Hacker News new | ask | show | jobs
by randallsquared 1911 days ago
Why would you want to revisit these debates, having progressed beyond them? :)

Instead of rehashing "Is OOP good, actually?", we should explore the boundaries of the paradigm by making a popular language that's based on prototypes, or that privileges composition over inheritance, or pulls apart the various pieces of OOP into standalone concepts that can be mixed and matched. If we get inspired, maybe come up with some new pattern or language affordance that's similarly groundbreaking to functions, objects, classes, message-passing, threads, or whatever. Maybe the obviously-better development paradigm of 2050 hasn't been articulated, yet.
2 comments
meheleventyone 1911 days ago
> Why would you want to revisit these debates, having progressed beyond them? :)

A moment of zen like grumpiness where I was motivated to complain about it.

And yeah I totally agree that the PL grudge match is better solved by experimenting with paradigms than drawing battle lines on a forum. I was aiming at being even more meta though.

link
jt2190 1911 days ago
> we should explore the boundaries of the paradigm by making a popular language that's based on prototypes, or that privileges composition over inheritance, or pulls apart the various pieces of OOP into standalone concepts that can be mixed and matched.

JavaScript? ;-)

link
randallsquared 1911 days ago
I can neither confirm nor deny that I had any particular languages in mind... but, yeah, having slightly or wildly different takes on various patterns can help us explore the practical consequences. I would suggest that the history of JS is good evidence that prototypes are harder to use and reason about than classes.
link
meheleventyone 1911 days ago
You joke but I think TypeScript's structural typing is really cool and has a lot of neat advantages of nominal typing.
link