The Semantic Web

Few people, in business or otherwise, would argue that the World Wide Web really has changed the way we think about and actually do business.  Even considering the Dot Com bubble burst and the odd few thousand failed business plans, it's clear that the World Wide Web really will change the way business interacts with other business, its customer, suppliers and the world at large.    Evidence further suggests that many everyday people going about their normal daily lives use the Internet extensively.

Unfortunately there is something seriously missing from the equation - information.  Data, in its raw form, is not information.  On the Web we have mountains of data. We've got more data than we could ever have imagined.  We’ve got everything from sales records to medical research to engineering specifications available online – by and large most of it unstructured or poorly catalogued. 

Finding the data and turning it into meaningful information is becoming worse than finding the proverbial needle in the haystack of the Web.  Search engines are very limited when it comes to sifting through a mountain of data that is increasing steadily.  The Semantic Web however could make finding and organising “real” information much easier 

How are we going transfer all of this data into meaningful information?

The idea is to build a new Web.  This one is based on the very meanings and syntax of our language but essentially remains transparent and invisible to the humans who speak it. The researchers and academics that are planning and developing this brave new Internet, have called this version the “Semantic Web”.  Strange as it may seem, in making the Web more useful to people, this Web is aimed more at machines. The ultimate goal is to allow computers to not only just process and move our words and data, but far more importantly, to understand them and give them meaning.  

For the last seven or eight years, the Web has been very focused on giving value to human visual interpretation – i.e. what we see and read.   Over the next half dozen years, interestingly, the eyeballs will belong to our computers.

So, is this a brand new version of the Web?  What will happen to the old one?

To be fair, the Semantic Web should really be characterised as being more an extension of today's existing Web rather than an entirely new one.   Tim Berners-Lee, the father of the original World Wide Web is the Semantic Webs greatest champion. 

Berners-Lee, believes ultimately this version of the Web will probably take the form of specialised tags inserted inside HTML documents by their creators.   Our current version of the Web is founded on these HTML documents, they are the mechanism used to build and identify the “pages” throughout the existing Web.  These specialised Tags will not simply identify the page but will also help computers understand its meaning, that is, what it's actually about.  It's a way of annotating and giving meaning to any form of content, any type of data, with additional semantic (meaning) relationships that are able to be machine processed.  Consequently a computer can infer far more about what's contained in that content.

How will the “Semantic Web” be able to identify and give meaning to this data on the Web?

The Semantic Web can do this because at its very heart are dictionaries that draw direct relationships between terms or descriptions. As an example the Semantic Web knows that a magazine is also sometimes called a publication, that people are also employees who work for a company, and so on.

Any system running a semantic search would see these Tags in a Web document and lookup a dictionary to define them and work out relationships before proceeding.  In a simple way a similar thing happens right now when you say, type abc.net.au in your browser, the browser program actually contacts a sort of dictionary, usually called a name server, to find out where to find the computer that Web site actually lives.

As an example, how would business use the Semantic Web?

Consider the hypothetical case of a manufacturer that needs to find the perfect part for a new product it's developing. The designers could instruct a semantic search tool to find, say, nuts that are lightweight, very rust resistant, of a certain size, cost less than a few cents, and can be delivered at the same time each week.   By accessing the relatable semantic tags in product catalogues from a variety of suppliers, a program will analyse, compare, and evaluate all the options, presenting the manufacturer with a list of nuts that best meet its criteria. That wouldn't be possible without semantic tags - not everyone says a 'cog' is a 'cog’.

The Semantic Web could also facilitate more automated monitoring of corporate financial performance, such as tipping off investors and regulators to incidences of insider trading or poor financial reporting.

Arthur Hissey
Computer Research & Technology
www.crt.net.au