About.com Co-founder Looks to Challenge Google in Search

Google has a potential search rival in Dorthy.com, a Web site that takes a different algorithmic approach to search. About.com co-founder Jim Anderson has been named Dorthy.com's new chief technology officer, and will use his experience in AI and natural language processing to try to refine the new site's search process.

Google may not be exactly shivering in its boots, but its search engine has a new challenger in Dorthy.com, a Web site that utilizes natural language processing and semantics to refine the search process. About.com co-founder Jim Anderson has been named Dorthy.com's new CTO.

Dorthy.com's search engine works by having the user input not a series of keywords, but an entirely formed question or statement such as, "I want to upgrade my data center."

"Keyword searches do not work; asking a question and getting an answer works," Anderson said in an interview. "So looking at natural language processing and semantics is what we heavily focused on. We looked at inflection and semantics, and applied a learning algorithm."

By having the user ask a question in its entirety, as opposed to using keywords, the site can supposedly take advantage of advances in natural language processing, a field of computer science that Anderson says has become considerably more robust over the last two decades.

So, in a hypothetical example, if an IT director inputs, "I want to search for virtualization conferences in Las Vegas," Dorthy.com will analyze that sentence and see that, semantically, much of the emphasis is on "virtualization conference" due to its placement in the sentence. It will also make note of "Las Vegas," and because you've set up a user profile, it will know that you live on the East Coast.

Your results, theoretically, would not only list upcoming virtualization conferences in Las Vegas but also present travel options for getting there.

Click here to read about how Lucid Imagination and other startups have been working to create open-source search engine technology and bring it to the enterprise.

According to the company, one of the main benefits of Dorthy.com's search apparatus is that it makes search more efficient by reducing the number of times a user will have to repeat different variations on keywords, or repeat searches altogether, in order to achieve a satisfactory result.

Dorthy.com will also allow members to network into a full online community. The site will store its users' search results, connections with other members and other data, and use that information as feedback to better target and refine subsequent searches.

While the site will initially be consumer-targeted, its creators insist that its search engine model poses substantial benefits to the enterprise.

Search continues to be a hot topic for enterprise users and developers, particularly when it comes to engines that will hunt within organizations for information, and big players such as Google, IBM and Microsoft have recently entered that particular arena. In January 2008, Microsoft purchased Oslo-based Fast Search & Transfer, which specialized in enterprise search, for $1.2 billion.

Dorthy.com's beta launch will happen at an unannounced date in the near future; the site's general rollout is planned for April 1. The service will be free, with revenue for the company coming from advertiser placement in search results.

The company's officers are already looking beyond purely site-based search into the realm of mobile technology, and plan to make the service available on handheld devices at an as-yet-undetermined point.

"Our biggest thing is looking at mobile as our future," Anderson said. "Being tethered to a laptop isn't necessarily the best thing."

"From a technology standpoint, if you can get people to write full questions, then the search engine has many more clues. With good NLP [natural language processing] search, it digests information just as a person would," Susan Feldman, an analyst at IDC, said in an interview. "Good NLP I haven't seen on the Web because it's so hard to scale. But it's time it was out there."

Editor's Note: This article was updated with commentary from an analyst.