Source: Wikipedia
Understanding context is one of the the big chasms that we have been trying to cross in internet applications over the last five or so years. Its right up there with understanding natural language. Search would be smarter if it knew the context in which you were thinking; online advertising would be more effective with context' music recommendations would be smarter with context and so on and so forth.
Many, many start-up have tried to to bring context to web applications. None have done an outstanding job so far. An early technology out of Northwestern University called Watson, attempted to watch what you were writing in a Word document and offer web links to research that might be useful. They are now called Media River and have changed their model. Personify attempts to classify web pages. Surf Canyon has developed a browser plug-in that refines google searches to make them more context sensitive. It watches what links in the search list you click on and re-orders the results accordingly. It attempts to solve the "Jaguar" search problem - am I looking for the animal, car or operating system?
Zemanta is one of the latest context tools that attempts to recommend contextually related content and pictures to bloggers. They have a implemented their solution very elegantly using Firefox extensions. All you have to do is install the plugin, go your supported blog platform and the tool magically appears. As you can see from the suggestions below, the technology has a way to go. It seems to be only using key words, not context. The nice implementation helps overcome the questionable recommendation.
Comments