about their search engine technology. AltaVista, they hoped, would pay as much as $ 1 million for the soon-to-be-patented PageRank system. It would improve AltaVista's search results, and Brin and Page would then be able to resume their studies at Stanford. AltaVista, they said, was just the beginning; Google was the future.

Flaherty agreed that the guys had a cool concept. He also warned them that problems happen on the Internet after you become popular; people try to break into your network or attack your site. But Page and Brin weren't afraid. Instead, they brimmed with confidence and wanted their technology in the hands of more people.

A few weeks later, though, Brin and Page heard that AltaVista was turning them down. Search was not a priority for AltaVista; it was only one of numerous offerings they would provide Internet users in addition to news, shopping, email, and more. Brin and Page tried unsuccessfully to sell their PageRank system to other search engines. It didn't seem to matter that they had something better. Everyone around them seemed to be focused on selling as many ads as they could.

Yahoo!, seemingly a logical buyer because it relied on directories edited by people and didn't have a fast way to scour the entire Internet, also turned town the chance to buy the Google technology. Part of the reason was that the Google search engine was designed to give people fast answers to their questions by sending them to the most relevant Web site, but Yahoo wanted users to send more time on the Yahoo site, where they could shop, view ads, check their email, and play games.

Yahoo! cofounder Davd Filo advised Brin and Page that if they wanted to realize the potential of their unique search system and believed in it, the best thing for them to do was to take time off from Stanford and start their own business.

If it was as good as they claimed, it would catch on.

© 2008 Carus Publishing Company Provided by ProQuest LLC. All Rights Reserved.