Jennifer Zaino
SemanticWeb.com Contributor

Alex Iskold, founder and CEO of smart browser and personalization vendor AdaptiveBlue, has some problems with the idea of the classic semantic web. They boil down to something like this:

1. It lacks memory and is not iterative in nature.
2. Its ultimate goal is to deliver perfect answers, which are unattainable.
3. It is technologically impractical to achieve.

More From Jupitermedia

The Semantic Web and Your Intranet The Business Case for Semantic Web The Semantic Web: Are You Scared Yet? BSM and the Semantic Web: Parallel Courses If you want to comment on these or any other articles you see on Intranet Journal, we’d like to hear from you in our IT Management Forum. Thanks for reading. - Tom Dunlap, Managing Editor.

This is not to say that Iskold doesn’t believe in the possibilities of semantic technologies. (For more on Iskold’s views, see his column on Semanticweb.com, “The Road to the Semantic Web”.) His main argument is with the “bottom-up” approach in which each web site needs to annotate information using RDF and OWL formats in order for computers to understand it.

He’s in favor of a “top-down” approach, evident in AdaptiveBlue’s smart browsing technology, which applies specific, vertical semantic knowledge leveraging existing information on the web to searches of various categories. It’s a noun-verb equation — for instance, in the books arena, the specific thing you might want to do is search for a book, find similar books, find those by the same author, add it to a wish list on Amazon. Or for restaurants, you might want to find them on a map, find similar ones in the neighborhood, or reserve a table.

“Given an object there is a relatively small set of meaningful things we do over and over on the web,” says Iskold. That is the essence of smart browsing, he argues, and you don’t need the semantic web and its formats to make that happen.

AdaptiveBlue encodes metadata in XML format that is relevant to some 20 consumer categories (book ISBN numbers, music genres, etc.), and uses algorithms to perform relevant sets of actions in each category, and personalizes shortcuts to results based on the browser history it collects when installed (such as, that you are a Barnes & Noble vs. an Amazon buyer).

“One angle you can take is that it becomes a world wide vertical search,” says Iskold, “And it leverages memory in a simple way to create a personalized experience.”

It’s that lack of memory that Iskold calls a “huge missing bit from the semantic web story.” A famous example of where the semantic web will take us is to book a user the perfect trip for a budget-minded family with small children, but “the big fallacy in this thing that people miss is that this assumes the computer has a memory of who you are,” Iskold said. Logic and inference are just part of the story. Not to mention the problem of “anti-complete” problems, which are computationally expensive and inherently unsolvable, because there are far too many pathways and too many possibilities to compute to find the “right” answer, he says. “The problem of figuring out the perfect vacation location for an individual — the perfect answer — is unattainable, and it’s a ridiculous way of answering the question.”

In terms of its ability to help you make connections, Iskold says 30-year old relational database technology and web service interfaces can manage that. And while he admits it’s exciting to think about connecting all these bits and pieces of information seamlessly, “the big problem is the Tower of Babel. There’s the famous problem between two companies trying to exchange information. The classic semantic web says there will be a centralized generic repository where all this mapping is maintained. In the computer industry, this has never been done, and so I am skeptical this will ever be done. All integrations are always done on the B2B level. Imagine maintaining this centralized repository of meaning and mapping, so how could that possibly work? Mapping between different services — the way it is described — is unnecessarily complicated and unachievable for many purposes.”

In Iskold’s eyes it’s about being visionary but pragmatic. The classic semantic web hasn’t mastered the second half of that.

“I am a very big believer in a simple and pragmatic approach to problems. The semantic web is already 12 years in the making, and no real end user tools are available,” he says. “We are 1.5 years old, and we have semantic technology out there on the market and close to 1 million downloads.

“People use it every day and it slices through the problem of recognizing meaning and having utility to the end user and helping them get things done faster,” Iskold said. “That is the kind of proof of value we need to do before we demand that people convert their stuff into RDF.”

SemTechBiz is Less Than 2 Weeks Away

The Semantic Tech & Business Conference (SemTechBiz) is coming to San Francisco on June 3-7! Join us for case studies, innovative panels, tutorials, and keynotes that will provide you with practical advice, hands-on guidance, and breakthrough approaches to solving business problems with semantic technology. Passes go up $200 at the door. Sign up now and save !