The second proposed legislation from the article seems to require that at least 5 years of updates must be provided after the device was last sold. This is far more than many low-end and mid-range Android phones, moreover many Android phone manufacturers start counting after the device was released, not last sold.

Interestingly, they also seem to put upper bounds on how quickly these updates should be rolled out. Eg. functionality updates must be rolled out within 6 months after another phone from the same brand has that update. So eg. if Samsung makes Android 12 available on their flagship phone, they also need to make it available within 6 months on the other phones that were sold to up to 5 years ago.

https://ec.europa.eu/info/law/better-regulation/have-your-sa...

This is such an improvement over the status quo.

Is it? It is going to meant that Android can't take advantage of the high end features of the new phones that are being released, and that old and slow phones are going to get updates that makes them run even slower.

To top it of, it is not like manufacturers are into throwing away money, so this will likely mean that prices on the cheapest phones will increase more.

> It is going to meant that Android can't take advantage of the high end features of the new phones that are being released

That's perfectly ok, what you don't have you can't use.

> and that old and slow phones are going to get updates that makes them run even slower.

It would be up to the manufacturers to ensure that this is not the case. One way to do it would be to reduce bloat.

> To top it of, it is not like manufacturers are into throwing away money, so this will likely mean that prices on the cheapest phones will increase more.

This makes no sense: manufacturers are competing with each other, if the prices on the cheapest phones will increase more then surely someone will exploit that.

Maybe it'll play out differently on the low-end, but on the high-end it seems there's been some "soft collusion" where one manufacturer raised prices, and others saw that as a signal they can also raise their prices.

Yes, that is a problem, but that's a different kind of regulation (anti-trust). The EU also tends to be pretty good at busting such things, but they take their sweet time to do it.

In the low end, surely you'd get more standardisation. Same processor, screen, Bluetooth and mobile chips.

So that from an update pov it's easy. They can always differentiate with cameras, batteries, cases etc.

So a communist phone for the lower class?

That’s the not-nice way of saying the same thing.

I’m not sure if I am for or against that without more thought.

I'm still on a dumb phone and that's because I think spending $800 on a phone is ridiculous. And it's not like I can't afford it.

An iPhone SE is half that. A budget Android half again. Both completely usable, just won't replace your mirrorless camera just yet.

There are other reasons.

I spend money where I spend time: phone, chair, monitor, mattress.

I rarely drive so my car is 19 years old.

I use my phone daily (right now!) so I have the newest one.

Why explicitly the newest one?

I’m on the apple upgrade program. So I just pay monthly. I could skip a year and own the phone, but don’t really care. I want the best cameras. Life is too short.

Exactly. Normal people have already replaced their computer with phones so how old is the PC that HN commentators use? Five years?

Phones are appliances that are being used up to 10 hours every single day. And sure you can keep repairing them but at some point buying a new one will actually be cheaper.

I'd love some honest mattress advice.

I'm on a pixel 6A. It cost me under £200 factoring in the trade in deal that let me trade in a 10 year old phone (if you don't have one eBay one for £5).

And there are plenty of reasonable smart phones for less than £100 these days. You can get something like a Motorola E13 for under £70 that will do everything most people use their phones for well (beyond the camera being nothing special, but we're comparing to a dumb phone).

Standadized parts are an idea from the industrial revolution, not communism.

White-labeling an oem part is not exactly the kind of communism Marx was known for

I think it's more that higher price = more flagship. A higher price is in itself a sales argument to a certain audience. If you just want a competent phone there's no reason to buy a flagship one.

Yes - at the very high end, smartphones are a luxury goods market. There's an old adage in luxury goods that "if your product isn't selling fast enough, raise the price".

IDK, I always save up and buy flagships, and convinced my wife to do the same, because that's the only reliable way to guarantee neither of us will have to suffer through a phone that chokes under its own stock software, and is a constant pain to use. So yes, I'm one of those for whom "higher price = more flagship" is an argument, mostly because so far, "most flagship = bestest specs" and "= most care taken by the vendor".

