RDF strategy:

Use the RDF Data Cube vocabulary as a basis (used by data.gov.uk and others, has W3C workgroup).

Schema (URI & URL):
http://seriousgaming.tudelft.nl/vocabulary#Survey
http://seriousgaming.tudelft.nl/vocabulary#Scale
http://seriousgaming.tudelft.nl/vocabulary#Question
http://seriousgaming.tudelft.nl/vocabulary#hasQuestion

URL to specific version:
http://seriousgaming.tudelft.nl/vocabulary/20120401

URI & URL to dataset:
http://seriousgaming.tudelft.nl/researchdata#c9770577-2cda-440a-9375-34b654f20176

Architecture:

Alt 1:

Domain model between client and server
Server uses RDF/SPARQL to create domain objects
Writebacks from Javascript with diffs or function calls?
RDF libraries wel developed

Alt 2:

Use SPARQL directly from Javascript, server is only endpoint.
No consistency checking on model maybe (unless ontology is strong enough?)
Libraries in JS not that well developed

Loosely structured frontend (maybe not visible?).
* Build experiments (surveys, sessions etc)
* Conduct experiments (generate statistics data)
* View statistics (browse and visualize generated statistics)

Respondent < foaf:Agent