It is really great that people are discovering the utility of triplestores through their ubiquity in the Clojure ecosystem (e.g. Datomic and its many clones).

I do feel saddened that few seem to realise that the fundamental idea actually comes from RDF (i.e. the semantic web) and that Datomic's Datalog dialect is actually very similar to SPARQL. Datomic's primary innovation is the immutability/time travel feature, not the fact that is uses triples and works like a graph database. If you find triplestores fascinating I implore you to explore the semantic web stack too.