AUSALITO, Calif. -- By Silicon Valley standards, it was an unassuming debut.

Last weekend at a Marin County conference on biological and cultural diversity, Groxis, Paul Hawken's new software company, was sandwiched by the booths for Birkenstock sandals and PowerBars.

Mr. Hawken is best known for as a co-founder of Smith & Hawken, a California catalog and store that sells high-end gardening equipment, and for his writings, which blend environmentalism and capitalism.

Advertisement

But at a time when the valley's digerati are bemoaning a technology industry recession and the death of innovation, Mr. Hawken's Grokker software, which is intended to allow personal-computer users to visually make sense of collections of thousands or hundreds of thousands of text documents, is creating a buzz. The software is attracting significant interest from large corporations and universities.

At the conference, Groxis employees began taking orders after people told the company that they were willing to pay for the program, which had not yet been released; 200 copies were sold that weekend.

Groxis, based in Sausalito, is trying to solve an old problem, but it is growing exponentially with the explosion of electronic information brought about by the Internet.

"The computer world has been alphanumeric, but we perceive things visually," said Mr. Hawken. He said he is now working on a new book and has used his program to organize a collection of more than four million Web bookmarks. "It's the first time as a writer I have too much information, but it's manageable," he said.

Groxis (pronounced GRAHK-sis) is named for a term in "Stranger in a Strange Land," by the science fiction writer Robert Heinlein. To "grok" something is to understand it completely.

The reference is driven home by the fact that today, according to the Groxis designers, 84 percent of Web surfers go no further than the first page of document titles in searching for information. While that may be appropriate for some searches, other queries, like one for the phrase "exotic vacations," may leave Web sites of interest deeply buried by search engines.

The popular search engine Google uses ranking techniques to sort through thousands of possible Web pages displaying what are often the most desirable pages first.

In contrast, Grokker builds a visual map of the general categories into which documents fall by using what computer software designers call metadata, which describes each Web page or document. The program currently works with the Northern Light search engine, the Amazon online catalog and as a tool for scanning a user's own PC file collection.

Previously, there have been thousands of research efforts in text visualization and dozens of commercial attempts to go beyond traditional directory listing approaches for organizing information.

The Xerox Palo Alto Research Center did work in this area in the 1990's and spun off Inxight Software, a company in Sunnyvale, Calif., that now offers tools for the management of text data. Early on, Alta Vista, the first search engine, offered a so-called refine feature that permitted a graphical view of information returned after the query.

There are also Web search tools ranging from KartOO (www.kartoo.com) to Vivisimo (www.vivisimo.com) that provide alternative ways of viewing or categorizing the expanse of information accessible on the Web. Tim Bray, a designer of the XML Internet protocol, a commonly used format for storing and exchanging information on the Internet, has created Antarctica Systems (www.antarti.ca), a company based in Vancouver, British Columbia, that offers a geographically oriented visualization approach.

So far, none of the visualization tools have succeeded commercially.

"The universal reaction is that people say, `Oh, how pretty,' " said Danny Sullivan, editor of Search Engine Watch, a British Web site (searchenginewatch.com) that tracks the search engine industry. "However, they don't continue to use them. People forget that text is visual and reading a list is a very helpful way of organizing information."

The basic ideas underlying the Groxis technology were developed by Jean-Michel Decombe, a French computer researcher, who in the late 1990's worked for the Silicon Valley start-up Metacode, which was developing automatic categorization. When Metacode was acquired in 2000 by Interwoven, another Silicon Valley content management concern, Mr. Decombe joined with Mr. Hawken and R. J. Pittman, a computer scientist and venture capitalist, to acquire the visualization technology he had been working on.

Mr. Decombe argues that Grokker is a more universal approach to the problem of visualizing textual information than what has been found in previous tools, which focus more on navigation than on categorization.

"The difference is that we have no single paradigm," said Mr. Decombe, who is Groxis's chief technology officer. Instead, Grokker offers a wide palette of different ways to view information. Under Mr. Hawken's guidance, Groxis is also hoping to refine a business strategy that will make Grokker a widely used research tool.

The Grokker software, offered in versions that include a basic $149 program, will appear first as a preview beta version available at the Groxis Web site (www.groxis .com), according to the company.

At Stanford, the library has modified its online Socrates catalog system to be viewed by Grokker. "It shows sufficient promise that we're thinking of giving it to our entire community," said Michael Keller, the university librarian. "Potentially, it will not only give our students and faculty a way to view their own information world but also our wildly multifarious array of thousands of e-journals and hundreds of databases."

Grokker has also drawn the attention of Interface Inc., based in Chicago, a big carpeting company. In March, the company plans to begin a division, Interface Flor, offering a Grokker-based catalog that will permit Web customers to view, organize and select from 50,000 samples.