& Server ▶ Websocket Client & Server ▶ Rich eco-system ▶ Python 3.4 (but 3.5 recommended) Why? Performance Is it worth the trouble? Actually: ▶ It’s very similar to regular Django/Flask ▶ It’s easier when you need realtime features
(asyncpg, asyncpgsa, aiopg, …) ▶ Templating → jinja2 ▶ REST → you don’t need a framework for that And you have proper nested app support eg. aiohttp-debugtoolbar.
new async/await keywords that you must drop here and there the code looks like its synchronous counterpart. ▶ True: a web handler can be cancelled cf. aiojobs for a way to handle that case.
HTTP and websocket. ▶ There’s a low level server API and high level server framework. ▶ You get to choose your prefered tools (databases, template, validation…).
very good thing. Flask and Django are very keen tendency to rely on globals (like PHP 3) which aiohttp fully avoid. This makes the code of aiohttp much more readable and future proof.
language its usuability is good and will improve. ▶ aiohttp is not monolithic like its big brother Django. It’s a good thing. Also for the common cases libraries are available.