Radar Networks is hoping to have the first Semantic Web killer app in its just-announced product, Twine. In case you haven't been following Sir Tim Berners-Lee's pet technology, the Semantic Web will be better than the boring, old, plain-vanilla Web in that everything will be tagged using Resource Description Framework (RDF) and ontologized with Web Ontology Language (OWL) so that pages and content will be meaningful. The newly added meaning will make it possible for other computers to use the content intelligently: by giving you more-useful and context-aware search results, for example. Twine extends this to all of your information--email, sites visited, feeds--and will automatically generate the tags to "semantify" the data.
Radar's Nova Spivack treated PC Magazine staff to a phone-and-Web demo of Twines this past Wednesday, though we weren't able to start actually playing with the product yet. Spivak stressed the importance of the product's motto,"Tie it all together," and likened the Twine interface to "a Facebook, except for information." Each user of the hosted Twine service will have his own "twine," which analyzes and stores all the collected metadata for all of his information. Twine will offer several ways to enter content into its uber index of all your information--via a Web button, by importing from documents, receiving it from collaborators, or sending email to your twine. Connecting your twine to other users' will result in a network effect benefit to knowledge management, claimed Spivack.
Twine is a hosted service that uses open semantic standards RDF, OWL, and SPARQL (Simple Protocol and RDF Query Language), but will probably become available for large organizations. It will be free, supported by personalized advertisements, but premium versions will become available that have storage and admin options.
I'm skeptical of the idea that machines can find the meaning in Web pages without the page's creator specifying the information, as is Twine's proposed goal. Say a Web page for a product contains this data: "This is a consumer electronics product. Here's the product name. It's a television. It's a widescreen television. It costs $999. This is the description." Without someone supplying the metadata for each of these pieces of information, how can an algorithm know where on the page the relevant data is? Maybe that's just where Radar's secret sauce comes in, and you have to give the company credit for tackling this challenge.
Though even its creator seems to have given up the idea that the Semantic Web will replace our tried-and-true Web, there are rumblings in the developer community that the idea will still be used for something alongside the existing Web--most likely intra-organizational collaboration. Twine seems to fit that scenario, and seems suited to use by trusted team members who need to share knowledge for their projects.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment