byteblog

The search giant Yahoo, amongst others, is experimenting with a new version of the internet called the “Semantic Web”. In other words, the aim is to try and get the internet to behave more like person, rather than just a collection of dumb terminals.

At the moment, when you do a web search for, say, “office equipment”, the search engines use the specific words in that search request, and try and match pages that they have previously indexed. This can sometimes throw up all sorts of anomalies and weird pages. Basically this is because people are still much better than computers at associating between words and phrases, and as people’s associations differ wildly between age groups, countries, races and so on, it is very difficult for a search engine to get it right every time, for everyone.

The solution, say the techies, is to have the web think for itself, and instead of just dumbly returning a list of pages, it should actively search for relevant content by understanding its meaning. Of course “understanding” is a relative term - we think it is doubtful whether a computer will ever pass the Turing test in our lifetime, as the meaning we as humans attach to information is vastly complex and, ultimately, any computer will have to be programmed by us, and we do not fully understand this relationship ourselves.

We all saw what happened in Terminator 2 when SkyNet started to think for itself (it fought back, apparently, when someone tried to unplug it), so maybe going as far as a fully self-aware internet wouldn’t be such a good idea. However, anything that makes our daily trawl through the internet a little easier and quicker is a good idea in our book.

Check out:
http://news.bbc.co.uk/1/hi/technology/7296056.stm
http://en.wikipedia.org/wiki/Turing_test