The problem with complex systems is that you can be dead long before you realize you're dead. Things can continue to seem "pretty good" for a long time just from the inertia of past good decisions and built-up infrastructure. You see some fraying or cracks but everything looks fundamentally sound, until it isn't.
Apple's inability to deploy a new UI framework that's better than the previous one is troubling. This is the company that shifted from Toolbox to Carbon to Cocoa and each step was better than the last. Could Apple build MacOS X and Cocoa today if they didn't already exist? Microsoft's two decades of failure to ship a true successor to Win32 suggests that Microsoft (at least, the operating system side of the company) died a long time ago. I wonder if we'll say the same thing about Apple 10 years hence.
I think this opinion is heavily shared by Swift developers now, but the messaging every WWDC is always "Swift and SwiftUI is the best way to build apps for Apple Platforms", etc.. If you have to keep telling everyone what they don't believe is true, it's a sign there's a problem.
It feels like someone with a lot of organizational power is disconnected from the pulse of the community. SwiftUI is undeniably clean in a lot of ways, it presents beautifully and fits on slides well, but that matters less and less, and this all wasn't really working out even before LLMs disrupted things.
It’s a seductive idea, but it displaces who is responsible for the poor framework development onto users of the framework.
It’s easy to say they were just trying to be trendy and thus the real flaw was the trend is bad and the fault of worse devs than us (the young web devs mentioned)
The thing that obviates that is the trendy stuff works, yet, SwiftUI doesn’t.
(source: I wrote ObjC/Cocoa as early as 2006, and switched to Flutter as my primary dev kit some years ago: it simply doesn’t have the performance issues mentioned.)
I think the main issue is the functional approach. Having a functional approach to state can be quite elegant, but ultimately the computer does not do functional. You have to implement a lot of plumbing to have stuff that is a bit performance (clojure), or just add a veneer of functional over what is essentially a normal state machine (emacs).
React works well, because it's only an abstraction over the real DOM. React only handles your app state. But the DOM mechanism is still very performant and very much imperative. But I don't think something like React would work well in a mobile app, because the UI tree is often very simple on iOS and macOS.
Imo the biggest issue with this functional model (at least in React), is that it handles things like virtualization, async, etc. poorly. Which is kinda ironic, because in a true functional language, it'd be feasible to provide 'a world model' - that is act as if the entire state is always available, and let the engine decide when to evaluate pieces of code - without any effort from the part of the programmer.
UI is also very much not functional, and in fact the lack of progress in UI the last 30-40 years can largely be traced to trying to create UI with procedural/functional programming languages, an instance of linguistic-architectural mismatch.
Further reading:
Programs = Data + Algorithms + Architecture: Consequences for Interactive Software Engineering -- Stéphane Chatty.
Functional doesn't mean stateless. I'd argue functional programming is superior for representing state in user interfaces. For a success implementation, see Jane Street's Bonsai:
> functional programming is superior for representing state in user interfaces
It's always going to be slower than using something imperative. Trying to process the entire world's state for the sake of purity can feel elegant but it isn't free. And the complexity that gets added to make things performant is worse than just accepting that UIs are going to require you to jump around the tree and modify state.
Every time I go through the trouble of understanding the latest web technique (React, Elm, Signals (the newest solution), etc.) to deal with state management and the DOM, I end up walking away disappointed. There's nothing new in them that you can use to improve what we've been doing for ages in native GUI toolkits.
And to jump back to the original topic, yes, Cocoa was pretty decent, and SwiftUI while nice in many ways tries to Reactify native macOS development and made it worse. And it made Swift incredibly more complex and worse in hindsight.
The goal of the reactive/declarative approach was never to be more performant than imperative code. The goal is to more easily build UI that is performant enough and functions correctly. With imperative UI code it is incredibly easy to forget an edge case in your update logic.
User interfaces representation are mostly trees. And with functional programming you basically have Tree2 = f(Tree1). Until f is done you can't do anything really. React has a lot of escape hatches to improve performance, but they are escape hatches, not an endorsement of the architecture.
With imperative programming (and OOP), you only have that single `Tree`, which you update at will. Less elegant yes, but we have modularization to help us there. What Emacs does is to keep that `Tree` as a single mutable object, but have the code be functional, while the results are imperative.
I’m no huge apple fan, but I didn’t really think the 17 was more than a regular boring spec upgrade, which seems fine all things considered. (Currently on an iPhone fucking 17)
This seems to be a popular ‘controversial’ topic. I’ve been using SwiftUI for production applications and games since 2021. I drop down UIKit, Metal, or core animation when needed. But that’s no different than when I was making games in UIKit and would drop down to core animation or glyph renderers written in C etc.
Data Flow: The author claims there’s no way to know when things update. Not only does experience help here but there are profile tools that tell you when and where things are updated. This isn’t black magic. Keep Views small, be careful how you hand around data. @environment is super cool but can have a cascade effect. This was greatly improved iOS 17+ and I wouldn’t support anything older than iOS 17.
GeometryReader: Occasionally I’ll use this. It’s kind of a necessary evil when dealing with certain view complexity. It can also be a sign that you’re doing something wrong.
API Stability and performance: Apple users upgrade. There’s no reason that you should be supporting iOS 17 at this point — even iOS 18 is roughly 2%% of our user base across several apps. I’ve been using SwiftUI without major performance issues but I also don’t early optimize. I profile and fix as needed. One of the early studios I worked for wrote all our games in UIKit as prototypes, when performance tanked we’d switch to the appropriate tools (eg. OpenGL) where it was necessary — like in the core game.
I could go on but in the end just use the right tool for the job, if you’re not proficient in SwiftUI or it isn’t going to work for your cross platform project, you have a lot of other alternatives. For me though, it’s been amazing to work with. I stepped away from iOS programming for 3 - 4 years because I was burnt out using storyboards, dealing with massive view controllers, and all the boiler plate it’d take to get a view up and going in UIKit. SwiftUI roped me back in.
