Yes, but I am trying to understand what is the most natural use case for Logica solution-wise. Logical programming brings very convenient data models: I do not know about this Logica on top of BigQuery, but for example with semantic web technologies it’s possible to just put triples together to combine the data. I suspect this is not so easy with Logica. One can create “schemaless” in relational database as well with about 7 tables (there is a three volume cookbook, which describes it) — no need for Mongo. The question however is: What Logica is good for? Is it just for Prolog-savvy developers’ brains? Or does it have some technical superiority? And as it’s primary storage is BigQuery, is it good only in combination with it?

Logical programming is very powerful mechanism, but it has been historically computationally extensive. Maybe, the whole reason to push Logica is to sell BigQuery better — I do not know, but it’s intriguing.

Also about application code. With logical programming domain model can entirely in the “rules” part of the knowledge base. When powered with powerful enough generic inference engine, it’s basically all one needs there. It’s next level compared to relational databases or Mongos. This means, that application code needs to be really decoupled from the domain if one (as you describe) is not to change it when switching from relational/document/graph db to semantic web/logical programming like solution. This may mean, that at last there is no need to leak domain into persistence layer: It’s rightfully there. And that is my main doubt about Logica: Does it really provide this kind of possibility or is it just a different query language?