New build almost ready, great performance enhancements

Quick update, we are now preparing the next build. We are quite excited since we finally finished implementing several optimization that where missing on the mapping manipulation algorithm. The use of "OR" in WHERE clauses is much more robust now and we now are able to exploit FOREIGN KEY contraints.

The result is that in virtual RDF Graph mode, with real databases and usual SPARQL queries we rarely generate more than 1 query SQL query one which often matches what you would have written yourself in SQL. An it is FAST (also for more complex SPARQL with variables everywhere, even if its more than one query), even with inference enabled.

We'll be posting benchmark numbers soon. Stay tunned!

Documentation and Examples

Quick update. He have re-organized the documentation section of the main site and of the technical wiki . Things should be much easier to find now.

We are also extending the "Examples" section with several of the scenarios that we use for testing, demonstration and benchmarking. We have already included the BSBM and FishMark scenarios that we use for benchmarks, and the IMDB-MovieOntology scenario that we use for demonstrations. Take a look at them, they are nice examples of very natural mappings.

The IMDB-MovieOntology scenario is particularly interesting since the database is real, very large and mappings turn out to be interesting. Take a look!