The Semantic Web

[bascule]

Will the Semantic Web ever be successful? For those who don't know, the Semantic Web, pioneered by Tim Berners-Lee himself, is an attempt to move the web from a loosely structured collection of documents to a type of machine-understandable knowledge structures called ontologies. To put it simply, the Semantic Web tries to add meaning to the link structure of the web. It's like what the Xanadu project was trying to do way back in 1960, although they were waaay ahead of their time, and some worry the Semantic Web might be too. So far, no one has managed to make a "killer app" for the Semantic Web's underlying technologies, but this may be in the pipe in the form of Semantic MediaWiki, a Semantic Web-enabled version of the software which powers Wikipedia.org. No longer will the information in Wikipedia be locked up in a way where it can't be used by other applications (except by linking back to the Wikipedia web site), it will be directly accessible to all in a way where it can be cohesively integrated into other sites.

 

The idea is that there's all sorts of functionality spread all over the web which, using the Semantic Web and other "Web 2.0" technologies, you can begin to integrate into more cohesive packages of underlying functionality. In this way the web can begin to take on more of an instantly searchable universal archive of all human knowledge type of feel, when structures that overlay our own understanding of the knowledge they describe can begin to map connections between related fields, and those exploring these connections may soon discover interrelationships between their own fields and similar ones, especially in the sciences. When you look at how interconnected and multidiciplinary interpretation you need to continue scientific advancement (e.g. biologists need to know how old their fossils are, so they turn to geologists to determine the age of the rock layers they were found in, which in turn use techniques created by nuclear physicists) the semantic web will become a great tool for revealing connections in multiple domains of knowledge which we previously couldn't have seen existed because precise descriptions of the interrelationships of ideas in realms of knowledge are typically inaccessible to those outside the field. Searches for similar knowledge structures could automatically reveal such interrelationships in disparate fields, and thus we can begin using computers to automatically reveal similar ideas stumbled upon independently by two different researchers in two different fields, so we don't have to depend on chance encounters to do that job for us (e.g. Geneticists had discovered a 70,000 year old bottleneck in the human genome at the same time geologists had discovered evidence of a 70,000 year old supervolcano eruption in Madagascar)

 

Can you imagine being able to produce a correlated timeline of everything we ever knew of happening all the way back to the dawn of the universe? Think of how many mysteries that could solve, by having instant access to all contemporaneous events...

 

Anyway, it seems like stupid crap like social bookmarking has garnered a lot of interest and in the process usurped a niche that the Semantic Web was aiming to solve. I just wish the Semantic Web would get more attention, and hopefully it will when wikipedia.org itself starts running Semantic MediaWiki...

 

(Ed: There's a Semantic MediaWiki running you can check out if you want: http://wiki.ontoworld.org/index.php/Main_Page)