* Quick addition edit: Cross platform for iOS, iPad, macOS has never been good. I’ve found recent updates have made things better to the point of tolerable and it’s nothing like when we had to post-fix an ~ipad to our Nibs — There has never been a ‘glory days’ of cross platform Apple UI libraries.
I'm sure many will disagree, but I have doubts that pure declarative-reactive is the "right" shape for an all-purpose native UI framework. In my experience, Kotlin+Compose shares many of the same warts… its main redeeming quality is that it's better than Android Framework (most of the time), which is low bar to clear.
These frameworks have a number of good ideas but they don't necessarily combine in a way that transcends high quality traditional imperative frameworks with declarative-reactive bits sprinkled throughout, at least for more complex apps. SwiftUI and its ilk work best for super simple tabs-and-flat-lists sorts of apps.
Do you have an example of where "the ideas don't combine well"? Or where an imperative approach is truly better?
In my experience the solution is often to use the appropriate architecture underneath your UI layer. Basically you build a data structure that represents your UI and always hand this entire structure to the declarative UI layer. Underneath the UI layer, you can still use imperative code to manipulate the data structures. The benefit of the reactive/declarative approach is that you don't have to think about how changes in the data need to be reflected in the UI. That's the framework's job.
Note: With "data structure" I don't mean "build a shadow DOM". I mean something specialized to your use case.
I think open source is the right move here. The great thing about Flutter is you get the declarative UI to build a lot of standard screens easily but you can always drop down to lower levels, inspect the standard components, mix and match with your custom implementation.
You can mix UKit and SwiftUI but there’s way more friction due to the closed nature of the SDK.
Fully agree (been an early Flutter adopter!). It's always the right too for the right job, declarativness makes a lot of sense for most UIs. UIs are surprisingly more complicated than one might think, there is accessibility/ theming/ keyboard traversal etc., just declaring UI and letting the framework figure out the rest has held up pretty well.
And because of the open source nature you can always drop down in layers, just a few weeks ago I made a very specific engine optimization, all while the application code on-top can stay declarative!
Perhaps. Certainly source visibility makes life easier as a dev who has to live with these frameworks on a daily basis. That said, I think frameworks should strive to have a wide and deep enough set of widgets that in most cases, devs won't need to resort to custom implementations to do what they need to.
It also doesn't fix problems inherent to declarative UI, like readability breaking down easily and certain things that would be a cinch in an imperative framework being extremely awkward to implement.
Apple had a real winner with ObjC and AppKit. Swift is horrendously complicated for the benefits it offers over ObjC, and SwiftUI is a massive step backwards from AppKit.
Just MHO, and it’s not going to stop the juggernaut, but there’s just no appeal in moving there for me :( If (when ?) Apple drop ObjC, that’s the day I move to Linux
Apple will likely never drop Obj-C, but it's been a secondary language for years now. New frameworks are all written in Swift, Obj-C is only used for existing Obj-C codebases.
I've noticed that developers who started with UIKit really have a hard time working with SwiftUI. I think this is because it's not just new syntax, it's an entirely different way to think: state is the source of truth, views are ephemeral, and you describe UI instead of managing it.
Apple has been adding Observable support to UIKit as well, so it's easier than ever for state to be the source of truth in UIKit too. Anyone writing more than simple UIKit apps has known that for a long time, it's just big a huge pain to actually implement without a lot of custom code or a dependency like RxSwift. If Apple had embraced reactive programming 15 years ago, SwiftUI would be UIKit's new layout system, not a whole new API.
Personally, SwiftUI makes the 80% so much easier that, even if the remaining 20% requires dropping down to UIKit, it's worth it.
SwiftUI it is the type of framework that makes the easy things easier to accomplish but the harder things harder.
It is a newbie trap. It is great at producing simple apps, or things that don't require intense scrolling, or anything with heavy animations and precise layouts, but when you do something smooth, it is not it. It feels more of a React Native competitor, than a true UIKit replacement.
Also, almost everything that Apple has re-wrote with SwiftUI, came out worse as before (Settings, Spotlight, etc), and that doesn't help.
With AI coding... SwiftUI lost its edge/advantage (easy to layout screens), as now you code less of that directly, and might as well just go fo the framework that allows the most flexibility and gives you the best results for your users.
Apple really needs to either double down on it, and make it such that it has everything that UIKit has (match both features and performance), or just turn it into another optional framework, just as Interface Builder (.xib files) back in the day of Objective-C.
Right now it heavily promotes it as a first class citizen, meanwhile the features are not even parity with UIKit. It is so misleading to newcomers to the platform.
Ps. The other upsetting thing about SwiftUI, is that it has hurt the Swift language as well, as the team has had to introduce all kinds of hacks, or obscure features to the language in order to make it work, which made even the Swift language experience worse and more complicated than it should have been.
> With AI coding... SwiftUI lost its edge/advantage
On the contrary, I think AI has made it EASIER to approach SwiftUI now, because AI makes up for SwiftUI's biggest weaknesses: the many ways of doing everything, Apple's wonky documentation and the need to know when to fall back to AppKit/UIKit.
Maybe for you. But SwiftUI promise was to make coding UI easier for the programmer similar to ReactN, but with AI coding that is not necessary anymore, you can go straight to UIKit which is most stable and feature complete.
In car terms:
UIKit - Manual Transmission - you control everything, but a bit of a chore
SwiftUI - Slushbox Auto Transmission - easier to drive, sucks for performance
UIKIT+AI Coding - Dual Clutch Transmission - both great performance, and easy to control/drive
What does this even mean? There are hundreds of thousands of apps in the App Store that are 100% SwiftUI. They are real. And they are very much "production grade".
The only people still complaining about SwiftUI 7 years later are the UIKit holdouts that never took the time to properly learn how to use it.
Agreed. SwiftData is terrible for anything more than like 2 "Models". I have a bunch of apps written with SwiftData and a bunch more with GRDB... I hate touching the SwiftData ones.
Yeah, I like WPF and I still use it for my own apps. In fact, my own software is the only thing keeping me on Windows now. I really should take a look at Avalonia and port my stuff over to Linux.
