I dived (dove?) into Catalina on day 1, the very first developer beta, because of SwiftUI.
It's an amazing framework. Sure Flutter and others came first and already did parts of what SwiftUI+Combine do, but being able to take an app from design mockups -> interactive prototype -> final version right in the same editor and using the same language for UI, logic and model (with native APIs and look-and-feel) is something I've fantasized about since forever.
There are many rough edges (just like Swift itself which took like 3 years to become good enough for "serious" use), you still need AppKit/UIKit for some things, and Apple's developer documentation remains appalling as ever, but I was able to turn a bunch of ideas into usable apps in a matter of hours, that traditional frameworks would have taken me weeks to do.
I hadn't even planned or expected to make "proper" apps with it, I was just playing around, iteratively adding to mockups, and boom, there was an actual app on device.
Catalina itself however was the buggiest macOS beta for me. I permanently lost random files on iCloud Drive, CPU and network usage was always high, and I had to use a temporary local user account that was not tied to iCloud to avoid messing things up.
And it's annoying that they still haven't fixed some minor bugs I reported since the Mojave beta, like unreadable white text on brightly colored tags in Dark Mode and Books jumping around between desktops on its own (the kind of stuff that would have infuriated Steve Jobs I assume but is deemed unimportant now.)
My SwiftUI experience is quite different. For me it worked ok for everything they do in the tutorials but everything else is underwhelming. Animations are really cool though.
Some of my issues where that sometimes the buttons seem to have a very small hit-box (yea, I changed that but it didn't help), you have to use lists because tables mess up certain animations, lists always have this separator line, the documentation is not really there, ....
I started to use it from the 3rd beta and still, the API changed with every update and generally it felt more like an alpha. Overall I wouldn't do much more than a simple example ToDo app with it and definitely not rely on it for a product.
Yes, there are bugs and documentation has always been Apple's Achilles' Heel.
I definitely managed to make something a little more complex than a To Do app though. See the demos from other people like @MengTo [0] to see what's already possible.
Lists etc. will be customizable once they open up the -Style protocols. Try ScrollView -> VStack -> ForEach for now.
Yeah...and SwiftUI is by default an iOS only program. Flutter you can plug any phone in that's been made in the last 5 years, type `flutter run` in the terminal, and get hot reloading.
SwiftUI is walled-garden bait. Hard pass.
EDIT: Not sure why I'm being downvoted - (one of) Flutter's main selling point is that it's cross platform by default. SwiftUI doesn't do that and yet is marketed to be very similar in coding style to Flutter, which yeah great, but no cross platform so you still have to code everything twice, unless all your users happen to have iPhones. I mean really?
and macOS, tvOS, watchOS (and soon I suppose, AROS?), all counting for over a billion devices.
You can literally run the same SwiftUI code on all of them, then make small changes to adopt each OS's unique paradigms (like menus on macOS, the crown on watchOS, the remote on tvOS) and get the most native performance via Metal etc. for free.
It's ideal for publishing to services like the new Apple Arcade and macOS/iPadOS Catalyst.
How is Flutter for accessibility features? Those seem really effortless to adopt in SwiftUI.
> ideal for publishing to services like the new Apple Arcade and macOS/iPadOS Catalyst
I was just looking into this yesterday but found no info - so you're saying it's now possible to write an iPadOS app with SwiftUI and use Catalyst to port it to macOS? It used to be a "known issue" in the beta that this was not supported... does this work now?
I haven’t tried it but it must be one of their top priorities to get it working soon, as Mac/iPad hybrid apps are one of the use cases where SwiftUI makes the most sense.
You can of course write regular (non-Catalyst) projects that share a lot of the same code between Mac and iPad. In that case you may not even need Catalyst, so maybe Catalyst is only UIKit-on-Mac.
I guessss.....except if I want to make an app (and not a web page) I would probably make it in electron, because again I don't have to code everything twice. Making apps for apple watch or an iPad just doesn't appeal to me I guess - they always seemed like devices that were invented so people could spend more money on needless crap. Is a watch with a touch screen really that much more useful than a smart phone? Seems like just another way to suck people into a walled garden. Your watch sinks to your iPad, iTV, iWhatever...Yeesh, who cares?
> they always seemed like devices that were invented so people could spend more money on needless crap
As a view on the other side; I develop business apps. They are distributed via a local enterprise store, and never appear in the public App Store. Often, my apps replace paper in old-fashioned businesses. It's very honest and satisfying work.
No, RN isn't bundled with any browser runtime. You're possibly thinking of Apache Cordova (which leverages the device's browser runtime for rendering UI).
RN is Javascript so it will execute on the device's JS runtime, which I guess is connected to the browser in some way, but RN doesn't use the browser for UI. The rendering is true native UI.
The way React does this is by decoupling the core library from UI rendering. The core React library contains no UI rendering code, it just handles UI lifecycles, internal state, events, etc. for logical UI components as object instances. You then plug that into a renderer library (most commonly the "ReactDOM" lib for web browser rendering).
If I do this my apps are limited to the lowest common denominator which means supporting either 6% best case of Android users, or going back further, supporting say 30%, if Flutter will even support doing so, and giving up many features thus hurting not only my Android users, but also my iOS users who could be way ahead and who stay updated. And tying the future of my project to Google? No thank you. Have you seen any Google projects lately? Shallow and not impressive at all, and often abandoned. OMG I have to code everything twice you say that like coding is not fun, or like it's the end of the world. How about this: code everything once, and target iOS. If cross platform is really needed, go with something from a serious company, like Xamarin.
It's an amazing framework. Sure Flutter and others came first and already did parts of what SwiftUI+Combine do, but being able to take an app from design mockups -> interactive prototype -> final version right in the same editor and using the same language for UI, logic and model (with native APIs and look-and-feel) is something I've fantasized about since forever.
There are many rough edges (just like Swift itself which took like 3 years to become good enough for "serious" use), you still need AppKit/UIKit for some things, and Apple's developer documentation remains appalling as ever, but I was able to turn a bunch of ideas into usable apps in a matter of hours, that traditional frameworks would have taken me weeks to do.
I hadn't even planned or expected to make "proper" apps with it, I was just playing around, iteratively adding to mockups, and boom, there was an actual app on device.
Catalina itself however was the buggiest macOS beta for me. I permanently lost random files on iCloud Drive, CPU and network usage was always high, and I had to use a temporary local user account that was not tied to iCloud to avoid messing things up.
And it's annoying that they still haven't fixed some minor bugs I reported since the Mojave beta, like unreadable white text on brightly colored tags in Dark Mode and Books jumping around between desktops on its own (the kind of stuff that would have infuriated Steve Jobs I assume but is deemed unimportant now.)