Chris Lattner Formally Steps Down From Swift's Core Team

[phoronix]

Phoronix: Chris Lattner Formally Steps Down From Swift's Core Team

LLVM project founder Chris Lattner who began developing the Swift programming language back in 2010 during his time at Apple is now leaving that programming language's core team...

https://www.phoronix.com/scan.php?page=news_item&px=Lattner-Leaves-Swift-Core-Team

[Ironmask]

Figures the only news this language ever gets involves some kind of drama. It's such a pathetic attempt at gaining a monopoly. Literally just Go 2.0, nothing but a soulless corporate husk.

[grigi]

The whole discussion of management trying hard to understand the issues for "future" reasons, but not being able to do anything says volumes about the management structure. The management he dealt with themselves feel powerless, and they are hunting for ammunition to help them.
In my experience the chances of getting a silver bullet that actually works is close to zero, but it's better to try than to give up.

I feel sorry for Chris.

And the dynamic of the features being planned behind closed doors also feels like a small group of people have the power, and they consider everyone else beneath them is another common toxic one I had to deal with.

I really feel for Chris, having something you cared about taken away from you and twisted into something unrecognizable is really hard.

[gggeek]

The rebuttal quote from Chris is something I could have very well written myself a few years ago, at a time when I was working as part of a company developing an oss product, in close association with but but not being embedded in the core development team.

I see this pattern a lot, and not only in corporate environments: every community has a natural tendency of developing an "inner cabal" group, made of the people who invest more in the project. And it takes a lot of wisdom, humility and foresight to identify and possibly break free from the dark patterns that develop because of this.

Projects managed by a corporate entity will suffer both from that, and from the naturally occurring friction between the loose nature of community involvement, and the desire for control of managers of any level of experience. Not involving the community in reviewing the design before starting implementation is a cardinal sin in my book, but I also know that it is hard to have the community interact with on tight, precise deadlines - after all, they are not getting paid, so why would they even attend all the boring meetings where such things are discussed?
The corporate developers are the ones who will generally suffer more from the situation: they will naturally empathize with the community devs, but will not be able to act on their desires, as they will have to follow the corporate plan instead - often with an explicit recommendation to not even talk about it! Having a big community also means a big backlog of tickets, and, unless someone is willing to invest a lot in bug triaging, that will also be stress inducing.
Last but not least, having a huge community can be a boost to morale, but it also means that there will be many who act as spoiled children, demanding constant attention, complaining loudly when their pet peeves and crazy ideas are not taken into account, and giving back very little in exchange. This is particularly true when the relationship between the parties is skewed, with the community ending up acting the "children" dependency relationship with the "parent" company.

Having said that, not all companies are the same, and even within the same project, management comes and goes, and the tide can shift pretty quickly. So, never give up something as long as there is a glimmer of hope!

[timofonic]

Apple management seems even worse than Microsoft and Gnome...

At least he didn't suffer depression. Good luck, sorry for the past and good he left that shitstorm.

[Alexmitter]

That seems to just be how apple manages projects. Apple as a company is used to a community that just eats whatever apple serves and be proud about that. A cult like following. But the moment Apple managers feel real disagreement, they go full bezerk and behave like someone insulted their mom.

Has no one ever noticed why there is zero users of Darwin outside Apple and why nearly everyone runs away from Webkit as fast as possible? Its horrible to work with Apple.

[cesarcafe]

Designing Swift was a horrible decision. Both C and Objective C are solid rock. If the time spent in Swift had been invested in doing other things, developing for Apple devices would have been improved instead of becoming unfriendly.

[skeevy420]

lofty goals, fixed schedules

Doesn't matter the industry: pick one, you don't get both.

There's a reason the Fedora schedule is more like a loose guideline than a fixed schedule. And it really sucks working with managers that just do not get that.

Best of luck going forwards, Chris.

[amxfonseca]

Designing Swift was a horrible decision. Both C and Objective C are solid rock. If the time spent in Swift had been invested in doing other things, developing for Apple devices would have been improved instead of becoming unfriendly.

I don't think a lot of people agree with that statement, unless just for the sake of finding something to complain. I have done some work with Cocoa/UIKit back in the iOS 4 days, and now recently revisited the more modern framework using Swift.

I can say that almost everything is an improvement compared to the old way of doing stuff. Objective-C was just old and outdated. It didn't provide any modern language features that we can currently enjoy with Swift. I wouldn't call it unfriendly, it's actually a super simple language to get started with.

It's safer (not Rust kind of safe, but way better than C/Objective-C), supports more expressive functional patterns, proper structs, functions as first class citizen, better string manipulation. I mean, the list would go on and on. And most of these improvements would not be possible to implement in Objective-C, because that would break the compatibility with C.

With this I am not saying that swift is the best language. But it's miles ahead of any of the other C derivatives for its intended purpose.

Now Xcode is quite a bit cumbersome, I don't like that at all. But that's not related with Swift at all.

[kpedersen]

With this I am not saying that swift is the best language. But it's miles ahead of any of the other C derivatives for its intended purpose..

With Swift, developers lost the ability to directly consume C and C++ APIs via Obj-C and Obj-C++ respectively. Now they need to find and maintain bindings which tend to rot quite badly. To be fair, Rust suffers from this too; the FFI / bindgen stuff is not a substitute for direct inclusion of a header.

So whilst C and C++ are not the most comfortable languages to develop with, I really do see Objective-C making a comeback and ultimately outliving Swift.