|
Page 1 of 2
By Lauren Bielski
CIOs have been hearing about the semantic Web for years now, at conferences or in the context of business intelligence. But if you and your company aren't using it, the method might have slipped under your radar. At the same time, periodic articles hailing -- or quietly jeering -- the semantic Web movement might leave you wondering what the fuss is about.
While it can seem esoteric, with its requirements of syntax and call for ontologies, the semantic Web has a basic core: It is in one sense about systematically adding standardized tags that can help computers read information like statistics, weather data and TV programming contained in Web pages and act on them intelligently.
Advocates, among them Tim Berners-Lee and his W3C -- which envisions an epic effort of re-crafting annotation of the existing Web to support semantic queries -- say that a semantic Web methodology can offer businesses more valuable information than that obtained by keyword searches and statistical inference engines.
In the consumer realm, the semantic Web promises improved searches, because, in effect, machines are reading tags that allow them to act as if they understand the meanings behind the search terms (no more searching for "jaguar" the animal only to have half the results refer to the elite auto). It may not be AI, but those tags, and the Web services and other tools that can work with them, support some basic inference logic and are able to mimic thinking.
While emerging standards have made it possible to set Web pages up to work this way, it isn't easy. There is a huge quantity of pages involved, as well as a human tendency toward endless variation.
Using the semantic Web is about leveraging metadata to make Web pages machine readable. Page builders do this by using a uniform resource identifier (URI) -- usually a "triplet" -- modeled on the subject/predicate/object construction found in many languages. A URI representing the sentence, "This stamp is from England" would include the noun "stamp" the verb "from" and the proper noun "England." Other rules of usage would support a more complex chain of reasoning for the machines handling the pages.
Why all the effort to enhance the machine-to-machine interface? If used correctly, says expert John Hebeler, the B2B semantic Web can support better integration -- lifting differently formatted data from multiple databases with the aim of running queries on more massive data sets. Robots and other tools could also leverage a Web laced with meta-tags.
|