WPF was stamped out by lack of dogfooding = inability to inspire confidence, and then Microsoft spazzing out with that clusterfuck that was the Windows 8 "Metro" bullshit.
Back when I wanted it to succeed, Microsoft was barely using it in any of their own apps.
Apple has used SwiftUI in their OS more than MS/Windows did, and there's quite a few apps on the App Store that -seem- to be made with SwiftUI.
I fell in love with SwiftUI the day it was announced, but as a solo dev I still haven't been able to make a full app with it yet, mostly because of the lack of documentation, and the gaps where you still need to drop down to AppKit/UIKit.
I went through Visual Basic, .NET, WPF, Cocoa, and other random frameworks all professing to be the promised panacea for UI, but I think SwiftUI+SwiftData is the best environment ever.. IF only it could reach its full potential, i.e. do everything that Apple's "legacy" APIs can do.
The biggest/worst hurdle in the "modern" dev experience is Apple's insistence on a yearly update cycle and the way they advertise those updates: You have to wait for the next WWDC and suffer through videos of uncanny-valley presenters, hoping to catch a glimpse of something that fixes the shit that was bothering you since the last WWDC.
3rd-party sites like hackingwithswift.com & swiftwithmajid.com provide invaluable info that Apple's own docs should.
At least the Swift language has been getting more regular updates since it went open source.
There's no way I'd dare to take on a full Apple-platform app project alone on my own, but I've started dabbling in it again thanks to AI: Codex even converted an old app I made in Visual Basic 900 years ago and had it running in SwiftUI within minutes!
I've even tried to get AI to sift through the WWDC video transcripts so I won't have to waste my mortal lifespan on that.
Yeah I don't really understand their insistence on these videos. I don't have the time to watch an incredibly hard to watch present babywalk me through the code, hoping I'd catch one modifier I missed that fixes the issue. Just write some docs for god's sake.
> There's no way I'd dare to take on a full Apple-platform app project alone on my own
And that's really sad, because earlier Mac OS X (and even OPENSTEP to a degree) was an Eldorado for boutique developers, not least due to frameworks like Cocoa and Objective-C (despite all the hate it seems to get).
If you knew C, it was not hard to grasp Objective-C (at least it was much easier than say C++ and MFC on the Windows side of things).
That's actually why I jumped from Windows to Mac in the first place!
My first Apple purchase was an iPad I got for my aunt, and I loved how streamlined yet powerful it was so I installed a VM on Windows to dabble in iOS development, and fell in love with the sheer consistency of macOS, iOS and their APIs.
This was during the era when Microsoft was still figuring out whether to keep the Start Menu and replace the whole Windows UX with a tablet UI, so it was an easy jump.
Almost never looked back, and whenever I do I'm glad I jumped.
Then came Swift and I loved it, but then it took literal years to catch up with everything you could do in Objective-C with the Apple APIs. SwiftUI is having the same problem but for much longer :(
What I like about Kotlin Multiplatform (KMP) vs Flutter is that KMP is not a framework. For instance, I can write a KMP library and a "pure" Android or iOS app can pretend it is a "pure" Android/iOS library.
That is, pure native apps, can integrate KMP libraries. They cannot integrate Flutter libraries.
In other words, if I write a KMP library, I can ship it for multiple platforms and developers can use it in their Flutter project, or in their pure Android/Kotlin project, or in their pure iOS project.
SwiftUI is good, albeit not without issues, as long as you tolerate them. The same is true for any approach. My most recent app was built in SwiftUI because I have been using React Native in my professional work for so long I wanted a change of scenery.
Each approach has its issues, and it all boils down to a matter of preference and which tradeoffs are appropriate for the project.
I’ve worked on some award winning apps in the past, our experience with SwiftUI is it makes it really easy to develop a basic average looking app. As soon as you need something non-standard, you have to drop down to UIKit and figure out how to play nice between the two, which is often non trivial and hard to estimate. You file radars, which at best will be fixed in the next OS release and rarely back ported if at all.
In the end some of the more complicated apps just ditched SwiftUI altogether or applied a myriad of workarounds.
It's never "just working", you'll still have to debug the vibe-ported apps. And you won't be good at it if your devs don't actually know the target platform.
All that to say, I don't think that vibe-porting will give you more than cross-platform. With cross-platform, at least your engineers understand the codebase.
I know people would dismiss this, but I'm going to try to make the case for Qt - I'm currently converting my note-taking app to mobile and it's going very well.
I have been involved in two non-trivial Qt mobile apps. Never again.
I don't know what your experience is, and I mean no offence or anything, but the only people I have seen who were happy with Qt for mobile development were people who had never tried to do it with a modern language, like Kotlin.
With Compose Multiplatform and Kotlin Multiplatform, you can write a cross-platform app in a modern language, and that is far better than C++.
I heard this repeated many times. You can create closed-source Qt apps without paying any money. You cannot statically link, though; you must obey the LGPL license; that is all.
And yes, you cannot use commercial modules, but that is kinda expected.
Yes, Qt is mature and high-quality software. If you’re doing OSS, then Qt is truly a no brainer. The amount of C++ you have to write is truly minimal, and it’s not “raw” C++. There’s a lot of tooling and libraries. You can target web now too, and QML is legitimately good.
If KDE can develop a high performance stable desktop, you can design whatever.
Same here, but Liquid Glass has been a headache lately. Google really dropped the ball here, but to their credit it almost feels like Liquid Glass was an intentional move by Apple to cause as much friction for cross platform frameworks as possible.
Define right, it’s hard to replicate the native UI especially Liquid Glass on iOS because it’s rendering everything using its own rendering engine, making it hard to feel right on iOS. I have also seen a lot of companies move away from flutter to either native or React Native.
Well we all love Liquid Glass. Most loved UX update in iPhone and macOS history. Massive dealbreaker for an app not to look like Liquid Glass. I definitely choose apps based on whether they look like proper Liquid Glass applications instead of the developer taking shortcuts and focusing on working features instead.
