The Fairphone 5 is official and full of surprises. As you might expect, it’s the usual repairable phone from Fairphone, with parts available to order and easily installable with just a screwdriver. A new phone means faster components and a more modern design. What you might not expect is Fairphone opting entirely out of Qualcomm’s consumer upgrade cycle thanks to its choice of an “industrial IoT” SoC that promises longer support times. With a longer window from Qualcomm and a commitment from Fairphone to keep going even after Qualcomm’s industrial support cycle, Fairphone says this device will end up with a jaw-dropping 8–10 years of OS support.
The Fairphone 5 is not for sale in the US. Europeans, though, can get the device for 699 euros (~$753), with preorders starting today and a ship date of September 14. For the basic specs, we have a mid-range loadout, starting with a 6.46-inch, 90 Hz, 2770×1224 OLED display. There’s 8GB of RAM, 256GB of storage, a side fingerprint reader, and a microSD slot. For rear cameras, it offers a 50 MP Sony IMX800, an anonymous 50 MP wide-angle sensor, and a time-of-flight sensor. The front cam is a 50 MP Samsung JN1.
Such a support cycle should be legally mandated for every OEM.
“Such a support cycle should be legally mandated for every OEM”
I am not a big fan of the current state of affairs, and it is good to see an option in the market that can let the market speak, but this comment made me roll my eyes a little. You are suggesting that every maker of something like a smart phone should be legally required to offer 10 years of support?
First, from a “what people want” perspective, the useful lifetime of these devices is 5 years ( or much less ) for most people. So that would be an extreme duration to demand legal protections.
More importantly though, almost none of the participants in the supply chain for these kinds of devices are setup for this. As the article states, to do this, very specific SoCs from Qualcomm have had to be selected which are being targeted for long-term availability for industrial uses. The majority of Qualcomm’s product line is going to be available for sale for a fraction of this period. Some of these chips will be pulled from the market after as little as 3 years ( because they have been replaced by newer chips are are considered fully obsolete ). I am not talking abut the phone maker but about the chip makers. The OEM would be forced to go the grey or black market. At scale, these things are just not going to be available at all.
So, to do what you want, you would have to change the way the industry works. “Good!”, you might shout. Well, this would come with consequences. It is not this way because companies are evil ( not only let’s say ). This is how industry players give customers what they want. If you want slower and more expensive technology that evolves more slowly, we can give everything, everywhere 10 years of support. The thing is, as a group, nobody wants that.
Again, I am happy to see options in the market. If it turns out that everybody DOES want this, the market will tell us.
I am not a “free market knows all” person which ( reading above ), it may sound like I am. Mostly, I believe in smart regulation. I guess what I am saying here though that mandating 10 years of support for a device like a smartphone would not qualify, in my view, as “smart” regulation.
I realize my reply was a bit “hardware” centric whereas the Thom’s comment may have been more about OS updates. These are not as disconnected as many think though. Again, for these kinds of SoCs, there is a lot of “secret sauce” and the OEM is dependent on the chip maker to keep drivers and OS current.
Moving towards options like RISC-V will really help here but, even then, individual designs will be fast moving and proprietary.
tanishaj,
Ideally RISC-V would help, but since it’s being commercialized by the same industry players that have made ARM difficult to support, While RISC-V a great opportunity to offer an alternative to x86 that is equally accessible and portable, I am very concerned that RISC-V could turn out more like ARM. x86 may be alone in being able to offer relatively strong platform standards across vendors. I wish I could choose from more architectures without having to give up easy portability and standards.
tanishaj,
I would say absolutely yes. There’s no reason we as a society should tolerate having to throw away working hardware on account of bad software support. 10 years is reasonable especially considering the market is mature and new hardware basically has the same features, same form factor, same architecture, and so on…. users shouldn’t be forced to throw out hardware that works on account of lack of support.
I’ve said this before, but manufacturers should have a choice:
1) officially support devices a full 10 years
2) provide users and FOSS community with the resources necessary to provide their own support without the manufacturer.
The justification here is that It would matter much less that the official OS isn’t supported long term if users weren’t so dependent on the official OS in the first place. This dependency is the crux of the problem!
There’s no technical reason mobile devices have such poor support windows compared to PCs. But when the manufacturers control the operating system, it creates perverse incentives to shorten lifespan to maximize hardware sales. And this is what we’re seeing today. If Dell, Compaq, HP, etc controlled the OS, their incentive would be to do the same thing. The reason computers remain usable so long after the manufacturer drops support is that they don’t control the operating systems and the parties that do don’t have the same incentive to drop support.
