Semantic Technologies & Knowledge Management

According to Tim Berners-Lee, the inventor of the dumb Web as we know it today, "The Semantic Web is an extension of the current web in which information is given well-defined meaning, better enabling computers and people to work in cooperation." link

There is much work being done, led by W3C, to define necessary standards and methods to realize the dream of Semantic Web. Unlike the current dumb Web, Semantic Web promises to deliver meaning ("semantics") along with the content as we download it.

I find it exciting, but am skeptical of all its glory promises. It sounds like AI of the past decades, promising, but no one knows where it is heading. A demonstration of small successes in limited contexts here and there lifts our hopes to the skies, only to get beaten down by the reality. There are hopes that the "network effects" will enable the Semantic Web proliferate, just like the dumb Web did in the past decade. Unless there is an agreement on the semantics, either predetermined or evolving, across the network, I don't see the network effects affecting anything. Luckily for the dumb Web, everyone readily agreed on its dumbness and embraced it quickly.

The dumb web accomplished trivial things, but offered greatest value. The semantic web is trying to acccomplish the hardest things, and the value it offers is not as great to ordinary beings. There the problem really lies - the 80/20 rule - 80% of the benefit came from the first 20% of efforts on the dumb Web, and now we will be trying to squeeze that additional 20% benefit with the remaining 80% effort. So, I expect the players to embrace it much more slowly than they did the dumb Web, which in turn will diminish the network effects. And, many just won't care for that additional 20% benefit.

That said, while I don't see wide ranging success of Semantic Web in the broadest terms in the near term, I do think many interesting applications will evolve in specific contexts. For example, comparative shopping could be a cake walk, if all vendors marked up their e-commerce sites with standardized ontologies and metadata (presumably using OWL and/or RDF). Similarly, news articles can be marked up with meaningful metadata to enable easy classification and categorization. The standardization of metadata and ontologies in specific domains and contexts could lead to very useful and interesting applications.

I also think the intra- and extranets of corporations can get very smart by offering targeted search, navigational and personanization capabilities. It is easier for a corporation to standardize on ontologies, controlled vocabularies, thesauri, and metadata to enable consistent markup of documents and to develop and deploy capabilities that use them.

I think Semantic Web technologies will complement other KM technologies (text mining, classification, taxonomies, topic maps, etc) very well in the corporate world.