Article - Issue 46, March 2011
Dr Scott Steedman CBE FREng
Dr Scott Steedman CBE FREng
The potential of the internet to transform human behaviour has been argued for some time, but in the last few months we have seen for real just how fast the world can be changed by information. Arab states have been experiencing the power of social networking sites to transform traditional power structures. Back in the UK, a new Home Office website crashed as five million visitors an hour tried to access crime data by neighbourhood address.
As Professor Nigel Shadbolt FREng commented in his lecture on the future of the World Wide Web at the Foundation for Science and Technology in December, the Web is the most successful information architecture in history. Growing at an astonishing rate, the Web holds an unimaginable number of documents at any time. Last year, the digital content on the Web grew to over one zettabyte (one zettabyte is a million million gigabytes). However, although every document has a unique address, it is more difficult to determine what a document actually contains.
Search engines such as Google plough around the Web constantly hunting for answers to human and machine-generated queries. Search engines are increasingly sophisticated, but the data in documents has no context, so automated searches have to assess the relevance and likely importance of the information in Web pages based on string matching and the number of pages linking to it. We humans then have to interpret what the search returns to us.
Our intelligence is also used to help computers to solve problems that they can’t solve themselves. Humans recognise millions of words lifted from poor quality scanned documents every day for Google for free using ‘captchas’, an onscreen test originally developed to help tell humans and computers apart for security purposes. The captcha works by getting people to re-type distorted text that they see on the screen. One word will be a security check, but other words are commonly strings of text that the computer’s optical character recognition software could not recognise. Such are the problems and opportunities of a network of linked documents.
The emerging discipline of Web Science predicts that the Web could soon become a network of linked data that complements the familiar network of linked documents. A Semantic Web, where every piece of data was uniquely identified when it was first uploaded, could help transform the current search process, enabling machines to process the information far more efficiently.
This vision of a Semantic Web – one where machines could analyse the content to help assist human interpretation – was originally set out by Sir Tim Berners Lee FREng in 1999, but it has been slow to materialise. Last year, though, the UK government published large quantities of linked open data from sources as diverse as the Ordnance Survey and the London Borough of Camden. A new Web is emerging.
For engineers, a web of linked data would transform how we work. Data might be public, like that released under the UK open government initiative, or held on a secure intranet, perhaps related to a project, supply chain or asset. The concept of the Web as a library of documents is giving way to the Web as a network of resources.
For engineers, the Semantic Web will be an information-rich environment, capable of automated exploitation using machinebased systems. The real advantage to engineers will be the ability, for the first time in history, to find and access on demand accurate information across the supply chain – information on materials, products, markets, regulations, everything.
Over more than 150 years, engineers have relied on layer upon layer of codes and standards which set out to protect us from making misguided decisions in the absence of information. The opportunity to link information directly from a wide range of sources could transform our ability accurately to predict the cost and performance of engineered products throughout their life.
The Semantic Web is a big idea for engineering, an idea that could herald a step change in opportunities for efficiency and optimisation in our systems and products, but it will demand a radically new approach to our understanding of the nature of engineering resource. An Academy report could investigate the opportunities across our industrial base. In the UK, in the quality of our academic research, for example, we have a great deal of the underpinning knowledge of the challenges and opportunities that a Web of linked data could bring to the engineering profession. Surely we should exploit it? Dr
Scott Steedman CBE FREng