The "collusion" was 5G. Everyone raised price, thanks to Qualcomm?

>This makes no sense: manufacturers are competing with each other, if the prices on the cheapest phones will increase more then surely someone will exploit that.

You assume it costs nothing to maintain software for old phones.

It wouldn't if they opened up the hardware.

So release a handful of budget models and support them well instead of 60+ per year.

If the hardware was more open it wouldn’t even be a problem. People could pick and choose ROMs best suited for their phones and use cases, just like Linux. Newer phones could use bleeding edge features for their HW, and older phones could use more stripped down versions to preserve battery life.

That is the first good suggestion I have seen in this thread. We just need to make it illegal to not allow you app to run on an open device.

I mean realistically there's never been a good reason that phone OS's are as specific as they are. The driver situation is absurd. The most common highly integrated feature - the camera - doesn't even really run on the phone OS anymore since there's usually an image processor of some sort inline.

I have absolutely no problems with running a USB-C on the phone and component manufacturers to get their act together and figure out an open standard to support common components (i.e. basebands, cameras, audio routing, security chips) so hardware can be supported after release (which I would hope in practice would just be "in the Linux kernel tree").

What percentage of consumers want to pick the ROM their phone runs?

This is great because the more expensive phones get, the less likely people will want to replace their 2-year old phone with a new one.

Plus, those new high end features...there have been barely any new features in a smartphone in the last 5+ years (I'm exaggerating a bit of course). Check out Fairphone. That's an interesting project!

Isn’t this the ultimate “let them eat cake” argument? There are many people in the world who can’t afford anything but the cheapest phone.

Those people don't live in the EU. And anyhow the higher cost of the phone are balanced out by the increased longevity and support of the device. Lifetime cost may very well be lower when people end up replacing their phones less frequently and just have to swap a battery.

Has the EU also made poverty illegal?

Sort of? Almost all EU countries have social welfare nets and public services strong enough to have people's basic needs covered. I think in most countries welfare even directly covers basic phone plans.

We aren't in a situation where we need to allow environmentally damaging products here for the sake of the poor because they're not out on their own. This sounds like the old pro-sweatshop arguments people sometimes made about developing countries.

https://www.slowboring.com/p/they-have-homelessness-in-europ....

Poverty is a relative thing, not absolute. As for your link downthread - homelessness has many more causes than just not being able to afford a place to stay.

So what are those many causes where people have money and choose to live on the streets?

IDK, cheapest phones don't work, and I don't think the store price was ever that clear a differentiator, as telcos routinely disrupt this signal. You can easily get the newest flagships, and phones from the next price group below, cheaper than you can get mid-range phones - as long as you're willing and able to sign up for a 1-2 years contract.

Outside of the US and a few other companies, people pay full price for phones.

Eventually, always, but not up front. Poorer people "prefer" somewhat higher price that gets spread out over years, to paying full price up front.

(I put "prefer" in quotes, because it's less of a choice, and more the only option they can afford.)

Most countries don’t do two year contracts - people have to pay the full price up front.

It means you'll be getting a "This phone isn't available in your country" pop-up when shopping online and all SKUs sold in Europe by the maker's subsidiary will have different model numbers. Of course, because of the need to support the phone for longer on patched kernels and the impossibility of charging for updates, the phones will be more expensive. The removable battery will also mean that the phone isn't waterproof.

Wasn't the Samsung Galaxy S5 waterproof despite having a removable battery?

It was less so than later phones, but it absolutely handled its intended use case of browsing reddit in the shower

In recent years I think phone hardware has managed to outrun software bloat. My latest phone doesn't feel any faster than the previous one and it didn't noticeably slow down after three Android updates. Five years of updates would probably be fine.

Phones nowadays are pretty powerful, even on newer softwares older devices are pretty usable. I actually still daily drive an MSM8974-powered Android phone with an Android 11 custom firmware, and it still runs much better than some of the cheapest phones on the market.

