Miscellaneous Questions about tartiflette

[tsunammis]

Hi syrus :)

Thanks for the testing & feedback. Here are some answers to some of the questions and miscellaneous remarks.

First off, please bear in mind that the current implementation is still under heavy development: many things are expected to change in the coming weeks as we're still exploring and trying out various options. Benchmarking the framework now has little meaning as pretty much everything is expected to change in the following weeks.

Now, about your questions:

• It's true some type definitions are similar to GraphQL-core, but as many things are still under heavy development we need this flexibility to try out things. We also expect some of these type definitions to be tightly integrated with the execution engine which might not be as straightforward with GraphQL-core.
• We expect to add many layers of optimizations on the current tartiflette engine so most probably we'll find ways to reduce this.
• We haven't fixed the immutability vs mutability subject yet as we're still changing many things. I guess this will depend on performance and developer experience.
• This comes from our experience running GraphQL under high traffic and with the fact that most of our resolvers are asynchronous to work with our micro-services architecture.

Finally, we're always happy to discuss this once our framework will be stabilized :)

[achedeuzot]

Closing this for now.

[jedie]

I noticed Performance issues with large data sets with graphene, too: graphql-python/graphene#268
So i test tartiflette by making a small benchmark and compare with graphql-core...

But I am surprised that tartiflette (~57ms) is even significantly slower than graphql-core (~20ms).

My benchmark code:

tartiflette: https://gist.github.com/jedie/45ddf8ee7e24704c9485eb8cbcf9ba13
graphql-core: https://gist.github.com/jedie/581444e02e784ff7c2b9fb1e763759fa
Django REST-Framework (~8ms): https://gist.github.com/jedie/1d658a184eb4435383820aa0c647d7e9

Any comments?