Hacker News Reader

Comment by freehunter | Hacker News Reader

Yeah, you really have three options now. Either (1) never introduce features that require more processing power but newer phones are capable of handling, (2) have those new features but lock older phones out of being able to use them, or (3), stop supporting older phones with your updates.

The problem with (1) is now you've completely stagnated the market, and if your competitors choose either of the other options you're left lagging behind and you lose sales.

The problem with (2) is now customers of older phones will be upset that they didn't get the full upgrade for reasons they don't understand. The iPhone 4S and Siri or the iPhone X and Animoji show the reaction you should expect.

The problem with (3) is now you have planned obsolescence and old phones just become throw-away leading to an ecological disaster.

None of these are good options.

View on HN

> (3) is now you have planned obsolescence and old phones just become throw-away leading to an ecological disaster.

I don't see how planned obsolescence via missing bells and whistles present in new phones is any worse than planned obsolescence via all functionality getting slower.

Written on my iPhone 5S (2013) running iOS 8 (2014) - it is still a phone and is just as fast as it was in 2014. Is that planned obsolescence?

I'm OK with manufactures not providing full-version/new-feature updates to old phones, as long as they provide frequent security fixes - which is one thing Google is doing right with Android security fixes.

Samsung could trivially provide security updates on old version as Google does most of the heavy-lifting so they have no excuse for recommending upgrades on older, slower phones. Apple doesn't presently have a mechanism for this on iOS, to my knowledge.

Yes, but the problem is that the casual user will view this as planned obsolescence rather than a delicate tradeoff between performance and new features for older devices. Slowing down old iPhones was done to preserve battery life, but users did not see it that way, which led to massive backlash. Sure, Apple, should've communicated its reasoning and decision making before (which they have now done), but the casual user just doesn't care enough about the nuances here.

I think slow phone is more planned-obsolencence-y compared to "no OS upgrade, but phone works just as it did since the last supported upgrade - features and speed". I don't think the majority of users would know there is a new version out if they didn't get the upgrade nag-screens.

That sounds good in principle, except that 3rd party apps would also be frozen and start to fail.

> 3rd party apps would also be frozen and start to fail.

Only if they use online services that are taken down; not all apps are like this. Even if they are, it is still a better scenario than throwing the entire phone away because it's unusably slow or unsafe to use due to lack of security fixes.

A customer won't see it that way. Their apps will nag them to update and they won't be able to.

Just as fast, sure, but holy cow the number of WebKit exploits in iOS will leave myself scared shitless that you’re running that thing on the internet still...

>Written on my iPhone 5S (2013) running iOS 8 (2014) - it is still a phone and is just as fast as it was in 2014. Is that planned obsolescence?

No. But it probably has known unpatched vulnerabilities, and you're putting yourself and people around you in danger of a security hack.

Yup.

It's unfortunate that I can't get security patches. I'm not willing to update to iOS 76 to get the fixes because I don't want to run code that is tested on an A12 on my 5 year old phone, and I don't want to buy a new phone because this one works fine and isn't larger than my hands.

FWIW I installed iOS 12 on my mom's iPhone 5s a few weeks ago, and I can tell you it ran much faster than iOS 10 or 11 on that device.

2 seems the reasonable choice. If old hardware can't support a new feature, don't add it. I don't think customers will have trouble understanding "Your phone is old and it can't handle this new thing". That isn't difficult to understand. For older phones just provide necessary security updates for the older version of the OS.

Apple already doesn't turn on new features on old phones that "can't support" them.

The new features being talked about are ones that make an old phone slower. I owned an iPhone 5s and updated to the latest iOS version for 4 years and never noticed any slowness. So, basically, (2) means I would miss out on new features because other people complain about slowness that I don't notice.

Yeah, some people seem to have much higher expectations of speed than I do. Having lived through dial up and actual hard disks (rather than SSDs), I really don't notice any slowness in my iPhone SE. I just notice all the awesome new features.

Where's (4) Don't write slow software? Modern phone's have better specs than 2000-era PCs with software that needs to do less; they shouldn't be having these performance problems in the first place

> Where's (4) Don't write slow software?