Not necessarily. Software support can just mean security updates, which would already be great. Newer versions can also roll out features dependent on hardware strength only to devices that have the necessary hardware for them. And (if the manufacturers don't do the exact opposite on purpose) could lead to longer usable operation without doubling ram every 3 years.

It would be interesting if a side effect of this proposed legislation, Android manufacturers stop pumping out new models as aggressively. It would make sense if that's how it turned out, because these manufacturers would effectively be crafting their own support nightmares by continuing to put out new models at a high cadence.

This in turn should make for better parts availability since the window that any given model is manufactured in is wider. It'd also incentivize sharing parts between models where possible.

All in all, positive for the end consumer. It might've been a problem in the earlier days of smartphones, but the category has more or less plateaued by now and so the benefit of high model churn with each new model sharing little with its predecessor is dubious at best.

Or it actually accelerates the model pumping. If the software has to be supported for 5 years after it’s last on sale there could be a perverse incentive to roll SKUs like crazy even if there are not any actual changes so you only have to support for the shortest possible time.

That would still effectively be 5 years of updates which is a hell of a lot better than what most manufacturers currently offer.

Sure it would be, but I think the implied point above was that including the period the device is sold (what's that? say three years), the total support time is much closer to the ten years that was being discussed up top, and pumping out more new devices may shorten that back to closer to five years.

Does having 3 devices that you are supporting for 10 years make more work than 10 devices you support for 5?

Because while that doesn't make sense in my head, I could easily see that being the methodology used by the Android device manufacturers.

If these ten devices are mere SKU updates, then maybe yes, they create a sliding five year support window (plus however long they're being sold), instead of a ten year one. Keeping software working for a five year old hardware platform sounds a lot easier than doing that for a ten year old one. A lot less Git branches also.

If the ten devices are mere SKU updates made solely for the reason of getting around the regulations, then the company can expect a rather huge fine :^)

Of course, but any increase is an upgrade to the current situation.

Fantastic point. Chinese manufacturers pump out SKUs like crazy - several new models every six months or so.

Of course, it might turn out as theluketaylor says but I doubt it

Maybe that's just waiting on some settled precedents on what's really a new model. Could this be brought before a judge who'd then possibly rule that SKUs 1, 2, and 3 are essentially the same device and should thus count as such for the purposes meant here?

I would love if this software support needed to be printed on the box. Sort of like EnerGuide or the Nutrition Facts. Feature updates: N years, Security Updates: M years ... This way consumers can actually try to factor in the lifetime of the device when purchasing. Oh this device costs 10% more but will get security updates for 2/3 longer, so it is actually more economical. Then if the company doesn't comply with the label they need to provide full refunds or a replacement that will get updates.

In my experience, the current bottleneck with (Android) phones is that the get very slow without any apparent reasons after a few years.

I've had multiple reports of acquaintances complaining about their phones getting slow after a few years, and I've always assumed that they were just full of apps... until it happened to me. I can't find an unambiguois explanation; my guess is a mix of CPUs deteriorating over time due to heating, and disks due to cell degradation, but it's certainly a mystery.

That interpretation is incredibly generous to the companies that benefit from you feeling a need to replace that phone.

It's not a mystery at all. Planned obsolescence is a thing and manufacturers have all kinds of perverse incentives to ensure that your perfectly serviceable phone gets traded in for a newer model.

I like to call it "premature obsolescence". I don't believe that the manufacturers actively try to make your old phone get slower.

Rather the opposite: they don't try to keep your old phone working. The devs probably have the latest phone, and write code that works well with them. This code can be inefficient, as long as it works well on the newer phone, even if it makes older phones slower.

Maybe I am being pedantic, but I think it is different: it is _not_ that manufacturers actively pay their devs to make the phones slower (which would be planned obsolescence), but rather they _do not pay_ the devs to make sure the old phones _do not get slower_. The latter is "passive", that's (to me) premature obsolescence: the manufacturer could fight premature obsolescence, but it has a cost (and obviously they don't want to pay for it).

