Hey, I just committed a first version of a SPARQL EntityFinder, which calls the SPARQL endpoint mentioned in the previous post. Basically, the Query is transformed to SPARQL just like the default RDFEntityFinder, but is instead posted to the SPARQL endpoint, and then the result is parsed using SAX (for performance).
So now clients can use the SPARQL EntityFinder to locate serverside Entities, and then use the REST EntityStore to access the state, and yet codingwise this is all hidden behind the UnitOfWork/Query interfaces. Me likes! One other thing that I forgot to mention about the SPARQL endpoint is that if you notice that some queries are done very often, and want to optimize these, then it should be possible to add a filter which sets the Expires HTTP header so that caching both in the client and intermediaries is possible. This may significantly improve performance in some scenarios, and since it can be done on a per-query basis (since the request URL includes the query as a query parameter, which is used as cache key!), it can be a configuration setting what, when, and how to cache. Same thing goes for EntityState from the REST endpoint of course. /Rickard _______________________________________________ qi4j-dev mailing list [email protected] http://lists.ops4j.org/mailman/listinfo/qi4j-dev
