I don't know any GraphQL, and when it came time to implement a GraphQL client, I spent half an hour trying to figure out what it's about, and then simply chose another supplier. It may be "self-documenting", but not for those that didn't go through the learning curve.
The learning curve may be short, or worth it, or fun if you're that kind of person. But don't tell me that all you need is a http client because that's simply not true, not "out there". And I'm certainly not about to subject hundreds of my clients' clients to having to learn something new just to implement a simple API.
Of course, you need a GraphQL API first. Apollo is good if you want to resolve all your entities manually, but there are novel integrated database approaches like Dgraph that generate everything based on the schema.
I hope you're joking. What is this "resolving all your entities manually" you speak of? Apollo? Dgraph? Is it Klingon?
What you're telling me that it's doable... once you go through the learning curve. And if you'll allow me a testy comment (not personal), it's exactly this lack of empathy toward end users that doesn't make me wanna try it.
What would help would be an already generated REST API I can use as an alternative.
I think the other person's point is that a GraphQL API is not a trivial leap from normal http; it's a significant abstraction layer of its own that requires its own learning process