There's some degree of active involvement with this. I've long suspected Apple to do this for example. The best version of iOS for any iDevice is usually the one just before the one that they use as the hard cutoff point.

Granted, my big example of this is quite old (the iPhone 4 would run iOS 6 really well, while on iOS 7 it was a glitchy, slow, battery-draining mess) but it does seem to be a common trend (my last Apple phone, the iPhone 5 ran quite well on iOS 9 but would just have a complete breakdown if you installed iOS 10).

They were IIRC also caught red-handed a couple times deliberately releasing updates to older iDevices that would just wreck their battery life and have a few lawsuits pending about that.

“ They were IIRC also caught red-handed a couple times deliberately releasing updates to older iDevices that would just wreck their battery life and have a few lawsuits pending about that.”

Are you going to just drop an inflammatory statement like that and not even try to provide any information to back it up?

Is google down or something?

https://www.npr.org/2020/11/18/936268845/apple-agrees-to-pay...

That has nothing to do with "wrecking battery life" though. They added a fix to prevent the device from turning off early due to battery brownouts at lower states of charge. It did result in reduced performance (and they got a lawsuit cause they weren't open about why/how they did it) but this wasn't causing excessive battery drain.

Lets not use this case as a scapegoat and pretend that slowdowns don't exist in iOS.

Pretty much from the first iPhone model it's well known that newer iOS releases will slow down your phone and/or make your battery drain faster. I even remember when they actually broke wifi on a lot of phones with an update and pretended that the hardware was defective all the time (which it wasn't). Hell I'm looking right now my old iPhone 8 which got a new genuine battery, and it's a lagging, hot mess so I only use it when I don't want to burn the battery of my new phone.

My good colleague resisted updating her iPhone 11 to any newer version of iOS from the one she got, but couple of months ago she couldn't install her banking app anymore so she broke down and updated the OS. Instant laggines, and battery lost close to 40%. Just like that. She got mad, and when the madness subsided, she bought a new iPhone. Mission accomplished for Apple.

All of this could be much avoided if they would allow you to install previous versions of iOS, but they don't for obvious reasons. And before anyone mentions security, that is a load of BS because even previous iOS versions get patched up. So planned or premature obsolesence, the user just gets frustrated with an old phone which hardware-wise might be perfectly fine, but software-wise an unusable mess and buys a new phone. I hope the forced user battery replacement works, and I'm sure they could look into these forced software updates which are also largerly contributing to larger problem at hand.

The specific accusation was about 'wrecking battery life' and the source was the 'batterygate' lawsuits. This has nothing to do with a scapegoat.

Honestly, "it's well known" is anecdotal at best. If you're actually saying it was a problem that battery life dropped 40% from going from one update to another, then media would pick up on this. That's not a small amount here, you're saying battery is now 2/3 the previous runtime. If that was actually reliable and measurable, show me an article that has done testing on before-and-after from one update to another.

The Batterygate was just a partial response at iOS significantly slowing down even one generation old iPhones, and I don't want to go into that because that's not what I want to discuss.

In your second paragraph you completely dismiss my experience I guess because it wasn't published by some online zine which is ridiculous by itself. Online forums like reddit are full of these, my experience says the same, and pretty much the big majority of my friends have iPhones in this neck of woods and there is no question on whats going on. Why there are no articles about it? Dunno, maybe because by now its an old story as this is going on for decades now so it's an accepted fact. Maybe because in US people buy iPhones more often then in Europe so the effect isn't that noticeable? I don't know, and it's not in my interest to care.

So I'll put my claim into negative and ask you this: are you claiming that iPhones are not slowed down by newer updates? Are you claiming that perfectly good iPhones even as new as 1.5 years and older are as performant as the lastest gen iPhones with the same iOS (even with brand new batteries)? If so, thats an extraordinary claim and the burden of proof rests on your shoulders.

