Open namedgraph opened 4 years ago
That paper is cool as hell 'cause it has practical implications for optimizers and query rewriters. They used Coq + Rosette to find bugs in deployed code and prove query equivalence, two capabilities which would could improve SPARQL implementations.
You could probably reuse their join semantics for both BGPs and joins, and all of aggregation, so I guess you'd have a fair amount to show after modelling triple patterns (though property paths might be tricky).
Why?
Formal provers are used to verify the correctness of programs and mathematical proofs.
Previous work
Proposed solution
Write down SPARQL algebra using the Lean prover
Considerations for backward compatibility
None