A short note to myself and the community:
Stop fumbling around with the semantic web, make quality science.
A prototype for a search engine with a bad user interface, an implementation of a rdf database that only works half, an ontology that is never used, we all know these projects.
Our discipline is a crossover, we need results from artificial intelligence, web 2.0, usability, personalization, databases, data integration, software engineering . . .
So – science would be to concentrate on one aspect and then improve that, for example to fix yourself on a scaleable rdf database. You develop a scaleable algorithm and prove in a test setup that it works – voila. But then, it takes YEARS until you yourself or others can use this result in their other projects. Saying “we need named graphs” is far away from having an RDF store that supports them in a scaleable way, but the distance is often underestimated by us.
so, I should concentrate on writing down the good ideas we have and wait -YEARS- until I can benefit from my own ideas using software written by somebody else. Like TimBl using a Firefox.