Another week, another submission that hopefully edges more people towards realizing that they've just created an under-standardized, under-powered SQL-with-braackets that forces you to bring your own query optimizer, and eschews the advancements of Swagger/OpenAPI, JSONSchema, HAL, JSON-LD and the flexibility of HATEOAS.
All for a little bit of a velocity gain from declarative syntax and horizontal/vertical filtering.
This code is probably very useful to the people who made it and those in the GraphQL+Swift ecosystem, but we were already here (and just starting to get drastically better) with OpenAPI -- it's not like all the effort is only going in one direction but GraphQL really feels like a less-than-optimal branch.
The final piece of the puzzle for me will be when someone starts to replace the dynamic properties of the more advanced REST-ful techniques (API discovery, link following, etc) in GraphQL. Then the circle will be complete -- we'll have rebuilt XML (in the "linked data" respect anyway, not so much the accidental complexity side) and it's related standards twice in under ~20 years.
All for a little bit of a velocity gain from declarative syntax and horizontal/vertical filtering.
This code is probably very useful to the people who made it and those in the GraphQL+Swift ecosystem, but we were already here (and just starting to get drastically better) with OpenAPI -- it's not like all the effort is only going in one direction but GraphQL really feels like a less-than-optimal branch.
The final piece of the puzzle for me will be when someone starts to replace the dynamic properties of the more advanced REST-ful techniques (API discovery, link following, etc) in GraphQL. Then the circle will be complete -- we'll have rebuilt XML (in the "linked data" respect anyway, not so much the accidental complexity side) and it's related standards twice in under ~20 years.