Starting from first principles, that debate only exists because we organize code as a file system tree. Maybe a tree is not the way we should store and organize code 🤷‍♂️

@sebastienlorber fb figured this out years ago with haste

@douges I'm not super familiar with haste, but for that alternative vision to become a reality, wouldn't this also require IDEs to support it properly?

@sebastienlorber yeah, whole toolchain 😄

@douges Why nobody tells us how all this works?

@sebastienlorber @douges TL;DR globally unique file names inside your repo/project, import just by the file name, you don’t need to worry about the path - even when moving files around. Many benefits to it. I enjoyed using it so much at work that I adopted this in my personal oss repos as well.

@evyataral @sebastienlorber @douges I made in one of the prev works this project that prevent two exports with the same name github.com/oriSomething/n…

@oriSomething @sebastienlorber @douges Nice. Yep, that’s a big part of it. Avoiding name collisions

@evyataral @sebastienlorber @douges Sadly, not many devs appreciate this. They prefer to save 5 seconds now over saving days later

@oriSomething @evyataral @sebastienlorber it's true but it's also a scale thing, meta has done fantastically scaling their engineering practices to a huge amount of people if you're a small person team it's hard to appreciate some of these things even if there is still a benefit to be had

@douges @oriSomething @sebastienlorber I think these practices and configurations also makes sense for smaller projects though. I’m the sole developer in most of my oss projects and I still find them highly beneficial in terms of personal productivity.