Jul 25, 2017 · 1 min read
It would be interesting to check using HTTP2, how much longer it takes for REST to do multiple requests in parallel compare to one.
On top of that, if you have a layer of cache on the client, I don’t think getting everything all the time would be a always the best strategy.
It is though way easier to interact with GraphQL than building around REST endpoints and having to implement cache and all.
Maybe the end solution is a REST Api (good for formalisation) with a GraphQL layer on top of it? I should stop here, I’m starting to over-engineer :).
