It takes "limit" and "skip" into account. E.g. you have a collection with 10 documents, want to retrieve all of them and skip first 5. Mongo Java driver's `cursor(q).count` will return 10, but mongo-scala-driver's "coll.size" method will return 5 -- as this is what you'd expect, because you have a Scala collection.