> This fundamentally does not work for anyone with more than 10M+ installs just like you can't write Mandarin and English in one script.
Well, I did work on a Flutter app with a tiny team between 2018-2021 and we had 15M installs, were app of the day, got featured multiple times in the App Store, got Google’s design award, were a preloaded app on iPhones in Apple Stores in several countries, and were overall doing quite well.
It was a very focused app though, so the size of our codebase didn’t really grow with the number of installs.
New iOS and Android features, sometimes backward-incompatible are introduced.
And now, you need your dependencies to implement them. Which might or might not happen. At best, this makes usage inferior. At worst, you can no longer publish an update in app stores.
This app was started in 2017 and it is still running today and making (way more modest than back then) revenue. It’s just that I don’t work on it anymore.
I only have one meeting a month nowadays so I might take a look if that situation ever changes (hopefully it doesn't!).
Regardless, congrats on the launch.
> the macOS app and Chrome extension communicate via WebSockets on localhost; nothing leaves your machine.
So the macOS app sets up a local webserver on some port that the Chrome extensions knows about in advance? Smart, kinda like how I had to do an auth flow for one of my CLI apps.
> I only have one meeting a month nowadays so I might take a look if that situation ever changes (hopefully it doesn't!).
I'm happy for you
> So the macOS app sets up a local webserver on some port that the Chrome extensions knows about in advance? Smart, kinda like how I had to do an auth flow for one of my CLI apps.
Exactly like this! Some caveats there since Chrome actually shuts down extensions after 30 seconds or so if nothing happens (so you need to keep firing off messages across the connection to keep the extension alive).
I actually don't have any experience in that, but I think that you can absolutely do it. From what I understand, the IDEs just call the Flutter CLI and Dart analysis server, among other tools to get stuff done. From the top of my head, "flutter run", "flutter test", "flutter analyze", "flutter packages get" would be the commands you'd be using the most and could survive without using any IDEs. Of course you'd lose the autocompletion along with other IDE niceties, but I really believe that could be done. One thing is you might have to open XCode once to do the provisioning profile setup, but after that you should be good.
I'm hoping someone with more experience could chime in.
Thanks. Is having XCode necessary in order to build and publish for iOS (I guess yes)? I don't use apple computers and use linux, so if that's the case I'd need to buy a Mac and XCode to do so, no?
For iOS apps, you need a Mac. Maybe (and this is theoretical) you could set up a combined Fastlane & CI integration that builds the apps using some cloud service, such as Macincloud, Bitrise, etc. Fastlane would then build and upload the iOS test builds automatically so can use them on your phone using the internal testing group in TestFlight. When everything seems fine, then you can just promote the internal test build to production. Since Flutter custom renders everything, you could do most of the development on a Linux machine and test on Android emulator/device.
WRT cloud building: yes, it'd be more ideal to have a Mac especially if you want any platform specific UX or features but you can definitely get around it using just a Linux machine, an iOS device and some CI system.
You can also get around it using just a Linux and an Android device. You can press 'o' after you flutter run to change the platform on the fly to get iOS scroll physics, page transitions etc etc. But again, testing on the devices your users will be using is more ideal.
Thanks, glad you find it useful! And that's exactly the reason I open sourced it - I thought there was a need for a quite comprehensive but still simple sample app for people to learn :)
For the difference in coding time, it's hard to estimate. I'd say once you get the hang of it, which for me was quite fast, it's much more pleasant and generally faster than working with native Android. Hard UI related things on Android (rounded corners or customizing the design library components) are pleasant to do and make a lot of sense in Flutter.
Flutter and Dart come with a nice HTTP library out of the box. In my experience, it does everything you'd expect from a networking library. If you're looking for something like Retrofit on Android, there's a package called "jaguar_http" that does the job.
For taking/picking images, you'd use the officially supported "image_picker" plugin. It exposes a nice API based on Dart's Futures and handles Android's onActivityResult dance for you in the background. If you're looking into embedding a camera preview in a Flutter app, there's a package called "camera" for that.
There's generally a plugin that handles the platform stuff you need to do. If not, you can always create one and use Java/Kotlin/Swift/ObjC code to call the native SDK APIs you need yourself. Having published a couple plugins myself, I've found the plugin system to be really nice.
Like I said before in another comment, there's only so much I can fit in a Hacker News title. :) I'd like to think that a lot of people use the term "a cross-platform app" to mean "an app that works on Android and iOS" though. Or "an app that works on more than one platform using shared code".
I've been using Flutter for a year now, and I've loved it since the first time I gave it a try.