Well I’m a heavy Flutter user myself, and the app I built Hacki often gets reviews saying it doesn’t feel native on iOS. I personally don’t have anything against it and I think Flutter has the BEST developer experience compared to other mobile app dev toolkit/SDKs.
SwiftUI was doomed from the beginning. Not only Apple completely blew the implementation, it was actually DoA by a bunch of super bad decisions that ultimately make it incredibly hard to work with and manage, especially on bigger apps.
1. It uses the builder pattern for views (familiar to those who used Java) but for some reason they decided to make it so that the order of the modifiers in the builder matters. Each modifier doesn't actually modify the main view but it modifies what the modifier before it decided to return. This makes no sense as the modifiers should each be modifying the main view to make everything predictable and easy to debug. I'm convinced no one (not even senior devs with 5 years of SwiftUI experience) understands how the ordering of modifiers works. It's just swap them until it does what you want it to.
2. The view lifecycles and code execution path seem random and hidden behind layers of "magic," making it incredibly difficult for developers to trace and debug issues.
3. It is practically impossible to set breakpoints for rendering and view construction, it's impossible to really figure out when re-renders happen and what drives them. I'm convinced SwiftUI apps are incredibly slow not because the SwiftUI implementation itself is slow, it's because, even Apple's own apps probably do a bunch of unnecessary re-renders and one no one seems to have any idea. This is unfortunately another design issue that can't be solved by just making SwiftUI more efficient. It requires simplification and tooling to help developers not footgun themselves.
4. Lots of issues start appearing later on in the development cycle because, for simple apps, bad SwiftUI design decisions and footguns have unnoticeable effects, until the apps gets more complex and things start breaking. Fixing these issues sometimes requires rewriting whole features or spending hours debugging.
5. There seems to be almost no documentation on Liquid Glass. It's laughable that after more than 1 year, Apple has simply refused to document or provide good examples for Liquid Glass, except for maybe couple pages that resemble the brain dump of an engineer that has never passed a writing class in college?
6. Stuff seems to be rapidly changing and breaking from version to version. It took days to make my app look and work the same in iOS 27 as it did on iOS 26, even though iOS 27 is supposed to be a minor bug fix release. We don't even use anything non-standard and don't do any hacks. This defeats the whole purpose of a simple UI framework that can be easily adopted to different platforms (this never used to happen with UIKit).
7. View debugger still has no SwiftUI equivalent. It used to make things so much simpler in UIKit when you could just see the view bounds, pick views apart and understand what's actually happening. SwiftUI has no equivalent other than `.background(.red)`. Terrible.
I don't know how Apple can salvage this beyond just undoing some of these terrible design decisions and making it 1. super simple to work with, 2. stop relying on magic, making things more explicit, and 3. providing actual 1:1 UIKit feature parity.
Apple seems to be loved by users, but hated by developers (and for good reasons). They just seem to not give a shit about the developer experience.
You mention some examples, I have experience with lower-level stuff: terrible. Poorly documented, and actually not really working well.
IMO they should focus on improving the developer experience instead of wasting money on the joke that is Liquid Glass (even users don't like it, right?). But anyway, too late for me: I'm soooo happy that I can develop iOS apps in Kotlin now, and just endure Xcode for building and occasional debugging.
1. It's not really the builder pattern, which, in Java, is what you describe as wanting: mutating a single instance. SwiftUI uses result builders to compute a single generic View from the "DSL", but you could do the same thing without the custom syntax, it would just be more verbose.
2. Documentation issue. The lifecycle is standardized, but Apple is terrible as documenting it all in one place. I found this year's "Dive into lazy stacks and scrolling with SwiftUI" WWDC video actually had a pretty good explanation, at least for the lifecycle within lazy containers.
3 and 7. Yes, Apple's tools suck pretty bad. Preview canvas is also still terrible, despite multiple attempts to improve.
4. Not something I've experienced, not sure what you mean.
5. Yeah, Apple's docs suck.
6. Not sure what you mean here either, our stuff looks pretty much the same.
Somehow you being the GNOME co-creator made me curious about what the GTK/GIMP creators (Kimball, Mattis) are up to. CockroachDB, it seems. It's weird how it's so easy to miss these connections.
I joked, but reading the post is even more clear - man this poor dude.
He is brute forcing his way through life and broadcasting it on YouTube.
Dunno man, Xogot is made up of almost half a million lines of code. Of those 727 files use SwiftUI, and 65 use UIKit, 119 use AppKit (I did not count the overlap, so they are not mutually exclusive).
Using GeometryReader makes it clear he is living in the past and the fact that he uses printChanges in this year without mentioning the dedicated SwiftUI instrument show a lack of curiosity on the space.
The problem with complex systems is that you can be dead long before you realize you're dead. Things can continue to seem "pretty good" for a long time just from the inertia of past good decisions and built-up infrastructure. You see some fraying or cracks but everything looks fundamentally sound, until it isn't.
Apple's inability to deploy a new UI framework that's better than the previous one is troubling. This is the company that shifted from Toolbox to Carbon to Cocoa and each step was better than the last. Could Apple build MacOS X and Cocoa today if they didn't already exist? Microsoft's two decades of failure to ship a true successor to Win32 suggests that Microsoft (at least, the operating system side of the company) died a long time ago. I wonder if we'll say the same thing about Apple 10 years hence.
I think the problem with SwiftUI in a way was (or is) that Cocoa was so good. And, at least IMHO, it didn't need replacing, just improving.
But I feel like Apple wanted to appeal to young web devs and tried to offer something more similar to what these are used to.
Totally agreed. I kind of figured things were going to get shitty when the guy who made autolayout got hounded on so hard that he left.
I sure hope they can figure out a way through. Whatever it is, SwiftUI isn't it.
> Whatever it is, SwiftUI isn't it.
I think this opinion is heavily shared by Swift developers now, but the messaging every WWDC is always "Swift and SwiftUI is the best way to build apps for Apple Platforms", etc.. If you have to keep telling everyone what they don't believe is true, it's a sign there's a problem.
