Database architecture about dannet HOT 1 CLOSED

Reading the documentation once again, it seems like any kind of live system with inferencing is not ideal:

Having bound a Model into an InfModel by using a Reasoner its content can still be changed by the normal add and remove calls to the InfModel. Any such change the model will usually cause all current deductions and temporary rules to be discarded and inference will start again from scratch at the next query. Some reasoners, such as the RETE-based forward rule engine, can work incrementally.

Basically, if any changes are made to the graph, subsequent queries end up dog slow until inferencing has caught up. This is not ideal at all.

It seems like the only valid solution is to have two graphs:

• A TDB-based graph for editing. This would have no inferencing capabilities. It would therefore only be a place to make edits to the dataset.
• A graph for viewing which does contain inferred triples.
  • This can either be an in-memory graph or TDB graph.
    • I guess TDB is the best trade-off as it will allow the system to be restarted without having to rebuild the graph.
    • An in-memory graph would be preferable in cases where rebuilding the graph is not an issue (and memory is plentiful).

Furthermore, when it comes to exporting data, it seems like we will also need the capability to create adhoc InfModel graph containing e.g. the inferred opposite relations. Otherwise, the alternative is simply to only have the base dataset available as an export file.

from dannet.