Raw

Most strategies for Web user interfaces follow the paradigms of existing data browsing: formatted text, forms, tables, spreadsheets, trees, spring-loaded graphs, desktops... But these paradigms have evolved through application to data that is ultimately either document-shaped or well, database-shaped, the presentation making it useful as information. The current Web is mostly human-readable documents (or media objects etc) so the interface tends to be oriented that way.

But RDF and OWL are built to describe things, not just documents. It can be used for data in general, not just metadata. A lot of SemWeb data represents real-world artifacts and relationships between them (or at least conceptualisations of these). E.g. the FOAF vocab lets you describe people and connections between them; Geo stuff represents a 2/3D space; iCal/RDF offers event timelines. These representations map the physical world more directly, not intermediated through the world of human-readable documents.



There is plenty of software around which tries hard to model physical world-like systems, with their user interface optimised for these. Games. I was just reminded of this when I stumbled on nebula device, an open source scriptable games engine. Could be useful for condensing vapourware.

Games do their utmost to create rich environments despite the actual "active" data and algorithms being thin and extremely simplistic (compared to real-world physics and social interaction). SemWeb data can be very sparse (maybe this is inherent), and again the logic has nothing like the complexities of real-world analog interactivity. Seems like some kind of "shading" techniques could be applicable.

I can't imagine a 1:1 mapping of the physics being much good - if you've got a 250 pound friend in New Zealand or a want to view a 5' artwork in Ecuador, they could take some locating if you're flying through thousands of miles of VR space. However it should be possible to stay true to the topology of the SemWeb graph whilst allowing interesting/useful interactions. This stuff is well out of my league, but presumably the VR folks have spent a lot of time thinking about such issues.

There's a crossover here with the tradition of 3D visualisations of chemical molecules and the ongoing work with biochemistry using SemWeb technologies. But in the VR view of the SemWeb you might still have atomic balls joined by sticks, but zooming out, you'd maybe have people joined by flightpaths.