It feels like someone with a lot of organizational power is disconnected from the pulse of the community. SwiftUI is undeniably clean in a lot of ways, it presents beautifully and fits on slides well, but that matters less and less, and this all wasn't really working out even before LLMs disrupted things.
It’s a seductive idea, but it displaces who is responsible for the poor framework development onto users of the framework.
It’s easy to say they were just trying to be trendy and thus the real flaw was the trend is bad and the fault of worse devs than us (the young web devs mentioned)
The thing that obviates that is the trendy stuff works, yet, SwiftUI doesn’t.
(source: I wrote ObjC/Cocoa as early as 2006, and switched to Flutter as my primary dev kit some years ago: it simply doesn’t have the performance issues mentioned.)
I think the main issue is the functional approach. Having a functional approach to state can be quite elegant, but ultimately the computer does not do functional. You have to implement a lot of plumbing to have stuff that is a bit performance (clojure), or just add a veneer of functional over what is essentially a normal state machine (emacs).
React works well, because it's only an abstraction over the real DOM. React only handles your app state. But the DOM mechanism is still very performant and very much imperative. But I don't think something like React would work well in a mobile app, because the UI tree is often very simple on iOS and macOS.
Imo the biggest issue with this functional model (at least in React), is that it handles things like virtualization, async, etc. poorly. Which is kinda ironic, because in a true functional language, it'd be feasible to provide 'a world model' - that is act as if the entire state is always available, and let the engine decide when to evaluate pieces of code - without any effort from the part of the programmer.
Not just does the computer not do functional.
UI is also very much not functional, and in fact the lack of progress in UI the last 30-40 years can largely be traced to trying to create UI with procedural/functional programming languages, an instance of linguistic-architectural mismatch.
Further reading:
Programs = Data + Algorithms + Architecture: Consequences for Interactive Software Engineering -- Stéphane Chatty.
https://link.springer.com/chapter/10.1007/978-3-540-92698-6_...
Can Programmers Escape the Gentle Tyranny of call/return?
https://2020.programming-conference.org/details/salon-2020-p...
UIs Are Not Pure Functions of the Model - React.js and Cocoa Side by Side
https://blog.metaobject.com/2018/12/uis-are-not-pure-functio...
Beyond Procedure Calls as Component Glue: Connectors Deserve Metaclass Status
https://2024.splashcon.org/details/splash-2024-Onward-papers...
Hmm, but React doesn’t own or manage the app state. And react native has been used for pretty complex applications
Functional doesn't mean stateless. I'd argue functional programming is superior for representing state in user interfaces. For a success implementation, see Jane Street's Bonsai:
https://github.com/janestreet/bonsai
> functional programming is superior for representing state in user interfaces
It's always going to be slower than using something imperative. Trying to process the entire world's state for the sake of purity can feel elegant but it isn't free. And the complexity that gets added to make things performant is worse than just accepting that UIs are going to require you to jump around the tree and modify state.
Every time I go through the trouble of understanding the latest web technique (React, Elm, Signals (the newest solution), etc.) to deal with state management and the DOM, I end up walking away disappointed. There's nothing new in them that you can use to improve what we've been doing for ages in native GUI toolkits.
And to jump back to the original topic, yes, Cocoa was pretty decent, and SwiftUI while nice in many ways tries to Reactify native macOS development and made it worse. And it made Swift incredibly more complex and worse in hindsight.
The goal of the reactive/declarative approach was never to be more performant than imperative code. The goal is to more easily build UI that is performant enough and functions correctly. With imperative UI code it is incredibly easy to forget an edge case in your update logic.
Not sure how you define "success" here. Is Bonsai used much outside of Jane Street?
> Functional doesn't mean stateless.
I didn't say that.
User interfaces representation are mostly trees. And with functional programming you basically have Tree2 = f(Tree1). Until f is done you can't do anything really. React has a lot of escape hatches to improve performance, but they are escape hatches, not an endorsement of the architecture.
With imperative programming (and OOP), you only have that single `Tree`, which you update at will. Less elegant yes, but we have modularization to help us there. What Emacs does is to keep that `Tree` as a single mutable object, but have the code be functional, while the results are imperative.
I only have to look at the iPhone fucking 17 to know Apple is long gone.
Genuinely curious, what’s wrong with it?
I’m no huge apple fan, but I didn’t really think the 17 was more than a regular boring spec upgrade, which seems fine all things considered. (Currently on an iPhone fucking 17)
This seems to be a popular ‘controversial’ topic. I’ve been using SwiftUI for production applications and games since 2021. I drop down UIKit, Metal, or core animation when needed. But that’s no different than when I was making games in UIKit and would drop down to core animation or glyph renderers written in C etc.
Data Flow: The author claims there’s no way to know when things update. Not only does experience help here but there are profile tools that tell you when and where things are updated. This isn’t black magic. Keep Views small, be careful how you hand around data. @environment is super cool but can have a cascade effect. This was greatly improved iOS 17+ and I wouldn’t support anything older than iOS 17.
GeometryReader: Occasionally I’ll use this. It’s kind of a necessary evil when dealing with certain view complexity. It can also be a sign that you’re doing something wrong.
API Stability and performance: Apple users upgrade. There’s no reason that you should be supporting iOS 17 at this point — even iOS 18 is roughly 2%% of our user base across several apps. I’ve been using SwiftUI without major performance issues but I also don’t early optimize. I profile and fix as needed. One of the early studios I worked for wrote all our games in UIKit as prototypes, when performance tanked we’d switch to the appropriate tools (eg. OpenGL) where it was necessary — like in the core game.
I could go on but in the end just use the right tool for the job, if you’re not proficient in SwiftUI or it isn’t going to work for your cross platform project, you have a lot of other alternatives. For me though, it’s been amazing to work with. I stepped away from iOS programming for 3 - 4 years because I was burnt out using storyboards, dealing with massive view controllers, and all the boiler plate it’d take to get a view up and going in UIKit. SwiftUI roped me back in.
* Quick addition edit: Cross platform for iOS, iPad, macOS has never been good. I’ve found recent updates have made things better to the point of tolerable and it’s nothing like when we had to post-fix an ~ipad to our Nibs — There has never been a ‘glory days’ of cross platform Apple UI libraries.