In my work, I've been using other cross-platform technologies, such as Xamarin and React Native briefly (actually currently doing a customer project with RN). In my experience, RN is not that bad and mostly gets the job done, but I've grown to like Flutter more. The tooling and documentation are top-notch ("it just works"), and since it custom renders everything, the UIs are consistent across devices and platforms. For example, with React Native I had to recently fight with a third-party library to have gradients (RN doesn't support them), but couldn't get the library working.
Another thing that I heard from colleagues was that React Native can introduce a lot of breaking changes between versions. I have a one-year-old hobby Flutter project and I was still able to run that without any pain when the first beta came out.
Being a native Android & Java guy for the last 5 years or so, having the strong type system of Dart also felt more natural to me than Javascript.
I've been starting work with Flutter over the past couple weeks. The tooling and general experience so far has been pretty nice, though it doesn't always "just work". The fact that flutter is so coupled to Cocoapods is a disappointment and it means your build system is dependent on Cocoapods not screwing up - quite a tall order. I already ran into an issue that required me to generate a new project, copy my source over to get a working app again, then inspect the source trees to see what went wrong. It turned out to be a Cocoapod symlink issue or something to that effect. Not entirely a Flutter issue but something to be wary of.
Some good points of Flutter so far:
* Hot reloading is fast, much easier to iterate on designs and experiments
* Plugins seem to be easy to make, they made an architecture that focuses on minimal boilerplate
* Being able to use async-await syntax on a whole UI view is a cool idea (like Android's onActivityResult() with much less typing involved). Push a view onto the navigation stack, let it collect input from the user, pop off the stack with a data structure representing the user input)
* Animations are easy
* Minimal opening of Xcode. Most days I can get by without ever opening it.
* Data flow architectures like the Elm Architecture or Redux are well supported
Some not-so-good points:
* Having null
* Material design is pushed way too hard - Cupertino widget documentation/examples are lacking
* Code generation needed for JSON parsing
* State for StatefulWidgets are generic over the Widget and not the other way around - I would have expected a StatefulWidget to be generic over a State e.g. `MyHomePage extends StatefulWidget<MyHomePageState>`
* Because of the previous point, the UI building function, build(), is usually on the State class and not the Widget class
* My relatively barebones app is around 33MB on iOS
Overall though I like it! Lots of good ideas and tooling. I one day want to create something with a similar architecture using a language like Rust instead of Dart, and native UI components instead of reimplementing the entirety of Android and iOS component behavior.
I don't really agree about pushing hard on Material Design. Sure, the current widget situation might be skewed towards Material components and Cupertino is lacking. For Material widgets, they've implemented almost everything, even the most obscure ones. But I think they're pushing more for customized UIs. Meaning, UIs that don't look Material or Cupertino, but are more of a branded experience, think Spotify, Tinder, Snapchat, etc. Designs that don't really conform to Material or iOS design languages. The Hamilton app is a good example of this. Sadly, I'm not a UI designer so I went with Material Design on this inKino app since that's an easy route for me.
With your other points, I generally agree. Didn't mean to imply that it "just works" 100% of the time. But compared to other cross-platform dev tools / SDKs I've used, I feel like Flutter does such a great job in the tooling department.
I must have been spoiled by Flow, Typescript, Kotlin, and Rust, because this (having null implicitly be a member of every type rather than a separate type) has been the major factor in my decision to avoid playing with Dart yet. It's so nice to not have to often manually re-check functions I'm editing to see whether or not each of the parameters are supposed to handle null, and to not have to document that in comments (the same way you usually end up documenting types in a dynamically typed language; this is what I'm using a typed language to get away from!).
I've been using React Native too and my experience is: I accomplish a lof of stuff in a few days, then lose a whole day trying to hack a custom Picker element or similar. Grrr.
Disclaimer: It's been over a year and Xamarin could've improved a lot during that time.
Coming from a Java/Kotlin background, I found C# to be really nice, but I had some issues with Xamarin tooling. I used Xamarin Studio and for the most part, it worked fine. Sometimes, there were some random crashes and project build errors that mysteriously resolved themselves. I also needed pie charts per the customer requirements and the best pie charting library I could find had licensing issues, so ended up implementing the pie chart myself which wasn't that bad.
To compare, I would say the Flutter tooling is really good, compared to Xamarin where sometimes I would not be confident with the tools. Also, with Flutter, I can share the UI layer entirely, compared to the Xamarin project where we only shared some plain business logic with written purely with C# without Android/iOS SDK dependencies. I know about Xamarin.Forms, but have no experience to compare it with Flutter. At the time the project team didn't want to use Forms, but I don't remember why.
There's only so much I could fit in the title, and I honestly didn't mean to mislead anyone. I can definitely see how it could be a little misleading now that you mentioned it.
Yeah I definitely thought this was sth. about making videos on mobile, but I don't think the title is purposefully trying to mislead anybody. Maybe titles should be allowed to have ~100-120 chars instead of 80, it's a bit too limiting.
