Future, akka-stream, scalaz-stream

@viktorklang @a4dev More like an abstraction induced limitation. :-) Useful tools that should work on Future but don’t understand EC.

2014-09-20 10:38:11

@djspiewak I'm perpetually surprised at the low standard there is for encoding protocols with state machines in modern libraries/languages.

2014-09-20 10:38:17

@viktorklang @a4dev I have application level code that runs into similar problems, but it’s harder to exemplify in a tweet.

2014-09-20 10:38:22

@djspiewak Make an ExecutionContextT and run over that instead? :-) /cc @a4dev

2014-09-20 10:39:22

@viktorklang It really is astonishing! It seems like such an obvious thing to make elegant.

2014-09-20 10:39:45

@viktorklang @a4dev Well yes, that would sort of do it. That’s effectively what Scalaz’s FutureMonad does internally.

2014-09-20 10:40:30

@vikasdp I could write a blog post on Scala Futures, @djspiewak is much more knowledgeable than I when it comes to scalaz Futures.

2014-09-21 00:13:36

@nuttycom 1)You create a promise, add computation to its future, you pass that future along, -> you complete the promise at your discretion.

2014-09-21 00:37:23

@nuttycom or 2)you create your initial future with an EC you control the execution of, and then pass the future on to 3rd party. then run EC

2014-09-21 00:38:19

@viktorklang @djspiewak Viktor, I don’t understand your statement here. RxJava is push unless the consumer is slow then it flips.

2014-09-23 02:08:03

@viktorklang @djspiewak Ah, if I understand, you’re saying that Akka Streams and RxJava are dynamic push-pull, is that correct?

2014-09-23 02:10:22

@benjchristensen @viktorklang We were saying that earlier, Ben, just imprecisely. Akka streams and RxJava both do the same thing there.

2014-09-23 02:10:55

@benjchristensen @viktorklang Yep! In truth, scalaz-stream is also hybrid push/pull, but explicit (not dynamic) and slightly obfuscated.

2014-09-23 02:11:55

@djspiewak @viktorklang got it, as this space evolves terminology changes too :-) and I’ve proven to be bad at naming things.

2014-09-23 02:14:38

@benjchristensen @djspiewak yes exactly :)

2014-09-23 02:15:53

@djspiewak @viktorklang “reactive pull” is the name @headinthebox and I have been using for this dynamic push/pull model

2014-09-23 02:16:01

@benjchristensen @viktorklang @headinthebox Oh great, *another* meaning for the word "reactive"… ;-)

2014-09-23 02:21:17

@benjchristensen @viktorklang @headinthebox I think hybrid push/pull is a little more descriptive.

2014-09-23 02:21:32

@benjchristensen @viktorklang @headinthebox The real problem is that basically all serious stream frameworks are hybrids in some way.

2014-09-23 02:21:52