Are you also claiming that you're casually able to install previous versions of iOS so you could get performance back? If car industry did the same after software updates whole industries would collapse. Does your PC even do this nowadays? If you install an OS which actually slows down your computer, you can still re-install a previous version to get your performance back. Not so on the iPhone, you're locked out of that option and we'll screw you, buy a new one.

“caught red-handed” implies an intent that is not supported by the evidence.

The design of the battery in the 6 turned out to be problematic once the battery reached a certain age. After lots of reports of iPhones shutting down suddenly, Apple issued a software update that slowed down those phones so that they didn’t shut down. They then redesigned the battery and this was not a problem in later models.

You talk about a general slow down of phones when the OS is updated. That is most likely due to increased processing needed in later OSs. The early phone CPUs had barely enough processing to do what they needed to do so the added demands of a new OS would drag them down a little. Over time the cumulative effect would add up. This is not unique to iPhones. I have heard the same complaints from Android users. Fortunately modern phone CPUs are powerful enough that they have the cycles to spare and slowdowns are rare.

I have an iPhone XS. For what I do, the hardware is still perfectly capable. I use basically the same apps I used 5 years ago. Except now, I often can’t run more than 1 app at a time. I’ll have an app open, switch to another app and it’ll reload from scratch. Then when I switch back to the first app, it too will reload from scratch. Multitasking can be very painful.

My understanding is that apps use more RAM now because newer phones have more RAM. The result is that older phones with less RAM will aggressively kill apps because there just isn’t enough memory to go around.

As a developer I can see both sides of the coin. On one hand, I know it would be possible to reduce the memory consumption. On the other hand, I realise this would also cost more money. But it isn’t malice. Just a choice.

Apple as a manufacturer have gotten a lot better over the years. I remember when iOS updates made the iPhone 3G literally unusable, but that isn’t the case any more. I’m running iOS 16 on my iPhone XS with no performance issues at all. I expect iOS 17 to be just as fast. They also implement features to increase the longevity of their phones, like reducing performance and maximum power delivery when the battery is degraded. They were lamented for this feature but it genuinely saved my iPhone 6S which kept on shutting off.

I haven't noticed this at all. My phone is 4 years old and feels just as fast as always. My previous one I remember thinking how snappy it was after a factory reset when I sold it on. The one before that died. The one before that kept its speed but the battery turned to shit and I couldn't be bothered trying to get it changed. I'm just about to replace this one but it's still in great working order so one of my kids wants it as their first phone.

There was one Android device my wife had that got slower and slower and it was a known issue with the disk.

Android seems to suffer from a phenomenon similar to "Windows creep" where performance decays over time for no reason in particular, even for users who are good stewards of their devices. It's more of an issue for some devices than it is for others for reasons that are unclear.

I used to think this at well, but then I removed all of the useless apps I'd gathered over the years. The improvements were almost instant. Installing apps doesn't have a noticeable impact on general system performance, but install enough of them and you'll find everything mote sluggish than you remember.

My phone feels as fast as the day I bought it, which so far must've been about four years ago now. It came out with Android 9 and is currently running Android 13 through a custom ROM.

Newer devices certainly feel snappier, but the old phone works as well as it ever did, except for the battery carrying almost half of its original capacity. I still have my OnePlus One and I can't say it feels like the phone is slower than when I bought it.

I think the slowness is caused more by the change in perception thsnt be degradation of hardware and the ever growing software bloat. Modern phones software comes with fluid, high resolution, high speed animations and transitions everywhere that don't work well on older devices. Once you turn those off, you'll quickly find that many older devices will work just as well with modern software as they did with their older software.

Bargain bin devices that come out with specs that can't even run the factory firmware are the exception, of course. However, that segment seems to have shrunk significantly since I last looked into buying a phone.

Regarding my own case, it was a Google Nexus 5X. Performaed a factory reset, but the phone is still super-slow.

100% sure that this is not perception. I'm talking about seconds-long delays.

