The “Developing Semantic Web Application on the Java Platform” session just wrapped up at the Moscone Center. Did the panel discuss Java to the audience of Java developers in attendance? No, not really. Did it matter? No, not really. The nearly full room remained nearly full through the hour-long session. And that was a good sign.
The panel discussed real-life applications and four out of five panelists related Semantic Web technologies to social networking. MySpace, Facebook, Del.icio.us, and iTunes were all mentioned (and Amazon as well) in regards to how the Semantic Web can take full advantage of the data on those sites. To paraphrase a panelist, ‘The data is out there and Semantic Web technologies will serve it up to you.’
There was an announcement of sorts, which I thought was rather old news, that Yahoo! will begin officially supporting Semantic Technologies (they have always supported Semantic Technologies) on May 15. I guess the date of May 15 is the new news. The excitement behind this is that Yahoo! should begin creating API’s with the Semantic Web in mind. It was touted that with Yahoo’s support, this should be the year for the Semantic Web to take off.
Using the well-established notion of social networking on the web as a foundation for the Semantic Web was a key point from the panelists. The formula for using Semantic Web technologies was, ‘where is the data, how do we process it, and how do we bring it in?’ GRDDL, RDF, and database technologies appear to be the answers for, ‘how do we bring it in.’
The hard part—which was quickly discussed at the very end of the session—was the necessity of logical links that can be accessed. All of the sample applications the panel discussed were basically the same thing: one-stop shop access to all of your friend’s data, no matter where on the web the data originated. This one-stop shop allows for single sign on access and eliminates the need to visit several different sites to catch up with your friends and family. The Semantic Web technologies use URL’s to gather the data. What is needed are logical links that can be accessed to compile the data. The data is out there—the technologies to compile the data exist—the tough part is putting the two together. Is it possible and will it work?