Starting on November 1, 2022, existing apps that don’t target an API level within two years of the latest major Android release version will not be available for discovery or installation for new users with devices running Android OS versions higher than apps’ target API level. As new Android OS versions launch in the future, the requirement window will adjust accordingly.
This is a very welcome move, since finding incredibly old and abandoned applications is not an uncommon occurrence in the Play Store. Clean-ups like this almost make up for Google removing the “last updated on” field in Play Store listings.
Almost.
There is always a trade-off.
Yes, it is very good for the average end user.
But for software preservation, it is not a good sign.
Hopefully, they will still allow a path to discover these applications (web search?). However, if history is any guide, I expect they will be completely removed after a certain period.
I see the software preservation point-of-view made often but it’s not very convincing in my opinion. I simply don’t believe that all software ever written needs to be archived for all time. Of the gazillion apps that do X, what benefit is there in keeping them all, forever? Stuff that has some kind of significance, sure, no problem, that makes sense. But, keeping things just because they exist? That’s what hoarders do.
friedchicken,
hoarding sometimes helps 🙂 Every now and then some collector comes up with a lost recording of older Doctor Who shows, or even NASA moon landing tapes:
https://www.cnet.com/science/features/nasas-lost-moon-landing-footage-and-the-man-who-brought-it-back-to-life/
But, yes, there needs to be some kinds of discipline. Maybe they can be put in some soft of “cold” storage.
Specifically for software, the issue is not with random stuff (which might easily be archived as a two minute screen recording), but those for opening obscure file types, communicating with proprietary hardware, or some other sort of unique software that no longer has the code, or would not built even if we had them in the first place.
I’ve recently come across a rare and unusual computer, that as far as i can discern, is almost entirely known purely because of documentation and emulation targets. Without that documentation, and archived software, i’d be left with a very shiny gold brick.
Said shiny gold brick: https://imgur.com/gallery/xwSvG8A
Yeah, there should be an archive, but it should be difficult to access for the nonskilled. There are likely to be all sorts of security vulns in them as time goes on and the libraries they were built on get exploited.
Right now I can spin up and play the dos games I played as a kid, but any generation that grew up on smart phone or tablet games, won’t have that ability.
I think there’s another way of looking at it. Google Play store should have never been considered a place for preservation. Now people will be motivated to make copies of these applications to real archives.
This is not that simple. Remember the google Android API has stuff like “Google Pay API” these are dependent on servers. There is a lot of software in google play store that does not work right because they have not been updated so are using a discontinued internet service this can be the google play stuff or other third parties.
Its part a practical thing how can you find software that does not work and remove it from the store. If it has not had updates in the last 2 years there high risk that its non functional due to depending on something internet that no longer exists. This is one of the modern problems for software preservation and software providing problem.
I love the people / anybody can author an App and get it published, I hate that people /anybody can author an App and get it published,
Removing some of the unsupported Apps from the Store won’t solve problems for those with them already installed, forcibly removing unmaintained Apps from devices will cause untold issues to end users, leaving unmaintained Apps on devices might cause untold issues for end users.
Having struggled for years to maintain and support legacy hardware and software, I’m not sure what the answer is to this problem. This is a problem of the industry’s own making.
Emulation and isolation is the answer.
But nobody got it right, so far. Windows has been trying to do this, since Windows 7 with their XP mode. They were recently supposed to bring Docker like containers for desktop apps. But those are no longer even in mention anymore.
Android can run containers. My old Samsung phone had a Linux DeX mode, which allowed running Ubuntu or other distributions when attacked to a monitor. But that was also dropped. Again, it was a Samsung only solution in the first place.
sukru,
That’s a good point.
I think apple may have got emulation right with rosetta. Of course a big problem with this is that apple are guilty of dropping support after a few years. They develop the technology for transitional periods and not with the intention of providing long term compatibility. But in principal once you have a working emulator and many applications that work under it, they can continue to be supported with relatively little work via the emulator. I don’t think apple has an incentive to support long term emulation. IMHO the ideal solution would be open source such that the community could support it indefinitely, although obviously this is unlikely to happen for rosetta.
If it works, why fix it though?
Ads. More ads! Even more ads!
