The Google Stalker

Once upon a time in the late 1960s, Arnie Schultz decided to come up with a way to make words and numbers compatible.

Back then, Schultz was an English literature student at the University of Minnesota. A friend told him about a job at the university’s medical school to electronically index complex medical abstracts alphabetically and by category, figuring that Schultz’s language skills would be helpful on the project. Schultz got the job, and soon learned an assembly language program, a simple “human-readable” computer language. That would lead to something far more complex which would become Schultz’s life work: assigning a numerical “weight” to every word in the English language in order to perform more useful, contextually correct electronic text searches and summaries.

"It's amazing to me that these search engines process billions of words a day, and none of them are capable of saying, 'That's a noun, that's a verb.'"

Nearly four decades after the medical school tapped him for that indexing project, Schultz’s years of work in the deep structural underpinnings of the English language have culminated in a software program called ePrécis. Though its chief task remains indexing and abstracting, the technology also can be used to retrieve information on the Internet. Indeed, it may do nothing less than revolutionize how online search engines work. “We’re showing Google, and all of them, how they should do it,” Schultz asserts.

For the past few months, ePrécis has, in a sense, been stalking Google. EPrécis has been using a technique called scraping to demonstrate how much more accurate its own search results are—and to get Google’s attention. Scraping involves downloading a Web page and extracting some of its information, usually in order to place it on one’s own site. (A familiar example: obtaining your company’s stock quote from one site and displaying it on your own.) EPrécis sorts through Google results for a search request and purports to make them more accurate and—thanks to its abstracting ability—handier.

Now the Edina technology incubator that owns ePrécis is in licensing talks with the world’s most dominant search engine. The timing is propitious. Thanks to their exploding use as gateways to Web content of all kinds—and their growing sophistication in attaching relevant advertising to search-result pages—search engines are coming into their own as revenue producers.

And that could mean that Schultz’s lifelong project will pay off—big.