I disagree. Now that the market is mature, most people I know aren’t nearly excited about upgrading as they used to be. Yay, a new phone, same as the old phone, only supported. The last upgrade cycle for most users was compulsive when their carriers mandated new 5G phones and blacklisted older phones. The math is easy: shorter upgrade cycle = more sales & more profits. Several of my family and relatives alone got affected by this with phones less than two years old having to be replaced.
I agree, Qualcomm and friends are central to this problem. I personally believe this will not improve because they are benefactors of a rapid upgrade cycle too. Products becoming disposable and hard to reuse is a good thing as far as they’re concerned.
What are you basing that on? My extended family collectively groaned at having to spend more money to buy new phones. The fact that they ended up being forced to is not evidence it is what they wanted. While some people are eager to upgrade, I personally get very little from new phone hardware and I suspect millions of others feel this way too. Replacing one slab of glass with another doesn’t do anything for me personally and I even limit charging to 75% to extend the battery lifespan as long as possible, but I can’t do as much about support.
I think a bigger change would be mandating a standard ABI and open drivers along with user replaceable batteries so that a bog standard Android ROM would run on just about anything, just as I have Windows 10 running on a Phenom II X4, then it will be up to the user to decide if it needs to be replaced because the new OS is too heavy or if they can stand it.
Because I am sitting here looking at an Octocore phone that will probably end up in a landfill because its stuck on Android 10 and never got a single update from the OEM which is one of the reasons I got a new OnePlus as I hear they are good for 4-5 years’ worth of updates.
Sadly the old phone, frankly along with the last 3 phones before it, works just fine, but without security updates I simply do not want to be doing anything requiring security and since my phone before it is my PMP and has better battery life I guess into the ever growing pile of e-waste it goes.
Not everything is or can be open source as of today. AOSP (open source version of Android) does not cover the radio roms developped by the SoC builders. This piece of code is totally filled with patented stuff and thus, harder to open source. On Fairphone 4 and earlier models, they had not addressed this particular issue, so some kind of security issues cannot be fixed on these phones. Using an industrial-grade SoC is an interesting workaround.
deedub,
Overlooking my negative view of patents, I don’t see why patented stuff should make anything harder to open source. Patents are explicitly supposed to be completely public. In fact this was the entire justification for having patents – the invention becomes public and there are no secrets.
Patents are the opposite of trade secrets and there is no reason that patented intentions should not be open sourced. If anything it would have very sensible and logical for the government to mandate open source code be submitted when they started granting software patents. Unfortunately patent holding corporations have not been held to account. They’ve essentially been allowed to have their cake and eat it too by issuing patents that are so cryptic and useless that almost all of them belong in the trash.
“the useful lifetime of these devices is 5 years ( or much less ) for most people.”
Based on what? Sorry, but this sounds like complete bs to me. My current phone is easily five years old and does everything it needs to in daily use. If you buy a relatively decent phone, most people probably won’t notice any difference when it comes to software, even in a few years. The problem are the providers, not giving a shit even when it comes to security-updates.
“almost none of the participants in the supply chain for these kinds of devices are setup for this. ”
And? There’s nothing preventing this from happening. But it won’t happen magically, they will have to be forced to do it, as is so often the case. Anyhow, it’s no natural law.
“If you want slower and more expensive technology that evolves more slowly, we can give everything, everywhere 10 years of support. The thing is, as a group, nobody wants that.”
Which development in the last five years was actually breathtaking and absolutely necessary in the mobile phone space? Mentioning “slower” performance these days is just ridiculous. Most people won’t notice any significant difference (if any at all) if you’d give them a slightly slower phone than they currently have. And that’s why most people wouldn’t care. Your statement “nobody wants that” is at best unproven, at worst, disproven by Fairphone. This is already the fifth version of their phone, so apparently there are enough people to buy their products.
Whaaa??? How???
There are two ways to support an Android phone for such a long time:
1. Stick to a single major version of Android and rely on Google’s long-term support for that major version
2. Rely on the SoC vendor to consistently release new drivers for new Android major versions (you see, when dealing with warranties, you can’t just combine old drivers with new Android versions like the XDA people are doing in those ROMs that come strictly without a warranty, you need an official “board support package”, aka drivers for the specific Android version you are shipping).
3. A combination of #1 and #2
So, here comes the question: Has Fairphone gotten the relevant commitment from the SoC vendor or Google, or a combination of both, so the total support is at least 8 years, or are they talking out of their posteriors (aka the 8 years of support is a “best effort” thing)?
kurkosdr,
Fairphone are unlikely to get much help from soc vendors, so #1 seems most obvious to me.
