Friday, May 28, 2010

Clouse Encounters of the Semantic Kind

Explorers are we intrepid and bold
It was bound to happen. Some time ago I got curious about the whole semantic web thing. Working on the geoportal extension at ESRI, we're looking for ways to improve connecting users with producers of geospatial resources. With the advent of systems of systems (although sometimes feeling like turtles all the way down), assuming that a single catalog will do the trick is not an option. So I embarked on a journey into the world of linked data, RDF, and all the fun that comes with that


Out in the world amongst wonders untold
A couple months ago, I got invited to participate and present in a workshop at WMO about information access enablers. Tim Berners Lee suggested to the organizer to look into the RDF model as a way to allow linking data across organizations.

Equipped with a wit, a map, and a snack
So after seeing data.gov experimenting with SPARQL I felt it time to do some experimenting myself. Got some content from data.gov through the REST interface provided by geodata.gov (all 270,000+ geospatial datasets in data.gov actually are registered in geodata.gov and data.gov reuses this content through a web service. how gov 2.0 is that!), downloaded joseki, generated a Turtle file of the catalog, and had my own SPARQL server up and running. All while flying from Amsterdam to DC on my way from WMO to the Gov 2.0 Expo.

We're searching for fun, and we're on the right track
At Gov 2.0 I got a unique chance to sit down with TBL and discuss some of our work. You just don't pass on an opportunity like that! INFORMATION.ZIP. Later that day TBL met with Jack and it suffices to say that SPARQs flew through the room (pun intended). How to model spatial relations in RDF? How to handle relations that aren't explicitly expressed but are determined on-the-fly as a result of some question? What does 'nearby' actually mean?

Like any meeting with your professor at college, you leave said meeting with more work than you entered... I loaded various w3c documents, RFCs, and more prior to board the airplane for California.We're just starting to learn the possibilities of RDF, SPARQL. Providing a text box for someone to fill out an obscure query is not enough. But there already are some good examples available, such as the site This We Know.

to be continued...  


PS: rdf:about="http://www.blogger.com/post-create.g?blogID=1395973636692118549"

1 comment:

  1. Nice post Martin. I like where you are going with this. I also agree that a text box for a query isn't enough.

    ReplyDelete