Yay another abstraction layer to add the mix
A few questions though, any reason your endpoints don’t allow for custom queries? like:
/resource/{id}?fields=field1,field2,field3&limit=10
and would it not be better to specify what you need server-side as to not expose vulnerable data?
You mention we can white-list certain queries, but does that not invalidate the point of the problem we are trying to fix? Not having a backend dude having to create our custom endpoint.
Instead he now has to white-list a ton of queries instead of simply adding a new endpoint with the exact data we need.
When you mention we have to take 6 round trips to the server to get 6 different film requests why oh why don’t you have a films/?get=1,2,3,4,5,6 endpoint?
It solves the problem of 6 queries, and takes very little time to implement
GraphQL seems to add a new layer on top of an already saturated ecosystem. It is indeed clever and comes with powerfull features, but it only seems to move the problem rather than solving it.