I'm sure many will disagree, but I have doubts that pure declarative-reactive is the "right" shape for an all-purpose native UI framework. In my experience, Kotlin+Compose shares many of the same warts… its main redeeming quality is that it's better than Android Framework (most of the time), which is low bar to clear.
These frameworks have a number of good ideas but they don't necessarily combine in a way that transcends high quality traditional imperative frameworks with declarative-reactive bits sprinkled throughout, at least for more complex apps. SwiftUI and its ilk work best for super simple tabs-and-flat-lists sorts of apps.
Do you have an example of where "the ideas don't combine well"? Or where an imperative approach is truly better?
In my experience the solution is often to use the appropriate architecture underneath your UI layer. Basically you build a data structure that represents your UI and always hand this entire structure to the declarative UI layer. Underneath the UI layer, you can still use imperative code to manipulate the data structures. The benefit of the reactive/declarative approach is that you don't have to think about how changes in the data need to be reflected in the UI. That's the framework's job.
Note: With "data structure" I don't mean "build a shadow DOM". I mean something specialized to your use case.
I think open source is the right move here. The great thing about Flutter is you get the declarative UI to build a lot of standard screens easily but you can always drop down to lower levels, inspect the standard components, mix and match with your custom implementation. You can mix UKit and SwiftUI but there’s way more friction due to the closed nature of the SDK.
Fully agree (been an early Flutter adopter!). It's always the right too for the right job, declarativness makes a lot of sense for most UIs. UIs are surprisingly more complicated than one might think, there is accessibility/ theming/ keyboard traversal etc., just declaring UI and letting the framework figure out the rest has held up pretty well.
And because of the open source nature you can always drop down in layers, just a few weeks ago I made a very specific engine optimization, all while the application code on-top can stay declarative!
Perhaps. Certainly source visibility makes life easier as a dev who has to live with these frameworks on a daily basis. That said, I think frameworks should strive to have a wide and deep enough set of widgets that in most cases, devs won't need to resort to custom implementations to do what they need to.
It also doesn't fix problems inherent to declarative UI, like readability breaking down easily and certain things that would be a cinch in an imperative framework being extremely awkward to implement.
Apple had a real winner with ObjC and AppKit. Swift is horrendously complicated for the benefits it offers over ObjC, and SwiftUI is a massive step backwards from AppKit.
Just MHO, and it’s not going to stop the juggernaut, but there’s just no appeal in moving there for me :( If (when ?) Apple drop ObjC, that’s the day I move to Linux
Apple will likely never drop Obj-C, but it's been a secondary language for years now. New frameworks are all written in Swift, Obj-C is only used for existing Obj-C codebases.
Autolayout, while flawed like anything else, remains the pinnacle of UI across all platforms.
SwiftUI is a laudable attempt to idiot-proof UI, but it sacrifices too much and ultimately fails.
I've noticed that developers who started with UIKit really have a hard time working with SwiftUI. I think this is because it's not just new syntax, it's an entirely different way to think: state is the source of truth, views are ephemeral, and you describe UI instead of managing it.
Apple has been adding Observable support to UIKit as well, so it's easier than ever for state to be the source of truth in UIKit too. Anyone writing more than simple UIKit apps has known that for a long time, it's just big a huge pain to actually implement without a lot of custom code or a dependency like RxSwift. If Apple had embraced reactive programming 15 years ago, SwiftUI would be UIKit's new layout system, not a whole new API.
Personally, SwiftUI makes the 80% so much easier that, even if the remaining 20% requires dropping down to UIKit, it's worth it.
SwiftUI it is the type of framework that makes the easy things easier to accomplish but the harder things harder.
It is a newbie trap. It is great at producing simple apps, or things that don't require intense scrolling, or anything with heavy animations and precise layouts, but when you do something smooth, it is not it. It feels more of a React Native competitor, than a true UIKit replacement.
Also, almost everything that Apple has re-wrote with SwiftUI, came out worse as before (Settings, Spotlight, etc), and that doesn't help.
With AI coding... SwiftUI lost its edge/advantage (easy to layout screens), as now you code less of that directly, and might as well just go fo the framework that allows the most flexibility and gives you the best results for your users.
Apple really needs to either double down on it, and make it such that it has everything that UIKit has (match both features and performance), or just turn it into another optional framework, just as Interface Builder (.xib files) back in the day of Objective-C.
Right now it heavily promotes it as a first class citizen, meanwhile the features are not even parity with UIKit. It is so misleading to newcomers to the platform.
Ps. The other upsetting thing about SwiftUI, is that it has hurt the Swift language as well, as the team has had to introduce all kinds of hacks, or obscure features to the language in order to make it work, which made even the Swift language experience worse and more complicated than it should have been.
> With AI coding... SwiftUI lost its edge/advantage
On the contrary, I think AI has made it EASIER to approach SwiftUI now, because AI makes up for SwiftUI's biggest weaknesses: the many ways of doing everything, Apple's wonky documentation and the need to know when to fall back to AppKit/UIKit.
Maybe for you. But SwiftUI promise was to make coding UI easier for the programmer similar to ReactN, but with AI coding that is not necessary anymore, you can go straight to UIKit which is most stable and feature complete.
In car terms:
UIKit - Manual Transmission - you control everything, but a bit of a chore
SwiftUI - Slushbox Auto Transmission - easier to drive, sucks for performance
UIKIT+AI Coding - Dual Clutch Transmission - both great performance, and easy to control/drive
> real, production-grade UI framework
What does this even mean? There are hundreds of thousands of apps in the App Store that are 100% SwiftUI. They are real. And they are very much "production grade".
The only people still complaining about SwiftUI 7 years later are the UIKit holdouts that never took the time to properly learn how to use it.
SwiftUI is gold compared to SwiftData.
Agreed. SwiftData is terrible for anything more than like 2 "Models". I have a bunch of apps written with SwiftData and a bunch more with GRDB... I hate touching the SwiftData ones.
A similar story to WPF.
