4 ms·
A Turing tarpit can do the same, too. Where are your arguments? There are exactly zero and now everyone in the thread has to work to debunk these baseless clai
by quonn 5y ago
A Turing tarpit can do the same, too.
Where are your arguments? There are exactly zero and now everyone in the thread has to work to debunk these baseless claims.
There are exactly two almost valid points in here:
The default implementation has some overhead (for query validation). That may cause performance issues.
If you deliver public data where the full response can be cached, then REST makes that slightly easier (GraphQL discourages transport by HTTP GET).
I have worked with both for many years, GraphQL solves many real world problems. However you might only understand that as a full stack or frontend developer. If you only write APIs on the backend it appears pointless of course.
- jlouis 5y agoIt can be argued you can remove query validation from the critical path entirely. Though in my experience, it's a very small fraction of the actual processing time by Amdahl.
- quonn 5y agoCorrect, it can be removed. It depends on what you do. I worked on an application that could be cached heavily and got a very large number of requests and as a result the query parsing and validation mattered.