If you don’t update an app, it will bit rot. If you use any third party libraries, they’ll be either exploited or will disappear off the face of the earth. If you aren’t constantly updating your app, when you finally need to due to a vuln or a desire to add a new feature, it will be pain. All the apis have changed, the old library you really liked and did its job perfectly no longer works with some other api you use, or the new version kills off what made it great. So that simple text change or what ever turns into a near rewrite, which may not even end up being as good as the original. So you have to then go to the company and say ” hey yeah it looks like changing that word should be simple, but it will take a week or two of dev time”.
I despise the planned obsolescence of the entire Android ecosystem. If there are tons of users of an old app it’s probably because they’re stuck on “old” hardware that can’t be updated due to Google’s inane decisions about how Android should (or shouldn’t) be updated.
It is more like lack of foresight: it is Linux, all should be easy, right, … right?
Pushing a security update requires collaboration from the chipset designer, phone maker, and mobile network operator, at minimum. And sometimes those are no longer around (remember TI OMAP fiasco?).
In a hindsight, Windows Phone supporting only a single chipset was the better decision. Then again, their OS was not up to modern standards.
sukru,
I think we need to distinguish between the notion of hardware & OS compatibility with OS & software compatibility especially since in the case of android it involves different parties. Google determines when to break OS & software compatibility, meanwhile it is the manufacturers who determine when to break hardware & OS compatibility. The app support problem faced by users is a consequence of both incompatibilities. Solving either of these would effectively solve the problem.
It is extremely unfortunate that android has never managed to beat the hardware & OS compatibility issues. This comes down to the lack of ABI combined with the dependence on long term drivers. It’s a another case of nothing getting fixed because we have different parties blaming each other while allowing android consumers to suffer the consequences. The manufacturers consider planned obsolescence a “feature”, so they’re not too bothered and Linux devs don’t want to provide long term ABIs. This creates a clear stalemate. I’ve long wished that google would use its muscle to rectify the situation somehow. It remains to be seen how it pans out, but ultimately Fuscia could be the solution that we were never able to solve for android/linux.
Yeah, that’s exactly what I had in mind, IMO it’s basically the entire mobile ecosystem that they got wrong from the start. Or maybe they got it right for their business interests by designing a system that would encourage/require frequent hardware upgrades with the plausible excuse of “it’s because the ecosystem requires it”. The typical PC is also a hodgepodge of components from different manufacturers, many with proprietary firmware. And yet I can still install any modern Linux distro with full hardware support on my 13 year old laptop, and that’s with the admittedly chaotic hodgepog of the Linux ecosystem somehow working on the hodgepog of components that the laptop has. Imagine what a much more organized large company with practically unlimited resources could do if they wanted to… except they don’t.
Just crossed replies with Alfman. You echoed my sentiments, except that I’m less than hopeful that they will suddenly value their customers over their profits with the Fuchsia ecosystem.
rahim123,
Haha. I don’t know about Fuchsia yet either. Google certainly has the ability to solve these longstanding support problems and I absolutely think they should. I do worry about the extent of their commitment to the FOSS community. They may put crucial features inside of proprietary blobs in ways that give google control owner freedoms and privacy while creating roadblocks for alternative services. I’ll be disappointed, but not surprised if enforcing owner restrictions is part of their plan. Google is a for profit corporation after all.
Alfman, rahim123,
In a hindsight we can see some decisions could have been better, rather much better.
However we should remember the environment Android was released in back in the day (2008). I had the original HTC Dream (G1):
https://en.wikipedia.org/wiki/HTC_Dream
Back then phones did not receive updates. Even smartphones rarely did (I had several Windows Mobile devices before). And devices were already being replaced by two year contracts as the norm.
It is not too far off seeing Google trying to quickly launch *something* that would go out of the door to compete with Symbian, Blackberry, Windows Mobile, and yes the early iPhone.
(Personally, I would also change Java to C#. Given almost 15 years of needless patent and copyright struggles, today it seems like the clearer choice).
sukru,
I wouldn’t mind leaving java behind. I like C# better too, or dlang is nice too. But in practice starting up a new ecosystem would be extremely challenging even for google. Not even microsoft managed to build up a phone ecosystem. So I suspect a Fuchsia phone will have to at least support APKs to have a chance at being viable. The more transparent this support is, the less customers will resist.
Alas this is the reason win32s have stuck around for so long. We have better replacements including the .net framework, but too much has been invested in win32 to throw it away and rebuild. Even microsoft themselves have trouble dogfooding their own replacements.
Is this planned obsolescence? App authors have a choice: make app for new Android (and be installable on flagships) or make app for majority of phones (older APIs) but unavailable on new headsets. This is impossible choice. This move won’t make old, still working Android phones disappear or be magically upgraded to new API.