WPF was abandoned at birth, it went years without any attention and then was superseded.
Apple has at least been consistent on UI frameworks.
Sort of superseded. In practice WPF is still very often the best option in the complete mess that is window GUIs
Yeah, I like WPF and I still use it for my own apps. In fact, my own software is the only thing keeping me on Windows now. I really should take a look at Avalonia and port my stuff over to Linux.
WPF was stamped out by lack of dogfooding = inability to inspire confidence, and then Microsoft spazzing out with that clusterfuck that was the Windows 8 "Metro" bullshit.
Back when I wanted it to succeed, Microsoft was barely using it in any of their own apps.
Apple has used SwiftUI in their OS more than MS/Windows did, and there's quite a few apps on the App Store that -seem- to be made with SwiftUI.
I fell in love with SwiftUI the day it was announced, but as a solo dev I still haven't been able to make a full app with it yet, mostly because of the lack of documentation, and the gaps where you still need to drop down to AppKit/UIKit.
I went through Visual Basic, .NET, WPF, Cocoa, and other random frameworks all professing to be the promised panacea for UI, but I think SwiftUI+SwiftData is the best environment ever.. IF only it could reach its full potential, i.e. do everything that Apple's "legacy" APIs can do.
I even tried using SwiftUI for games: https://i.imgur.com/5aTWbft.mp4
The biggest/worst hurdle in the "modern" dev experience is Apple's insistence on a yearly update cycle and the way they advertise those updates: You have to wait for the next WWDC and suffer through videos of uncanny-valley presenters, hoping to catch a glimpse of something that fixes the shit that was bothering you since the last WWDC.
3rd-party sites like hackingwithswift.com & swiftwithmajid.com provide invaluable info that Apple's own docs should.
At least the Swift language has been getting more regular updates since it went open source.
There's no way I'd dare to take on a full Apple-platform app project alone on my own, but I've started dabbling in it again thanks to AI: Codex even converted an old app I made in Visual Basic 900 years ago and had it running in SwiftUI within minutes!
I've even tried to get AI to sift through the WWDC video transcripts so I won't have to waste my mortal lifespan on that.
Yeah I don't really understand their insistence on these videos. I don't have the time to watch an incredibly hard to watch present babywalk me through the code, hoping I'd catch one modifier I missed that fixes the issue. Just write some docs for god's sake.
> There's no way I'd dare to take on a full Apple-platform app project alone on my own
And that's really sad, because earlier Mac OS X (and even OPENSTEP to a degree) was an Eldorado for boutique developers, not least due to frameworks like Cocoa and Objective-C (despite all the hate it seems to get).
If you knew C, it was not hard to grasp Objective-C (at least it was much easier than say C++ and MFC on the Windows side of things).
That's actually why I jumped from Windows to Mac in the first place!
My first Apple purchase was an iPad I got for my aunt, and I loved how streamlined yet powerful it was so I installed a VM on Windows to dabble in iOS development, and fell in love with the sheer consistency of macOS, iOS and their APIs.
This was during the era when Microsoft was still figuring out whether to keep the Start Menu and replace the whole Windows UX with a tablet UI, so it was an easy jump.
Almost never looked back, and whenever I do I'm glad I jumped.
Then came Swift and I loved it, but then it took literal years to catch up with everything you could do in Objective-C with the Apple APIs. SwiftUI is having the same problem but for much longer :(
Is the right way to develop mobile apps flutter? or Kotlin multiplatform plus native UIs?
What I like about Kotlin Multiplatform (KMP) vs Flutter is that KMP is not a framework. For instance, I can write a KMP library and a "pure" Android or iOS app can pretend it is a "pure" Android/iOS library.
That is, pure native apps, can integrate KMP libraries. They cannot integrate Flutter libraries.
In other words, if I write a KMP library, I can ship it for multiple platforms and developers can use it in their Flutter project, or in their pure Android/Kotlin project, or in their pure iOS project.
There is no right way :)
SwiftUI is good, albeit not without issues, as long as you tolerate them. The same is true for any approach. My most recent app was built in SwiftUI because I have been using React Native in my professional work for so long I wanted a change of scenery.
Each approach has its issues, and it all boils down to a matter of preference and which tradeoffs are appropriate for the project.
I’ve worked on some award winning apps in the past, our experience with SwiftUI is it makes it really easy to develop a basic average looking app. As soon as you need something non-standard, you have to drop down to UIKit and figure out how to play nice between the two, which is often non trivial and hard to estimate. You file radars, which at best will be fixed in the next OS release and rarely back ported if at all. In the end some of the more complicated apps just ditched SwiftUI altogether or applied a myriad of workarounds.
I wonder what experts think of this other option: going fully native for one platform then vibe-porting it to others.
It's never "just working", you'll still have to debug the vibe-ported apps. And you won't be good at it if your devs don't actually know the target platform.
All that to say, I don't think that vibe-porting will give you more than cross-platform. With cross-platform, at least your engineers understand the codebase.
I know people would dismiss this, but I'm going to try to make the case for Qt - I'm currently converting my note-taking app to mobile and it's going very well.
I have been involved in two non-trivial Qt mobile apps. Never again.
I don't know what your experience is, and I mean no offence or anything, but the only people I have seen who were happy with Qt for mobile development were people who had never tried to do it with a modern language, like Kotlin.
With Compose Multiplatform and Kotlin Multiplatform, you can write a cross-platform app in a modern language, and that is far better than C++.
It's licensing is expensive though unless the app is opensource.
I heard this repeated many times. You can create closed-source Qt apps without paying any money. You cannot statically link, though; you must obey the LGPL license; that is all. And yes, you cannot use commercial modules, but that is kinda expected.
Yes, Qt is mature and high-quality software. If you’re doing OSS, then Qt is truly a no brainer. The amount of C++ you have to write is truly minimal, and it’s not “raw” C++. There’s a lot of tooling and libraries. You can target web now too, and QML is legitimately good.
If KDE can develop a high performance stable desktop, you can design whatever.
Flutter is not the "right way", due to the nature of Flutter being a lousy reimplementation of a UI frameworks.
