Guido van Rossum

You could say that it spiraled out of control a little bit.

At the time, it was the largest Python enhancement proposal that was ever proposed.

At the time, I was very much involved with that.

I was like the lead architect of

I ended up talking to people who had already developed serious third-party libraries that did similar things and sort of taking ideas from them and getting their feedback on my design.

And eventually we put it in the standard library.

And after a few years, I got distracted.

I think the big thing that distracted me was actually type annotations.

But other people kept it alive and kicking, and it's been quite successful, actually, in the world of Python web clients.

There are a couple of different ways you can handle parallel I.O.

And this happens sort of at an architectural level in operating systems as well.

Like Windows prefers to do it one way and Unix prefers to do it the other way.

You have an object that represents a network endpoint, say a connection with a web browser that's your client.

And say you're waiting for an incoming request.

Two fundamental approaches are... Okay, I'm waiting for an incoming request.

I'm doing something else.

Come wake me up.

Or sort of come tell me when something interesting happened.

Like a packet came in on that network connection.

And the other paradigm is...