The steelman of their argument (coming from things as a database person) is thinking about the problem in a different way for a different domain, which can beget a "simpler" approach.
Something like a document db approach (one big json entry to rule them all) gets significant benefits in setting certain levels of abstraction as even possible.
Hey - we just get the user details and then we update some json and that's all we can do, and it kinda sucks, but it means we're going to put most of our computation in this website and really make something that looks like a modern web app.
It has a bunch of downsides, but you dont have to teach anyone SQL, its just a simple repository pattern + microservices + limiting the scope of any one thing.
Type checking is strong, tooling support is strong and you usually can stick in one language and go ham.
Something like a document db approach (one big json entry to rule them all) gets significant benefits in setting certain levels of abstraction as even possible.
Hey - we just get the user details and then we update some json and that's all we can do, and it kinda sucks, but it means we're going to put most of our computation in this website and really make something that looks like a modern web app.
It has a bunch of downsides, but you dont have to teach anyone SQL, its just a simple repository pattern + microservices + limiting the scope of any one thing.
Type checking is strong, tooling support is strong and you usually can stick in one language and go ham.