I developed a complicated app in Flutter and it’s one of the best engineering decisions I made. Flutter is great.
Flutter has been a great success for us.
Same here, but Liquid Glass has been a headache lately. Google really dropped the ball here, but to their credit it almost feels like Liquid Glass was an intentional move by Apple to cause as much friction for cross platform frameworks as possible.
Define right, it’s hard to replicate the native UI especially Liquid Glass on iOS because it’s rendering everything using its own rendering engine, making it hard to feel right on iOS. I have also seen a lot of companies move away from flutter to either native or React Native.
Well we all love Liquid Glass. Most loved UX update in iPhone and macOS history. Massive dealbreaker for an app not to look like Liquid Glass. I definitely choose apps based on whether they look like proper Liquid Glass applications instead of the developer taking shortcuts and focusing on working features instead.
You jest but yes, even as I do not love this new update, an application bucking Apple platform standards personally sours me pretty hard on it.
Well I’m a heavy Flutter user myself, and the app I built Hacki often gets reviews saying it doesn’t feel native on iOS. I personally don’t have anything against it and I think Flutter has the BEST developer experience compared to other mobile app dev toolkit/SDKs.
I despise Liquid Glass too but for most users there’s a broad feeling of “it fits” vs doesn’t. And apps with Liquid Glass fit today’s iOS.
SwiftUI was doomed from the beginning. Not only Apple completely blew the implementation, it was actually DoA by a bunch of super bad decisions that ultimately make it incredibly hard to work with and manage, especially on bigger apps.
1. It uses the builder pattern for views (familiar to those who used Java) but for some reason they decided to make it so that the order of the modifiers in the builder matters. Each modifier doesn't actually modify the main view but it modifies what the modifier before it decided to return. This makes no sense as the modifiers should each be modifying the main view to make everything predictable and easy to debug. I'm convinced no one (not even senior devs with 5 years of SwiftUI experience) understands how the ordering of modifiers works. It's just swap them until it does what you want it to.
2. The view lifecycles and code execution path seem random and hidden behind layers of "magic," making it incredibly difficult for developers to trace and debug issues.
3. It is practically impossible to set breakpoints for rendering and view construction, it's impossible to really figure out when re-renders happen and what drives them. I'm convinced SwiftUI apps are incredibly slow not because the SwiftUI implementation itself is slow, it's because, even Apple's own apps probably do a bunch of unnecessary re-renders and one no one seems to have any idea. This is unfortunately another design issue that can't be solved by just making SwiftUI more efficient. It requires simplification and tooling to help developers not footgun themselves.
4. Lots of issues start appearing later on in the development cycle because, for simple apps, bad SwiftUI design decisions and footguns have unnoticeable effects, until the apps gets more complex and things start breaking. Fixing these issues sometimes requires rewriting whole features or spending hours debugging.
5. There seems to be almost no documentation on Liquid Glass. It's laughable that after more than 1 year, Apple has simply refused to document or provide good examples for Liquid Glass, except for maybe couple pages that resemble the brain dump of an engineer that has never passed a writing class in college?
6. Stuff seems to be rapidly changing and breaking from version to version. It took days to make my app look and work the same in iOS 27 as it did on iOS 26, even though iOS 27 is supposed to be a minor bug fix release. We don't even use anything non-standard and don't do any hacks. This defeats the whole purpose of a simple UI framework that can be easily adopted to different platforms (this never used to happen with UIKit).
7. View debugger still has no SwiftUI equivalent. It used to make things so much simpler in UIKit when you could just see the view bounds, pick views apart and understand what's actually happening. SwiftUI has no equivalent other than `.background(.red)`. Terrible.
I don't know how Apple can salvage this beyond just undoing some of these terrible design decisions and making it 1. super simple to work with, 2. stop relying on magic, making things more explicit, and 3. providing actual 1:1 UIKit feature parity.
Apple seems to be loved by users, but hated by developers (and for good reasons). They just seem to not give a shit about the developer experience.
You mention some examples, I have experience with lower-level stuff: terrible. Poorly documented, and actually not really working well.
IMO they should focus on improving the developer experience instead of wasting money on the joke that is Liquid Glass (even users don't like it, right?). But anyway, too late for me: I'm soooo happy that I can develop iOS apps in Kotlin now, and just endure Xcode for building and occasional debugging.
1. It's not really the builder pattern, which, in Java, is what you describe as wanting: mutating a single instance. SwiftUI uses result builders to compute a single generic View from the "DSL", but you could do the same thing without the custom syntax, it would just be more verbose.
2. Documentation issue. The lifecycle is standardized, but Apple is terrible as documenting it all in one place. I found this year's "Dive into lazy stacks and scrolling with SwiftUI" WWDC video actually had a pretty good explanation, at least for the lifecycle within lazy containers.
3 and 7. Yes, Apple's tools suck pretty bad. Preview canvas is also still terrible, despite multiple attempts to improve.
4. Not something I've experienced, not sure what you mean.
5. Yeah, Apple's docs suck.
6. Not sure what you mean here either, our stuff looks pretty much the same.
skills issue
https://en.wikipedia.org/wiki/Poe's_law
Let me rephrase, he is not the sharpest knife in the drawer.
In my professional medical opinion, he is at high risk of being replaced by a year old low-end Chinese open source LLM.
Somehow you being the GNOME co-creator made me curious about what the GTK/GIMP creators (Kimball, Mattis) are up to. CockroachDB, it seems. It's weird how it's so easy to miss these connections.
yeah, I often check on them - love watching what they are up to.
I joked, but reading the post is even more clear - man this poor dude.
He is brute forcing his way through life and broadcasting it on YouTube.
Dunno man, Xogot is made up of almost half a million lines of code. Of those 727 files use SwiftUI, and 65 use UIKit, 119 use AppKit (I did not count the overlap, so they are not mutually exclusive).
Using GeometryReader makes it clear he is living in the past and the fact that he uses printChanges in this year without mentioning the dedicated SwiftUI instrument show a lack of curiosity on the space.