This has not been my experience (Source: Apple native developer, since 1986).

It is a pain to deal with their App Store, but it isn't intolerable. I don't charge for my apps, but I understand that the Apple ecosystem is very lucrative for devs that do.

> causing codebases to become complicated and littered with if/then evaluations for various SDKs

I have found that I almost never have to do this, and when I do, it's usually a bad code smell. Part of the reason, is that I develop native, for one platform at a time. It might be another story, entirely, if I have to develop for multiple ones.

I just got contacted by a friend, wrt an ObjC app that has been deprecated for over a decade. He reports that it still works fine, on iOS18.

Many years ago, I spent a few months, learning Android. The thing that really put me off, was the Android Pizza of Death. That's the pie chart that describes the percentage of users, running various Android versions.

It was sobering.

Apple basically allows me to write support for the current OS, and support just one back. The app I'm in the process of releasing now, works in iOS 15, but I am declaring support starting at 16, and it works fine in 18.

The actual work of mobile development is much easier now than it used to be. The tooling and APIs have evolved and matured and the hardware is so much faster now that you have to spend a lot less time worrying about performance.

The issue with mobile development is that web apps and PWAs can cover a lot of the same ground now and getting people to install new native apps is really hard.

Sometimes I am wonder why an app is an app, and not a page that can be self hosted.

I know. Sometimes it makes sense. Sometimes it is easier. Sometimes it is for work, and it is a requirement. Often it could be a page, and is not just to have access to your device data.

For myself? Most of the programs I write are pages. Works on any device, Writing guis is not a problem, there are a ton of useful frameworks. There is no need for synchronizing data between devices, not need to install it on client devices etc. Etc.

Interesting there aren't any Android developers here. I develop for Flutter so I do both and Google is a pain to deal with. Once they kicked my app out of the app store because the month currency overflowed to the next line which was 'misleading'. Plus anything Gradle related is a pain to deal with. It's not quite as bad as what the author suggests though. Apple is OK, once your app gets accepted which may take a few tries.

Related on the Android tip:

Ask HN: Why is Android Development so difficult/complex? (compared to Web and Desktop)

https://news.ycombinator.com/item?id=40742719

Ask HN: Am I crazy or is Android development awful?

https://news.ycombinator.com/item?id=41062292

How much does developer happiness really play a role?

Probably significant in terms of speed of development, time spent maintenance projects, etc

I was so frustrated by this that I quit Apple altogether. Bought a Windows laptop and an android phone.

WSL was good enough to keep me on Windows, but I am back on iPhone one generation later (android was not my jam).

If I had unlimited time, I’d be tinkering on PWAs (super glad to see rails pushing in that direction).

I currently only do web development. I figured if and when I needed to develop a mobile app, it would probably be easiest to use something like React Native. Most of my skills would transfer. And I wouldn't have to deal with majority of the issues the author of the article pointed out. This is likely not an option for games. But for non-games I thought it wouldn't be a massive challenge.

I have never developed a mobile app, so I may be oversimplifying it. Let me know if you have been down this path and agree/disagree.

I have a pretty simple game on the iOS app store. It is simple enough that I have probably written it about three times. One of the times I tried to do everything the "apple happy path", but the final iteration in the app store completely rejected it (and was written in C, primarily on Linux then ported to iOS as a final step). I know that familiarity plays a huge part, but the experience of just writing it in plain C (and using SDL2 as the platform abstraction) was so much simpler and nicer. The things I had to implement myself instead of "coming for free" (simple physics, app lifecycle stuff etc) were relatively easy to do.

I haven't ported it to android, but I expect it to be simple too.

What I don't understand from the article is the moving goalposts. I had one threat of store removal because of a lack of sales and updates, so I found some stuff I didn't need to ship and removed it and pushed that as an update. Not so. I think it has been in the store 4 years or so now, so this doesn't seem too much work. I guess it could be churn in the areas that I deliberately avoid (e.g. I have no networking and only use the screen and audio out)

The de-facto apps should have been PWA, free, open, accessible, cross platform.