|
Page 2 of 2
Other Dimensions
It may be early days, but there are technologists who believe the semantic Web has some heft as a technology approach. Alex Iskold, founder and CEO of Livingston, N.J.-based AdaptiveBlue, a smart browsing company that leverages semantic tools and methods, believes that the scattered "bottom up" progress of individual firms in rethinking and reorganizing the Web is already yielding gains in making browsers more interesting.
His start-up offers Glue, a browser plug-in that works to enhance sites such as Netflix, making additional content suggestions based on previous activity -- an extension of what Amazon does by making recommendations based on previous purchases.
An effort by nonprofit organization MusicBrainz provides information about artists and their various projects in a simple search format, with all the complexity hidden "behind the page." Twine, a knowledge management site that says it "understands users' interests," is another example of a consumer-facing site that is based, in part, on semantic Web technologies.
Blogger and technologist Marc Fawzi told CIOZone that RDF "is being used by vendors like Google and Yahoo to structure data for better presentation in the browser. But unlike OWL-DL, RDF does not have the structure to implement a logical knowledge model for the given domain, which is required by machines to reason about the information published under that domain." However, RDF and RDFa (and other variations) are, in Fawzi's words, "perfect for structuring the information itself, as opposed to the knowledge model."
The next step, says Fawzi, will be to use RDF to structure information for machine comprehension, not just browser presentation. The latter, he wrote, assumes the use of an inference engine to supply the domain-specific AI, which is fundamental to any attempt to make information "comprehensible" (not just "process-able") to machines, per the semantic Web vision.
The difficulty with the semantic Web is not in natural language processing, according to Fawzi. "The 'hard' problems are usability and practicality issues. When it comes to the vision of turning the entire Web into 'a semantic Web,' it's tough." Especially, he says, given the fact "that not everyone who puts information on the Web is an expert in description logic models -- used in building the ontologies -- and the structured information standards."
Fawzi likes the mostly trickle-up development path that the semantic Web has taken. "Instead of hoping to turn the whole Web into a massive pan-domain knowledge base," domain-specific vocabularies have developed.
Only registered users can write comments.
Please login or register.
|
