The power of ROS is consistency and API brings very clear error boundaries. GQL would destroy that model making the API actually very hard to use in a long run. Additionally deprecation of GQL is very tricky. However, the biggest hurtle is complexity and performance of the server: there are multiple implementation in majority of the relevant languages and all of them I’ve ever seen lack in multiple areas.
Don’t get me wrong: I would love to tinker with GQL and ROS, but it’s simply not a great fit. To provide some context I’m saying this from a position of a lead developer of a large medical platform extensively utilizing GQL. After couple of years I’m 100% sure the GQL was the best and the worst choice we made
Nice wish - but won’t ever happen IMHO. We can be lucky - very very lucky- should v7 ever reach stable-status. So I guess the API has not the highest priority. And I am 100% sure that Mikrotik would not see any benefit of a GQL API anyway ^^