…
Can the Internet be chartered like the London Underground? Leander Kahney investigates mapping cyberspace
A three-dimensional map of the Internet, showing the geographical location of machines on the network and the ways in which they relate to each other, is promised by the Silicon Valley firm Invisible Worlds next year.
The announcement has been greeted with some scepticism, but the company’s leaders have a strong track record: Marshall Rose invented the protocol that delivers e-mail, and Carl Malamud is the man who pressured the United States government into posting public records on the Internet, the founder of the Internet Multicasting Service – the Net’s first radio station – and the organiser of the Internet 1996 World Exposition.
Malamud says Invisible Worlds arose from a dissatisfaction with the Internet’s current navigational tools, search engines and directories like Yahoo and Snap: these do not convey any sense of the Internet’s vast global network of computers or the way they communicate with one another.
Sometime next year, he says, Invisible Worlds will deliver “a toolbox” based on a “flexible, open architecture” that can be used to create a range of 3D Internet maps. He is reluctant to go into detail. “We’re focused on impressing investors, not the public,” he says.
Philip Worob, vice-president of sales at Matrix Information and Directory Services of Austin, Texas, is dubious: “Even a site map can be a nightmare to create. The Internet grows geometrically every year. Creating a map for that would be incredibly difficult.”
Others have tried and failed, says Avi Rappoport, principal consultant with Search Tools Consulting of Berkeley, California. A major stumbling block is trying to convey three dimensions on a flat computer screen. “I’ve not seen many instances where they’ve found the right visual metaphor,” she says.
The other problem is categorising information on the Net. “Look at Yahoo: defining relevance, closeness and the concepts in text is very, very difficult for humans, for experts to do, and nearly impossible to do automatically. It’s like voice recognition. It’s a great idea but it’s just not there,” says Rappoport.
Martin Dodge, a cybergeographer with the Centre for Advanced Spatial Analysis at University College London, says: “[Malamud] is a guy with a track record, so I’m sure he can deliver and he’s been doing his homework.
“But they’ve been vague about it. If they are going to create a useful map, people want to see what it’s going to look like.”
Mapping the Internet is like any other mapping problem, he says. There will be no single, definitive map, but a range of maps for different purposes. “It’s like maps of the US. There are maps of roads and rivers, of mountains, of topography, of soil types. There’s no one map of everything in the US.”
Although there are many instances of good maps of individual sites, says Dodge, a map of a complex network is much harder to create. “The best analogy is the London Underground map. It’s a classic: a map of an incredibly complex underground transport network that people use to get about London. And it’s not particularly tied to the geography of the city.
“That’s the kind of map people need: something that tells them that to get from A to B you follow a big blue line.”