Technology Review – Tabulator Firefox Extension

Did you ever notice how many Semantic Web applications are so focused on providing graph views? Sure, graphing makes sense from the third sentence in any RDF tutorial you read. The basic building blocks of information: objects, predicates and subjects come together to form triples DNA. From this, all knowledge springs to life, interconnecting and weaving until everyone and everything are separated by nothing more than a chockfull of vertices and nodes. Graphs make it so easy to grasp the connections, so simple to walk the data, so intuitive to see how it all fits together.

Unfortunately, the graph view is not suited well for large sets of data. In fact, the more connections between data, the more difficult it is to make sense of the relationships and the graph as a whole. That’s the rub – reaching the point of putting it all together, the penultimate dream of the Semantic Web, only to have the display implode under its own weight. Sometimes the problem can be circumvented with displays such as hyperbolic or cone trees. but they all have their limits. I tend to be a fan of the keep-it-simple philosophy which lends itself well to the new Tabulator Firefox extension.

I remember playing with the first Tabulator version Tim Berners-Lee produced at the end of 2005. It was a simple hierarchical view that retrieved data on-the-fly, relying on a simple set of colored icons. The premise was a generic tool that could display any kind of RDF data. The extension enables users to go right to the data source, supporting N3 and RDF/XML data based on select MIME types. There’s no need to save the triples to a file and then use an RDF viewer to load the data. Laziness is a virtue.

I started with DOAP and Firefox loaded quickly. I moved on to FOAF and it didn’t load quickly. Returning to DOAP, the extension made it easy to walk ontologies via owl:imports, expand and collapse RDF hierarchies at will and explore cyclical relationships. You can choose to view the serialized data (limited to RDF/XML and N3, of course) as well as embedded DBpedia entries! I can’t see all the data most of the time but I don’t need to. Yet, I can’t search the data easily, and I missed that aspect almost immediately.

The best part? SPARQL Update. Yes, it seems they’ve incorporated a version of Andy Seaborne’s and Geetha Manjunath’s SPARQL Update specification via HTTP POST. Most users won’t even bother going down this avenue (especially since some configuration is required) but you can imagine the possibilities of haveing an AJAX-enabled application that both reads and writes RDF without your ever having to see the nasty details of RDF/XML. Add some inline auto complete functionality, and I was in heaven.

I wish that there was a better visual indicator of progress instead of just yellow dots. Yes, I was extolling the virtues of simple colored icons earlier but I just don’t have the patience. At least if I see a progress bar, I can get an idea of how long to wait before I poke my eyes out.

The Tabulator extension is a great little web tool for traversing the ‘web of data’ that is all around us – without employing the ubiquitous graph. Don’t get me wrong, I think graphs have a purpose. But when I don’t know the size of an endpoint’s data or its distribution, this extension provides solid results. I’m realizing the goal of the Semantic Web: endlessly connected information and it feels good.

Check it out at http://dig.csail.mit.edu/2007/tab.