From the early code snippets of the first chapter up to these final recommendations on the behavior of subscriptions and disposal, you now have all the keys you need to implement functional and reactive applications that are readable, extensible, and robust.

The creation decision tree (list) is very useful when you are not fully at ease with all the available operators. Even after some time, it is a good idea to go through it again to avoid the tendency of always using the same operators even when more appropriate ones could be used.

It has been explained previously in this book but it is worth reminding you here: side-effects should be isolated from pure code, and their footprint should be reduced to the strict minimum. This often mean splitting some existing code into several parts, but this additional work is usually rewarded quickly with code that is easier to test. This also paves the way for sans-I/O protocols whose implementations can be reused on other transport layers.

The method provided here to debug applications that do not behave as expected is a reminder of Chapter 9, Operators in RxPY, but it is currently the only way to track down issues.

Finally, the last recommendations are things that you should know, but that are easy to forget when you are focused on developing or debugging an application. Never forget the base principles of observables and observers, and your code will behave the way you want!

By now you know (almost) everything on ReactiveX and RxPY, and you have all the knowledge necessary to write reactive AsyncIO applications. Experiment, play with, and use ReactiveX in your applications; the effort is worth it, and you will be surprised how easy it is to implement complex behaviors with it.