Thinking Clearly .

Ah, my first ever blog entry. After years of flying under the radar, I am now bursting upon the blog “scene.” I have been an employee with C&P for 9 months now; previously to that, I worked with them as a contractor for about six months. And even before that, we were colleagues at the University of Maryland’s MINDSWAP group. So I’ve had the distinct pleasure of working with this merry little band of hackers for several years now, and I’m very happy the relationship continues.

I have been hacking lately on jSpace, our 100% pure Java clone of the mSpace system, which includes some improvements of our own. This past week I’ve been trying to track down interesting data sources to use as demos, and to help test and extend the little automated model creator we have that helps reduce the work in writing models for new data sources. Musicbrainz was, and is, my first choice for a data set. It is an nice chunk of data that should lend itself well to the jSpace way of looking at the world, and its data that anyone can find interesting. Unfortunately, I could not find a dump of the data anywhere. With some magic from New Mike, we were able to get a scrape of the latest data using the old, and unfortunately VERY outdated RDFDump.pl script. Looks like we’ll have to do some hacking on that to get some good RDF.

The CIA World Fact Book provided a nice dataset that worked with only a minimal amount of massaging, which as mostly adding some rdf:type triples to each of the Countries to identify them as such. That demo is online at the jSpace site, feel free to cruise on over and check it out.

Lastly, I am trying to make something happen with the wikipedia scrape from dbpedia.org. It’s a fairly sizeable dataset, so I wanted to use their SPARQL endpoint rather than hosting it here, but the fact that jSpace does not play nicely with literals has proven to be a big hurdle. The next version will have much better support for literals, but that’s not on the todo list in the near future. I’ve got some idea’s about having an adapter that will re-write the results as they stream back from their SPARQL endpoint, but that will take a bit of magic and some good old-fashioned elbow grease to get working.

I’m always interested in new idea’s for data sources to demo, so if you’d like to suggestion one, or just like what we’re doing and have some ideas or comments, cruise on over to the jSpace homepage and drop us a line.

Next project to tackle, creating the new C&P website on our new Solaris container. Will be a good chance for me to learn some Ruby and to brush up on my HTML and javascript. Until next time…Stay Classy Cyberspace!

This entry was posted on Monday, February 12th, 2007 at 2:55 pm and is filed under RDF, jSpace. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.