Those of us who build linux kernels will appreciate that at least when it comes to userspace APIs compatibility, linux userspace doesn’t tend to be that coupled to specific kernel versions. I’m not really sure about the long term stability of google’s own kernel customizations. But linux distros in general have a high degree of forwards and backwards userspace compatibility. So in terms of fairphone hardware, they’re probably stuck on specific kernel versions, which is not ideal, but at least they should be able to provide userspace updates on old kernels.
LineageOS faces the same restrictions in terms of being tied to an upstream kernel.
They’re using an industrial SoC from Qualcomm that has such a long support window.
That’s 5 years of support, so the remaining 3 will be Google long-term support for a given Android major version or a “best effort” thing?
Also, there is the issue of kernel vulnerabilities (think futex or towelroot), so anything after 5 years is “best effort” if there is no way to get a new kernel from the SoC vendor.
That’s why I am asking: Nothing in the nasty, nasty world of ARM is supported for 8 years.
futex or towelroot = futex aka towelroot
Considering the Fairphone 4 is not even 2 years old, let’s see how far this increment goes.
Btw, used a 2011 HTC Evo3D (Android 2.35 up to 4.2, removable battery, glued screen) for almost 9 years, now on a 2014 LG G3 (Android 4.4 up to 6, removable battery, screwed chassis with screen) since 2021, currently no much problem beside each Play Store update render it slower and slower, yet battery usage keeps reasonable.
While I can understand new iterations can bring “benefits” (more “power”, less power drain), the ecological footprint of the new phone remains important and not necessarily worth the “upgrade”. At least for my usage (phone most of the time, camera and 2FA applications).
If you need the new iPhone update every year or so, it’s up to you, but some people just live really well with a phone that doesn’t “evolve” on a regular basis. At least the latest Fairphone 4 updates did more wrong than good, the UI became more despicable (fuck the “hype”) and the camera is still damn slow (+5s to take a picture that is just average and barely better than the LG G3).
Hence, I’m all up for these “slow paced” phones, yet the price is huuuuge and on par with more regular brands you may change 2 or 3 times in the same (expected) lifetime of a Fairphone. Even the replacement parts are priced a bit “silly” : USC-C port: 19.95 ($22). Really ?
Sounds like Fairphone will finally beat the OS support of iPhone (7 years).
And I agree, support like this ought to be required! Phones pught to be usable (securely!) for 10 years.
Greetings.
Since I have quite a few VPN 2FA Applications running on my phone, I certainly would welcome a device running for 8+ years. I would keep it solely for the purpose of the 2FA and Password stuff, more like a keyring device.
@Thom:
You really can’t mandate long warranties or support cycles. The corporation would just bankrupt and there is nothing you can do about. Regulation makes only sense when you are able to enforce it efficiently. (The EU still has to learn that lesson.)
The real problem is not warranties or support, but waste and resource management. So find a way that each producer covers the actual cost of disposal immediately when selling the gadget. Then the consumer would automatically think about the purchase twice and opt for more sustainable models.
After 8+ years, how would migration from an old to a new device work though?!
The good:
– 8 years of updates
– repairable
– microSD card
– removable battery
– DivestOS and CalyxOS support hopefully coming? (They support the Fairphone 4 already.) GrapheneOS seems to not want to support anything except for Pixels even if the device supports bootloader relocking.
The bad:
– WAY TOO BIG, this thing is like 1.5 cm taller than the iPhone! I understand it’s easier and cheaper to make a large phone but come on, this is a phablet not a normal phone, It won’t be comfortable in your pockets, unless you wear cargo pants or a handbag. And forget about using it with one hand. It also weighs a lot more (212 g) than the competition (Apple/Samsung flagships ~160 g).
– Removed the headphone jack. Most likely to sell their stupid Fair-branded wireless buds.
– Only IP55 certified. I remember in the olden days, all those Samsungs with removable batteries and headphone jacks already managed to get IP68.
As expected this comes with a real cost.
The long term supported SoC from Qualcomm is roughly equivalent to 2 year old ~780G:
https://en.wikipedia.org/wiki/Kryo#Kryo_600_Series
This has always been the problem with the closed ARM ecosystem. For example, when Texas Instruments left the market, they also left a lot of manufacturers hanging out with the then famous OMAP chipsets. They could not receive official updates to the next planned Android versions.
Anyway, there is always a trade-off. You can get the latest and most performant CPU/GPU, at the cost of short support cycles, or you can get an industrial SoC like this phone with slightly older hardware.
For clarification: my main concern is the closed and divided nature of the ARM ecosystem.