# Asynchronous Programming

Coroutines again, this time in a very practical context: ‹asyncio›
(finally). We will be writing both clients and servers using a
modern approach based on an IO dispatch loop (bundled with
‹asyncio›) and suspending coroutines. I am sure you will be happy to
learn that ‹asyncio› is essentially what ‹node.js› is for JavaScript
(and we all know how popular that is). Low latency, high-throughput
applications, here we come (yes, I know it's Python).

Demonstrations:

 1. (to be done)

Practice exercises:

 1. ‹sem›      – semaphore synchronisation in ‹asyncio›
 2. ‹proc›     – ‹asyncio› processes
 3. ‹multi›    – more processes
 4. ‹tcp›      – a simple TCP echo server
 5. ‹http›     – an HTTP client with a subprocess
 6. ‹merge›    – process data from multiple sockets

Regular exercises:

 1. ‹sleep›    – sleep, running tasks in parallel
 2. ‹counter›  – two-way communication with a process
 3. ‹pipeline› – multi-stage asynchronous processing
 4. ‹tokenize› – another stream pipeline exercise
 5. ‹minilisp› – an asynchronous parser
 6. ‹rot13›    – listening on UNIX domain sockets

Voluntary exercises:

 1. (nothing here yet)