Skip to Content
Uncategorized

Better Search in Virtual Worlds

Second Life’s new search tool helps users find 3-D objects.
November 6, 2007

As Google’s stock price can attest, good search engines are what make the Internet useful and entertaining. But in virtual worlds, which are made up of not text but 3-D renderings of people, places, and objects, the search problem is harder. Residents of Linden Lab’s Second Life have long been able to perform large, general searches–for a listing of a Spanish-language event, for example, or for the location of a particular group. But while it was possible, with a little effort, to find a shoe store, there was no good way to locate a pair of red shoes. Since much of the entertainment value of virtual worlds lies in creating objects and trading them, that’s been a frustrating limitation. But Linden Lab recently released a new search tool that begins to address the problem.

Getting specific: A new search tool for Linden Lab’s Second Life allows users to look for individual 3-D objects and for information in each other’s profiles.

Building on a commercially available platform from Google, the new tool gives Second Life residents search results ranked by relevance, rather than ordered alphabetically or according to how much traffic they generate, as was previously the case. Jeska Dzwigalski, who is in charge of community and product development at Linden Lab, explains that the tool also allows residents to search much more specifically. In addition to being able to search for objects, residents can now look for information–about hobbies, for example–in each other’s profiles. Dzwigalski says she expects that being able to search profile information will improve Second Life’s social features.

The algorithm behind the new Second Life search tool will resemble Google’s: found objects will be ranked according to how well the data used to describe them match the search terms entered, how close multiple words are to each other, and how popular the objects seem to be, based on the frequency of references to their locations. Like the existing search tool, the new one will allow people to select whether or not they want to include mature content in their searches. The current release, Dzwigalski says, lets Second Life subscribers decide which of their elements they want the search engine to index, so that they will have a chance to determine their desired level of privacy. “The new search tool allows people to search more things and better describe them,” she says.

Linden Lab’s new search tool for Second Life, shown above, goes beyond the alphabetic and traffic-based ranking systems that were previously available; instead, the tool sorts results by relevance. Users can decide whether or not to receive results that include mature content.
Credit: Linden Lab

Before Linden Lab announced its new tool, third-party companies, such as Electric Sheep, were working on their own to improve search in Second Life and other virtual worlds. “The search capability in the worlds has been historically quite basic,” says Giff Constable, who leads the Electric Sheep’s software business unit. Constable says that his company was sending bots into Second Life to pick up virtual objects and extract data from them in order to compile search results. “The analogy would be to Alta Vista in the early days of the Web, before Google came around and became able to rank things for popularity,” Constable says. He adds that his company hopes to take advantage of the new search tool from Linden Lab and will focus on providing additional tools for social networking and e-commerce.

But both the new tool and Electric Sheep’s bots are dependent on tags labeling objects and places; there is no virtual-world search technology that can recognize objects without those clues. Constable notes that, when items aren’t tagged, the best his bots can come up with is “object.” “We can find them, we can see them, but we can’t make sense of them,” he says. Dzwigalski says that, while the new Linden Lab search tool filters out the useless list of “objects,” it’s no better at identifying them. As a result, residents are responsible for tagging items that they want indexed. Constable says that his company hopes to add tools that will let people label their items more descriptively and effectively. “People are looking for a very wide range of experiences and things, and it adds some complexity to the search problem,” he says. He points out that people need to be able to specify whether, for example, they want to buy a virtual replica of the Eiffel Tower or visit a virtual reconstruction of it.

But the search problem has consequences beyond whether or not users can shop efficiently within a single world. As developers work to let users pass easily from one virtual world to another, the problem of searching those worlds is becoming more important. (See “Moving Freely between Virtual Worlds.”) “This is the same sort of thing that we dealt with in the early days of the Web,” says Corey Bridges, cofounder and executive producer of Multiverse, a platform for virtual worlds. “If we’re enabling a whole network of virtual worlds, it’s critical for the user to find just the world that he or she wants. Odds are, it’s going to end up like the Web, where there’s a whole bunch of stuff out there, and you don’t want 98 percent of it.”

Although Linden Lab’s tool represents an improvement, it is far from solving the problem. “Search is important, but I personally don’t know of any answer in that space yet,” says Michael Rowe, manager of 3-D Internet and “intraverse” research at IBM. “I would say search [in virtual worlds] is going to be pretty exciting as it matures, since you’re not only looking for a link: you’re looking for a thing within a context of space and time.”

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.