Planned obsolescence is a strange argument considering 3G networks were recently shut down in the U.S. Even if you want to keep your phone around forever, eventually you’ll still have to buy a new one if you want to keep using it as a phone.
Sadly there are people who think retiring 3G is planned obsolescence. They probably also thought retiring VHS was a conspiracy for force people into buying DVD players. The fact that something may still work doesn’t somehow shield it from becoming antiquated, nor should it.
Planned obsolescence is certainly a disgusting and abused reality of capitalism. But, it’s not automatically a dirty term either. There are two sides to that coin.
friedchicken,
I understand your point. However I personally wouldn’t use “planned obsolescence” to describe changing consumer demand…
Planned obsolescence is nefarious because it’s designed to accelerate replacement by making old products become worse instead of making new products better.
No, app authors are not forced into that choice. An Android app can specify both a “target API level” and a “minimum API level”. This means that the app complies with all the new requirements of the target API level on devices that use that API level but it is still compatible with devices down to the minimum API level. At runtime, the app can check the device’s API level in order to disable app features that depend on APIs that the device does not support. See the documentation: https://developer.android.com/guide/topics/manifest/uses-sdk-element
This changes the situation completely! Thanks for the explanation.
It looks like but that would be way too sinister, especially with a paltry 2y window. It takes 1-2y for many phones to catch up to released Andriod version, that would make them totally useless in this reality.
Hiding them by default and making them visible if turned on somewhere would be less user-hostile and authoritarian.
Time for archive.org to kick into gear?
https://archive.org/details/apkarchive
Thank, yet I didn’t found “Forest” wallpaper by Tom Barrasso.
I was talking about specifically acting to archive apps about to be taken down.
Not a general APK gallery that randoms can upload random APKs to.
Then maybe you want this: https://www.apkmirror.com/
(Found via https://wiki.archiveteam.org/index.php/APKMirror)
dboddie,
I’ve needed to use sources including that one to install APKs on lineageos devices that don’t have google services. I’m very thankful to have 3rd party services like this around even if sometimes they’re missing things. Although I’m not sure about the legality of such mirrors given the fact that they’re redistributing software without permission of the author.
This is horrible, terrible and asinine as tons of older apps and games in Google Play perfectly work, don’t need Internet access and don’t shove ads in your face. Or, and many new apps are malware in disguise as they have features to update their behavior using control servers on the Internet.
New modern apps? The only reason Google does that is to increase their profits.
Lastly, goodbye hundreds of thousands of perfectly working apps and games which are no longer maintained.
You got it wrong Thom, terribly wrong.
I agree, perhaps this move made sense 7 years ago when google couldn’t make up their mind about visual language but now? Android is now a mature OS in more or less maintenance mode, changes between releases are far from revolutionary and most of the apps (at least those from version 5 upwards) fit in w/o issue. And that with platform that is so weak in the testing cost department while promoting price race to the bottom.
It will push more and more independent publishers below profitability level and will result in the pie split among less and less number of big cats and while limiting the choice of the customer. If anything forcing everybody to retest every 2 years on all devices puts less focus on google’s backwards compatibility promises and allows it to get away with less than stellar track record.
Google is a search company and it should have the abundance problem in app area solved long time ago.
Artem S. Tashkinov,
I hadn’t even contemplated that the quality of new apps might actually be statistically worse. Do we have any data on this?
I mean this trend does seem to happen in other industries as they mature & consolidate. There’s less focus on innovation and more focus on squeezing their customer base for profit.
Yep, and that’s the case as well.
More code, more bloat, agile development, less QA/QC, and a “we can fix it later” approach as perfectly seen with modern games which are buggy and bug-ridden as hell for months after the “golden” release.
I remember games of the past which worked perfectly right from the get go. Nowadays it’s considered luxury.
A race to the bottom indeed.
I think there is a possibility new apps are worse, but Honestly, there were a lot of terrible early apps too. I’m not sure if its gotten better or worse really.
Bill Shooter of Bul,
I’m not aware of any analytics having been done to measure this. Most apps are just “disposable” throw away apps anyways, loosing those wouldn’t be a big deal. But other times it’s very important to have a specific app, like those that are needed to interface to specific hardware. It would be very frustrating for google to delete the applications that go with said hardware. We have an O2 & heart monitor for example. Also printing & scanning, IOT apps, etc.