Nov. 6, 2006 | One hundred million nodes – that’s the number of servers now interconnected on the Internet. Citing the rapid growth of the World Wide Web over the last decade, a group of scientists from the Massachusetts Institute of Technology (MIT) and the U.K.’s University of Southampton believe the Web is now so large as to require its own field of study if it is to evolve systematically and serve humanity optimally.

At a briefing held at MIT in Cambridge, Mass., on Thursday, the scientists announced the formation of the Web Science Research Initiative (WSRI), the purpose of which is to produce the scientific advances necessary to guide the future design and use of the Web. Using a multidisciplinary approach, the WSRI will generate a research agenda for understanding the scientific, technical, and social challenges underlying the growth of the Web. WSRI researchers will address how Web-based information is accessed, and its reliability. Legal questions and social networks – how people use the Web to communicate – will be critical areas of study.

Such fundamental research is considered crucial, said Tim Berners-Lee, professor at MIT and the University of Southampton, and one of WSRI’s cofounders: “The Web is something we created, so we have a duty to make it better if we can.” Referring to the organic growth of the infrastructure thus far and the basic rules established early in its development, Berners-Lee said, “We’ve created the microscopic infrastructure of the Web, but little things by themselves are not sufficient to understand the big thing. We’ll be developing new ways of analyzing things and will be building systems that have completely new properties.”

Berners-Lee stressed the interdisciplinary nature of the effort needed, saying that biologists, neuroscientists, psychologists, legal experts, and economists must all be involved. “Huge number of fields must be linked together. We must think about and engineer the Web as one huge system.” The goal is social, he said, to make a system that serves humanity. “It’s about making the Web infrastructure a richer, more powerful space in which to do exciting things.”

“Accountable software” is an example of the kind of benefit the multidisciplinary approach could yield. In drug discovery, clinical trial data involves policy issues. Researchers need to know which patient data they can use, and what they can use it for. If, for example, epidemiological data were collected, could it also be used for anti-terrorism activities? By drawing upon data not only from the clinical trials, but also from federal regulations governing such data, “accountable software” would explain where the data is from and which rules it used to combine it, to help users figure out if they’re legally allowed to use it for the purposes they have in mind.

Blending the Boundaries
Underscoring the multidisciplinary approach, Nigel Shadbolt, professor of artificial intelligence in the School of Electronics and Computer Science at the University of Southampton, said he has always been intrigued by the boundaries between disciplines, “which is where a lot of the interesting questions live.” Because the Web has gotten so large, Shadbolt said, “We need new forms of analysis and new methodologies to understand some of its dynamic properties, how it evolves, and what makes certain aspects of it stable and resilient.”

He pointed out similarities between Web development and the life sciences: “One of the things that happens as we try to understand extremely complex systems, like living organisms, is the need to invoke what is called a ‘systems biology’ perspective. You can’t reduce everything to its very simplest level and understand the high-level principles, but you actually have to have tiers of explanation. And if one looks at the work in biology, at the level of understanding the genome and how proteins are encoded, the chemical pathways and communication pathways between cells, you see this process of the levels of abstraction, and the need to apply mathematical and statistical techniques to understand some of those irregularities. Intriguingly, recently, even importing back into life science some ideas from pure computer science.

“A particular branch of computer science – process algebra – is currently being used to understand how protein networks interact,” Shadbolt continued. “Here you’ve got a very direct analog where the interaction between disciplines can be very fruitful. When we look at how people are trying to understand protein networks, trying to understand communication in nervous systems and other biological systems, there are tools and methods of insight that we believe could be very valuable to our understanding of the structure of the Web itself and the Internet underlying it.”

Getting from Here to There
Wendy Hall, professor of computer science at the University of Southampton, addressed some of the practical aspects of the collaboration, including money and training. The WSRI intends to raise lots of money for joint research, she said, and passionately spoke of the need to train new generations of experts with the right combination of skills.

Citing a common complaint from employers that there are far too few people with the right skill set, she said, “We can’t do it with the existing cohorts of researchers. We need to build new cohorts of Ph.Ds who understand importance of the multidisciplinary way of working and can appreciate the best of other disciplines.” Down the road, WSRI plans to develop teaching modules for the undergraduate and master’s levels, as well. Teaching will be done at both MIT and the University of Southampton. Ultimately, “we would like people around the world to be calling themselves Web scientists,” Hall said.