Scoble interviewed Radar Networks' Nova Spivack for an hour about their new "Semantic Web" application, called Twine (twine.com). Nova demoed it, and for those of you who don't have an hour, here are my notes:
I watched the whole thing through. Nova said there are 20,000 people on waiting list for beta, will be releasing it slowly in Jan to foster scalability...
The app has REST and SPARQL api's, and utilizes OWL, RDF, and later will utilize GRDDL (Gleaning Resource Descriptions from Dialects of Languages; that is, for getting RDF data out of XML and XHTML documents using explicitly associated transformation algorithms, typically represented in XSLT. As of September 11, 2007[1], it is a W3C Recommendation).
Nova's more notable quotes:
"its a listserve, wiki, and group, with a semantic database under the hood"
-it's targeted to data intensive users, not general public. i immediately thought of basecamp from its interface and funcionality.
"our semantic graph is a superset of the social graph"
-apparently they have a patent pending on this concept, just like Friendster patented the social network..
"will spread faster than a social network, because designed to live on top of your email behavior"
-can invite people to app thru email, can cc content emails into Twine, it imports the content and enriches it with rdf.
"we now live on top of relational databases. scaling issues are what took us so long to figure out.
-probably b/c they're trying to rebuild the web.
"we're letting you build your own semantic web"
-instead of trying to index the entire web, nova is trying to get "the best of the web" into his system thru users importing everything valuable in their lives (bookmarks, rss, email, desktop, twitter, flickr, wikipedia, amazon). then Twine marks it up in rdf, pulls out entities and tags etc, and he mentioned that you might possibly be able to push it back out at some point. his leverage is that Twine is the only system (currently) that can make sense of all the connections.
"we will provide a higher signal to noise ratio than google"
-mk. but but it also requires a higher effort to results ratio on my part.
he also talked trash about facebook being closed and the lack of "semantic" relationships within, despite the richness of the social graph.
He mentioned that there are tons of opportunities for "visualizations" and data connection from other sites (i can't help but think Google here), and started talking about recommendations being a huge feature of the service later on.
I might see myself using it professionally. i already bookmark content/articles into iGTD or Basecamp that I don't want to forget, and if i could just type "tips for Facebook app developers" into Twine and it would find everything relevant, that would be useful. and then if it would automatically tell me everything relevant that other people like me read, that would be really cool. they currently capture all kinds of user experience data, including viewing time, click streams, etc. for the purpose of making better recommendations (and for advertising later).
and their 2 tier subscriber model actually works in their favor. the free account requires everything you import to be public. most users will be free users, so most info imported will be public, so i would have access to possibly millions of other people's enriched data.
but therein lies a bitter truth. the real PROBLEM is info overload. if all of a sudden EVERYTHING that I SHOULD read on a topic was recommended to me, I would probably need Prozac. i already have enough stress trying to go through the stuff that I find myself. the last thing i need is MORE stress from MORE unread stuff.
Tyler - great find. Thanks for posting your notes and for actually listening to it all. I've been waiting for the Twine invitation. I just hope it's ready for OS X apps when we finally see it in action.
If I could find my iPod I'd load up this interview for my trip.
Posted by: Blake Burris | December 19, 2007 at 10:33 AM