I think you might not be considering personal bugfixes (X interaction is great for the common denominator, annoying to me), long-term support (I want to use my device securely in 2045), or custom features (I want to be able to do X, but only 0.001% of people care so Apple doesn't). I use Android, so I can't be more specific. I have heard the unverified complaint that Apple sometimes slows down old devices as new ones are released, which would be impossible in FOSS.

Apple is ultimately going to work in their own best interest. Their particular brand aligns that with the consumer fairly well, but there will be disconnects.

I don't really understand this. You plan to single handedly replace the entire OS and security teams to maintain the OS on your phone for a decade? How will you update the kernel when the proprietary driver blobs stop being maintained?

I expect more people than just me are interested in reducing tech turnover or using old tech in the hobbyist space, ie. not single handedly. And yeah, as the sibling says, if we're relying on proprietary drivers, we're not really FOSS- my fantasy world doesn't just apply to Apple.

Yes, it would be very hard. But today it is impossible, so very hard is hardly a complaint you can make in comparison.

Did my other two examples seem more reasonable?

I imagine their ideal open source phone OS wouldn't include any proprietary driver blobs, but what do I know...

Perhaps the more generous take is that, with FLOSS software, you don't have to go it alone. My Pinephone Pro will not be running security fixes I write myself. But entirely because of its open nature, it doesn't need to - it will continue to recieve updates for years and years. For once, the hardware of a cell phone is likely to die before the software.

If it gets the support of the right people who know how to write such security software and have an interest in maintaining it, sure. But that's the reason many open source initiatives fall short of proprietary. It's either based on the altruistic whims of a few particular talents, or some company is paying to have them maintain it. And we know the latter is fleeting.

The caveat is that sometimes an initiative can be funded by charity or bounty to keep interest, but relying on generosity for 99.99% of projects is a fool's errand.

No. They want iOS to be open source, because they want to be able to modify a few bits of source code and use their modified OS for their personal use, and then keep pulling updates from upstream forever after that. Like a small fork.

People are still rooting Android and adding thousands of modifications to it.

Android is not proprietary.

https://source.android.com/

That makes it rather easier...

You linked to AOSP, not Android. AOSP doesn't have things like the Google Play Store; the shipped image is substantially different and isn't necessarily open-source just because it contains open-source components - Windows contains BSD networking code but that doesn't make Windows open-source.

Sure, but the vast majority of those mods and alternative distributions are based on the sources that are available.

It is a bit like a Linux distro running Steam. Yeah, there are proprietary bits, but the FOSS part naturally happens on the FOSS parts.

You know what the "A" in AOSP stands for, right? :-)

All I'm saying is that there's a big difference in jailbreaking an OS when most of the source code of that OS is publicly available for study and experiment. Surely that is not a controversial statement?

You don't root OSes, you root devices. I learned that the hard way when I got a Samsung Z-Fold 4 and learned that the US version isn't rootable.

All this terminology is muddied and often misused. The whole term is supposed to refer to getting access to the root user, which is definitely an OS-level thing. One thing people often don't remember also is that gaining root on your phone and unlocking the bootloader are not only different things, but entirely separate things. You can gain root in your vendor ROM without being able to unlock the bootloader (common case on Amazon tablets), and you can install LineageOS in your phone after unlocking the bootloader but you still won't have root access in the OS unless you do an extra step, setting up Magisk or similar. You also have to re-set-up Magisk after every LineageOS update, which often means once a week if you do all the OTA updates in a timely manner. Gaining root is usually the easier thing and the less useful thing. It doesn't help much with getting a nearer-to-AOSP experience. You can remove some apps but there are still limits to what can be done without just flashing a different ROM with less crap in it.

When picking a phone I like to see if it's on the LineageOS devices page (you bring up a good point that different variations of the same model aren't equal, definitely watch out for that), as that means both that the bootloader is unlockable and that someone else is already maintaining a ROM for it, and hopefully will for years to come. If I just go after the shiny new hardware, chances are the bootloader is locked and it will never be unlocked, generating e-waste.

There are several OS level things I wish I could change/override on iPhone: ability to share screen during a video call, ability to control how certain Bluetooth devices pair automatically, ability to throttle apps that I know are abusing the CPU and costing me battery.

You can share screen, Shortcuts supports Bluetooth automation triggers, turn off Background Apps or let the CPU race to sleep.

It's "commonly said" because it's true.

https://www.gnu.org/philosophy/rms-nyu-2001-transcript.txt

via

https://www.fsf.org/blogs/community/201cthe-printer-story201...

> These days modern software is so extremely sophisticated and refined that I can’t think of anything I could improve for say iOS.

> Still love foss, but the idea of “it’s open source, you can improve it yourself any time” has worn off for me.

If you look at major MacOS or iOS releases, you will often see new features being advertised that were copied from FOSS systems that experimented with them and proved them out often many years beforehand.

That is to say: iOS might have the edge on QA, but if you want to have impact and shape the future of how people use computers, contributing in the community is still a good way to do this. And in fact, commercial development tends to rely on this talent pool for survival, as companies do a comparatively poor job in educating new talent. It's where you get people with "job experience" in doing certain things.

When it comes to large sophisticated software the goal is not so much to improve it one self. Rather it is the knowledge that any anti-patterns will be removed by the community if it ever get added. Most of the time the community will not even need to remove anti-pattern from open source projects, since companies know that adding anti-pattern is just wasted developer time.