Posing this question to Google results in over a million articles. But are any of them right and in what sense does Google actually understand the question?
The Semantic Web was a phrase coined in the late 90s by the architect of the World Wide Web Tim Berners-Lee. It describes his vision of the future of the Internet as a universal depository for data, information and knowledge and it looks increasingly likely that the release its first fully functional applications may drastically alter the way we think about search engines.
The Semantic Web might be considered as the next–generation World Wide Web where human-facing, natural language content pages are enhanced by the inclusion of a hidden layer of ‘meaning’ that is understandable to software agents. In essence, the idea is to add context to the Web’s content and make it easier to locate, integrate and disseminate relevant information.
The project has been on the drawing board for a number of years now but is now re-entering the zeitgeist and has topped a number of prominent lists of ‘what will be big in 2008?’ There are currently two approaches to Semantic Applications:
The bottom up approach requires semantic annotations to be appended to existing data in a similar way to current meta-tags. This type of approach is already fairly well utilised in well-known Web 2.0 destinations such as Flickr and personalised retail properties like Amazon.
The top down approach attempts to analyze existing data and attempt to extract meaning from it. The overarching vision for the proponents of this approach is a natural language processor capable of reading and understanding written text. A number of start-ups claim to have made real strides in this area in recent months and a number of potential semantic search engines are now in the process of being tested.
One of the most interesting and potentially important pending additions to the Semantic Web space will be Polar Rose. This search engine fuses a mixture of top down and bottom up approaches with its own proprietary software and is aiming to index images and make all of the Web’s public photographs searchable.
Powerset and Twine are two of the most promising ‘top down’ search engines from the current crop. Both of these projects are in Beta testing and ultimately intend to become natural language search engines capable of understanding a search query by analyzing full sentences and returning a meaningful and ‘correct’ answer.
Google has so far shot down the potential importance of this type of natural language search by claiming that people themselves, when searching for things in any situation, tend to operate on what could be considered a ‘keyword’ level.
The search giant does, however, appear to recognize the prospective significance of image search and has even quietly launched face-search functionality. Conducting a search for Tom Cruise at the following appended Google URL – http://www.google.co.uk/ig?hl=en&imgtype=face – will result in only those sites which feature face shots of the pint-sized, Oscar-winning scientologist.
Further Reading
Wikipedia
