The school of electronics and computer science (ECS) at the University of Southampton is on a par with the prestigious information technology engineering schools of MIT, Stanford and Tokyo, and is where optical fibre technology was first developed, a field in which it is still a world leader.
Its main lab burned down in October 2005, and the £100m building that replaced it opened in 2008, jam-packed with state-of-the-art kit, computer labs, 10Gbps networks, unique equipment for esoteric experiments, and some remarkably tolerant scientists who showed me around.
A day visit cannot do justice to the school. We only had time to skate through some of the top-line issues, such as biological and molecular computing, intelligent autonomous robots, advanced optical fibres for telecommunications, high-energy laser guns, internet addressing systems, the semantic web, crowd-sourcing and open data.
The future of information processing
One of the biggest problems facing the IT industry is the coming energy crisis. Silicon runs hot, and is simply not good at some important information tasks, says senior lecturer Klaus-Peter Zauner, whose interests range from how to program a molecule to how to make a lot of smart autonomous robots "play nicely".
Not only is the search on for "cool computing", but also better ways to process information. Nature is good at solving problems with what Zauner calls noisy ambiguous information where there are no clear rules that you can program into a system to tell it how to behave.
Zauner says a seed has a complex information processing problem in deciding whether to grow or stay dormant, depending on light, temperature, soil moisture content and other variables, where getting it wrong is lethal.
A bacterium or seed is much more efficient at processing and integrating this type of information than computers today. This is because computer scientists never abstracted the principles of biological information processing to improve computing capacity, the way aircraft engineers abstracted the principles of aerodynamics by studying birds' wings, for example.
As a result we have to build in high-energy barriers to keep data discrete and enforce its programmability so that we can process it, which is why computers today run hot and are so big compared with the central processing unit of a fly, for example, says Zauner.
So what can we do now with molecular computing? "Extremely little," he says. "We are at the stage where we look at the birds but have no understanding of the aerodynamics. We can see information processing happening at the molecular scale, so the proof of principle exists, but we do not understand it. All we can do is marvel."
Applications for autonomous agents
Zauner's work feeds into research into software agents that control devices such as flying drones (unmanned aerial vehicles, or UAVs), which armies use to see what is happening on a battlefield. But there are other applications, such as disaster management, where incoming information is incomplete, out of sequence, ambiguous, self-interested, and where cooperation is vital.
Alex Rogers is working on Project Orchid, which continues the development of autonomous agents from recently finished Project Aladdin. Orchid could change the way online auctions are held, or control consumers' energy consumption via a smart meter. "Your dishwasher may soon look on the internet, check the energy generators' carbon intensity and hence pricing, and decide for itself when it will do the dishes to minimise the cost to you," said Rogers. He is already working with some firms to commercialise these ideas.
Crowd sourcing is another area he is interested in. But it is more subtle. "Who wants to be a millionaire? (the TV game) is really interesting," he said. "It is not that the crowd knows the answer; it is more like there are a few individuals who really do know the answer, and they take the rest of us with them."
This insight could have profound effects on public policy and firms' product development and marketing strategies. "The issue is how to develop a system so that, within a social network, the person who really does know can get his message out, so that the crowd can produce the right answer," he said.
Making data widely available
Crowd sourcing might be done through the semantic web, which one day, may link all the information on the web in ways that make sense, depending on the question you ask. How to catalogue that data is the interest of Thanassis Tiropanis.
His view is that the web has always been built from the bottom up, and this will continue. So people will develop systems for cataloguing information that make sense in the context that concerns them, but they would do well to leave open ways for other people to link to it, Tiropanis says.
"We may never get there, but what matters is the process," he said. Interest in the semantic web is rising because people can see the value of making data available."
This is perhaps unsurprising. Nigel Shadbolt, deputy head of ECS, led development of the UK government's Data.gov.uk initiative with visiting professor Tim Berners-Lee. This has seen more than 100 apps emerge based on more than 6,100 data sets previously closely held by the government, such as Ordnance Survey maps, which were made available to the public under licence, but free.
The rest of the tour included sights of holographic 3D images on large-screen TVs, insights into IPv6 addressing issues, and ECS's work to boost optical fibre bandwidth 1,000-fold. There will certainly be plenty to write about information processing in the future if ECS's work is any indication.
This was first published in April 2011