| > it's largely locked in the apple walled garden with tokens given to the outside So, the compiler, stdlib and runtime, core libraries, build system ... not enough? What else would you want? I feel the problem is not what's in open source, but that the open-source community cannot really form, since no outsider can significantly change what the Apple contributors decide. Some of the peripheral projects have relatively free rein, but they can't compete e.g., with server libraries elsewhere. Also, the Apple people have to track what Apple needs, so they'll put out stuff per schedule that works for them but falls apart on untested code paths. And they don't really deprecate stuff, so you end up with multiple ways to do the same thing. And there seems to be no budget for documentation, which is quite far behind esp. for concurrent programming. And so it goes. We'll see where they get with ownership and inter-op with C/C++/Java. Concurrency + ownership + legacy + FFI inter-op => combinatoric complexity... |
First class IDE support (Xcode is not that). Better documentation. The build system only half works. Better packaging. I can keep going.
The swift ecosystem often feels like just enough was done to lock in iOS devs (but not enough to actually provide a good developer experience) and then they stopped because Apple has no incentive to do more than that.
> I feel the problem is not what's in open source, but that the open-source community cannot really form, since no outsider can significantly change what the Apple contributors decide. Some of the peripheral projects have relatively free rein, but they can't compete e.g., with server libraries elsewhere.
So you agree. This is exactly the point I was making.