Powerset Demo Video from officialpowerset on Vimeo.

Powerset has launched a public beta of their new natural language search engine. What exactly do we mean by natural language search? First of all, you can enter keywords like you would with any other search engine. But Powerset can also handle phrases and questions. But Powerset's semantic tools go far beyond that. For example, if you search for "paintings by Dali," the search engine will understand that you are looking for paintings, and if it can find some images, it will put them at the the top of the page.

Right now Powerset searches exactly one site. But since that site is Wikipedia, you can still find a ton of useful information. But you're going to have more luck with queries like "who shot Lincoln?" than "when will the 3G iPhone be released."

When you click through to read a Wikipedia article, Powerset will bring up a sidebar tool that lets you view either an outline of the article or "Factz," which are typically some of the key points in an article. If you click on a fact, you'll be taken immediately to the relevant part of the Wikipedia article.

It's not clear whether Powerset is a Google killer, since the search engine currently doesn't index the web, just one site right now. So it's not clear how well it would be able to prioritize data from millions of pages. But it works quite well as an advanced search tool for a single site like Wikipedia.

[via GigaOm]

Radar Networks has expanded its private beta of Twine, a social networking, bookmarking, and discussion site built with semantic web technologies. If that sounds a bit confusing, let's break it down a bit. Twine lets users create "twines," or web pages around a particular item, whether it be a web page, a generic topic, an idea, or a person. You can create a twine for yourself, your blog, or a concept like "web 2.0."

Other Twine members can join public twines or any twine they've been invited to share. Once you're a part of a twine you can add comments, add links or share items with another twine.

So where does the semantic web part come in? While you can add tags to items you submit, Twine uses some intelligent features to locate people, places, types of items, and other tags that pop up in your twines. Over time, it develops a sense of the items you're interested in and will start recommending new twines that you might want to join.

For more info on Twine, check out our interview with Radar Networks' Nova Spivack. Twine is still in an invite-only private beta, with a public beta launch scheduled for later this year.

There's been a lot of talk over the last few years about the semantic web. While Google and other search engines do a pretty good job of ranking web pages and providing relevant results to keyword searches, the truth is that Google doesn't really understand what you're asking it. Not in plain language anyway. We've learned to tailor our search habits to search engines as much as they've learned to adjust to our preferences.

The concept of a semantic web is based on the idea that machines can be taught to understand language more like the way people understand it. For example, a semantic web search engine would know the difference between John F Kennedy the person and John F Kennedy the airport. But the truth of the matter is that Google does a pretty good job of meeting your search needs. And that's why we were interested in speaking with Nova Spivack, the CEO and Founder of Radar Networks, one of the companies pioneering the Semantic Web.

Today Spivack is announcing Radar's first product. Twine is a "knowledge network," built on a semantic web platform. Basically you can think of Twine as a cross between Facebook and Google, with a little bit of del.icio.us thrown in for good measure.

We asked Spivack why Radar would focus on a social application rather than a search engine, and here was his response:

"Google's mission is to organize the world's information. And our mission is really to organize your information. Your information is really the most important information in the world. It's the other 90% of the information that Google isn't indexing really. I think that Google is doing a wonderful job of crawling the web. And in fact it's something that we plan to leverage. Certainly I don't think that it would make a lot of sense for any company today to directly compete with Google. And in fact, there's lots of opportunity in areas that Google isn't really strong at. And this is one of them."