What's been the most surprising change since you wrote Wired Society in 1978?
The book was surprisingly accurate. Two things surprise me. One is how non-academic people talk to computers more than I thought they would. They may not be doing sophisticated things but they are using neat tools. The other thing is how fast the internet grew once it had picked up that initial positive feedback. The web was easy to use so the growth from 1995-2000 was phenomenal. I think we'll see something similar with the Grid. We're also seeing the growth of data warehouses. Along with them are various techniques for data mining and pattern recognition.
Could supercomputers be harnessed to help the environment?
The biggest problem is the pumping of gases into the atmosphere, which may trigger climate change. There are a few supercomputer models of the climate and they indicate it is metastable - you can only disturb it up to a point. The biggest supercomputer is Japan's Earth Simulator, which is working on models of climate. We need more research but lots of technologies could replace carbon-based ones.
3G?
There are many technologies by which you can do wireless broadband and 3G was just one . Other ways now look better. One problem was locking into one architecture before it was capable of dominating the market. We are beginning to see some of the things they said we would with 3G _ a lot of people have mobiles capable of taking and sending photos.
Does Moore's Law still hold true?
Moore was talking about the growth of transistors on silicon chips, and if we look ahead 20 years, we're going to be talking about something else such as carbon nanotubes. We have a lot of technologies in the research labs, which show we'll soon be able to have enormous memory on very small devices. There'll also be enormous processing power and a high-level of parallelism in supercomputers. Fibre-optics is gaining bandwidth at a phenomenal rate and when you have a web that operates at the speed of fibreoptics, you'll be able to have a huge amount of parallel computing on it.
What's next in the evolution of computing?
At Virginia Tech in Bermuda they decided they could build a supercomputer by lashing together a large number of Macintosh G5 processors. Normally, if you're building the architecture for a supercomputer, it takes up to four years and costs between $15m and $100m. They built this one in six months for $5m. They are replacing the desktop G5 processors with the smaller and more powerful server processors. It has now been ranked as the third most powerful supercomputer and means ordinary organisations can effectively have a supercomputer.
Visit: www.drjamesmartin.tv, www.headstrong.com &