Regarding the acquaintances, I don't remember (or know); I think a couple of them were old Samsung Galaxy S models. Again, factory reset didn't help.

I guess that as parent posted, this problem affects some models more than others. I'm still really curious about the technical cause :)

I believe several generations of Samsung devices have a hardware issue with their flash storage that will make the storage slow down to a crawl before bricking itself. The problem only appeared a few years after release and it doesn't seem to affect every device either. Google and Samsung were most commonly affected if I recall correctly.

My old tablet suffers from the same problem. It's quite frustrating to use because of it, but luckily I'm only planning on using it for postmarketOS and that can boot from the SD card slot.

Other brands didn't have this problem and the problem has since been rectified in the Pixel/Galaxy lines. I suppose some very cheap devices may still be using trash flash storage chips, but it's no longer an issue with Android phones from the past few years.

I have a 9 year old phone (Galaxy Note 4, yay for removable batteries!) as a daily driver, now on LineageOS and while it is true that removing the bloat will help a lot, it is very clear that the slowdown is real.

Among the most noticeable worst offenders is Google Maps, but many apps get slower update after update while not offering much more than cosmetic changes. Websites too, many apps are just websites bundled with a browser anyways. Apps that didn't receive updates for many years are as snappy as they ever were.

I had a Samsung Galaxy S2 and a Nexus 10 which both suffered from flash related problems making them incredibly slow. Samsung never fixed that nor was there ever a recall.

I also owned a HTC One M7 which worked well for a very long time but it also got slow even though there were no hardware problems. The M7 worked just as good from day one if I factory reset it, it only becomes slow again once all the apps are updated. What I think is happening here is that apps move to the newer Android API but keep compatibility with older Android versions through compatibility layers, and they compound so once the entire compatibility layer stack becomes "too thick" the CPU gets bogged down and the whole thing becomes slow...

I finally made the switch last year, but I used a moto G (rather popular model, at least in Europe: https://wiki.lineageos.org/devices/osprey/ ) for some 6 years.

It did slow down at one point, but it came back to life after installing lineageos, so now I blame bloat for the slow down phones suffer from.

The bottleneck on the Asus phone I replaced recently and I had since 2017 or so (before that I had T-Mobile G1 for maybe 8 years) was the glass - it breaks really easily because there is no plastic protection along the edges as it used to be on older phones.

How old? I'm still using two LG V20's from 2016 and don't experience issues.

> CPUs deteriorating over time due to heating

I might be wrong, but AFAIK this is not a thing within the usual timescales.

It can't be excluded. I remember a friend of mine overclocking his CPU (around 20 years ago!), and notificing that the maximum clock they could achieve, would decrease over time.

I guess that when put under thermal stress (which may be the case for some phone, for whatever reason) may actually have a measurable effect.

It's probably due to old battery causing imperfect thermal control then extreme throttling. Phones are not really well tested on aged batteries since it's hard to artificially make such one that fits well to real world scenario, and manufacturers also don't have good incentives to do so.

One contributor is that most models get a mid life major Android version update after 2 ish years and those don't work as well as the launch versions. It's unlikely to be hardware problems you suggest. Flash ftl slowdown due to filling up or fragmentation from use might contribute though.

There are some jank debugging tools on Android, would be interesting to see what they report on a fresh vs slowness afflicted device.

> But basically I think phones have now reached a place like laptops and desktops where HW capabilities are stable, and there's no good reason for devices to be obsolete in 5 years any more.

Unless you care about your phone working as a phone. 2G was discontinued years ago, 3G is rapidly being dropped by the carriers and how much longer before 4G is discontinued?

> streamlining the Android update process? (To Google's credit they have already done a _lot_ of positive work in that area as I understand it).

What exactly is Google doing? Apple released a security update for the 2013 iPhone 5s earlier this year.

3G is alive and well in IoT.

All of the US carriers are shutting down

https://www.pcmag.com/how-to/the-3g-shutdown-how-will-it-aff....

