Isn’t it better to approach async from coroutines ...

# datascience

[orangy]

Isn’t it better to approach async from coroutines point of view for Kotlin?

[thomasnield]

orangy: Yeah that might be a good go-to since it is native to Kotlin.

[thomasnield]

I wonder how coroutines interop with Rx. I need to explore that.

[orangy]

Suspension is a natural point where traveling across nodes can occur transparently, but that would need serialization of coroutine state, I suppose. @elizarov ?

[kyonifer]

agreed, especially since the stackless limitation was dropped, it seems like the go-to solution. a lot less complicated than quasar's post instrumentation

[elizarov]

@thomasnield Take a look at this guide. It explicitly addresses coroutines-Rx interop: https://github.com/Kotlin/kotlinx.coroutines/blob/master/reactive/coroutines-guide-reactive.md

[elizarov]

@orangy Coroutines are serializable good enough for datascience needs

[thomasnield]

@elizarov bookmarked, thank you. I probably should include this in my upcoming Packt book on RxJava, in the chapter covering Kotlin interop.