What I implied was it might not be a big thing for smartphones, but there are a lot of IoT devices out there that will fail the 3G sunset.

Even in your example you point out that 3G "is being rapidly dropped" but it's not dead yet. The iPhone 3G came out in 2008. So just today, 15 years later, it would still work on most networks (certainly I can still use 3G in Australia).

Then there's still wifi of course. Admittedly, a phone without cellular connections is less useful, but it could still be used to function like a tablet that's simply kept at home etc.

> Unless you care about your phone working as a phone

Most of that is a chicken-and-egg problem. 3G was dropped because very few phones were still using it, and you could not run 3G on the same frequency as 4G/5G. If phones lasted longer, more people would've remained on 3G, so they would not have dropped it. Furthermore, 4G and 5G can share the same equipment and frequency due to Dynamic Spectrum Sharing, so there is no technical reason to quickly get rid of 4G like we wanted with 2G & 3G.

3G was dropped because the same frequency could better used for 4G. Unless you somehow found a way to get around the laws of physics.

Not here in the US. The carriers wanted to discontinue 3G for years. They really only started it in earnest in 2018. It took them about 3 years longer from the first dates given to finally kill 3G. It sound to me as if 3G phones didn't bite the dust as fast as carriers assumed or customers refused to upgrade at the rate they expected.

That's the situation in the US. In most of the EU, you can expect either 2G or 3G to be available for at least five more years.

Nope. 3G is being shut down in Finland as we speak. Should be finished by early 2024.

It's pretty well explained here what Google is trying to do: https://www.xda-developers.com/android-project-mainline-modu...

That being said, Apple is atm way better in providing longterm software updates, but that doesn't mean Google isnt doing a lot of work there also.

Next to that, there is already a lot of apps that are updated through the Play Store instead of an OS update. Mail, Dialer, Android Auto, Photos, Camera, Keyboard, Contacts, Clock, Play services, ...

They’ve been “trying to” do the same thing since 2014.

https://en.m.wikipedia.org/wiki/Android_One

And none of those apps help when there are security vulnerabilities at lower levels.

> Unless you care about your phone working as a phone. 2G was discontinued years ago, 3G is rapidly being dropped by the carriers and how much longer before 4G is discontinued?

First 4g devices were released in 2012, in 2014 you could find sub 150€ 4G devices. If you bought the last 3G phone in 2015~ (there were still 3G phones after that, but nothing reasonable), 10 years of network support is still reasonably within reach.

> To Google's credit they have already done a _lot_ of positive work in that area as I understand it

They haven't. I've extensively worked on their Project Treble, and they did a beautiful work to perfectly versionize every behavior OEM might depend on, or version driver APIs. They could use that to provide long term support (which is what I do in my own project, the foundation is very sound), but in practice they use it as a way to obsolete code faster:

Before Project Treble, Android kept whatever code was in use. So if some OEM was still using a 5 year old code path, and it was still working fine it would remain.

After Project Treble, every driver API has a new version every year, and the older ones will be deprecated (read: purposely broken or completely removed) within 3 years.

Before project treble, OEMs could maintain the Android for their platform quite effortlessly for 5 years, nowadays they have to rewrite whole parts of drivers after 3 years.

someone will come and say that since Project Treble OEMs have been upgrading for longer. But I consider it's completely a matter of timing, the smartphone market stagnating. Look at nVidia Shield and fairphones for reference.

Qualcomm is still terrible at getting their SoCs supported in mainline kernel and as a result as soon as they stop providing updates to the kernel fork you're SoL.

This can happen as early as 3 years after initial SoC launch

>I think the next bottleneck is likely to be software longevity.

My girlfriend recently dropped (and broke) her iPhone X. She then bought a used Galaxy S9. The S9 is a few months older than the iPhone X.

Despite that, the app for our home alarm is not supported on her S9, because for some unknown reason it requires Android 11+, and the S9 only supports up to Android 10.