Says who, the government? We are talking about court-imposed fines here, not consumer wishes. Do we want a literal slow-software police?

> Do we want a literal slow-software police?

That obvious answer is no, but as I think back on how many times I've been frustrated by some poorly performing piece of shit application over the past year, it's starting to not seem so bad of a solution.

>Do we want a literal slow-software police?

Yes please!

> Do we want a literal slow-software police?

This court case shows that we kinda now already have one.

The phones from 5 years ago already had amazing processors with billions of cycles per second. They can probably do nearly everything.

But you know us software engineers: "make it faster till it runs wells on my test kit, then oh well ship it!"

I think this ruling is a nudge into a better way. Yes deplay updates for more than 2 years, but don't stop optimizing new stuff when the latest phone can run it pretty well - keep optimizing until devices from the last 5 years run it very well.

> This court case shows that we kinda now already have one.

That's patently wrong. The point of the fine was not that the phones were noticeably slower after the upgrade. The point of the fine was that the slowness was motivated as a way of inducing forced obsolescence. The manufacturers forced the updates with no rollback mechanism to render phones unusable and thus force their users to purchase replacements. If users could roll back the changes then the manufacturers wouldnt be fined.

Yes. Immediately. If it can be shown that a piece of software written today has major functionality which also was written into any software before 2000, and the modern software has twice as much or more delay in its functionality, the author should be publicly executed. Maybe a little extreme but hey it'll provide an incentive :)

I don't think anyone intentionally wants to write slow software. What would iOS and Android developers (both app and OS) stop doing or do differently?

Stop making everything animated or translucent?

I'm curious why you were downvoted. I have found animations do slow down my phone. When I press the home button, as the apps fly in, I can't scroll left/right until the animation is done playing. I have reduce animations turned on (accessibility settings, I think) as it makes my phone feel a lot faster.

Also, I remember the calculator app being unreliable due to animations: https://www.macrumors.com/2017/10/24/ios-11-calculator-anima...

Is this not a legitimate problem?

I know on Android there's an option in the developer menu to double the animation speed that can make an appreciable difference. Maybe something exists on the iPhone side too

Thanks for mentioning this. I've noticed the setting before but never played around with it until now. I didn't have any performance complaints (Pixel 2 XL), but disabling animations makes (nearly†) everything feel noticeably snappier.

† The google feed seems to have a consistent lag upon switching to it from the home screen (double tap on the multitasking button is normal) that the animations hide. It loads somewhere around 150ms slower than most applications, even taking longer than the first launch for unused apps.

Stop packaging mobile JS frameworks as native "apps".

Have you heard about the new fancy framework in JS ? It emulates x86 in css, then boots Windows Vista in to execute a Visual Basic Application in containing a webView with mobile react !

It's a bit slow on loading, but after that it's really fast !

Absolutely this. I am dumbfounded that some mobile applications are so painful to use when they really don't need to be. Take Snapchat for example. Why is Facebook and Instagram multitudes faster than Snapchat at taking pictures when that's 10% of FB or IG and 95% of SC? What is SC doing that requires so many resources?

And every Reddit client ever, or Reddit in general. Shameless amount of cookies and a horribly bloated UI just to do the same thing they were doing 10 years ago with double the client resource usage.

> And every Reddit client ever, or Reddit in general. Shameless amount of cookies and a horribly bloated UI just to do the same thing they were doing 10 years ago with double the client resource usage.

The amazing/sad thing is reddit with other clients can be blazingly fast. I just tried redreader from the f-droid store (android open source store) and using it feels like the content is loaded locally.

They presumably make it slow ("conservative"!) to limit server resources, to limit costs.

Give Apollo a go, runs great on my iPhone SE.

The whole point of this patch was to deal with old (~7 W h) batteries. Your 2000-era PC was usually plugged directly into the power grid, or if it was a laptop, it had a 50 W h battery and went a few hours away from grid power at most.

It's a lot easier to "not write slow software" if you're not constrained by power limitations. Apple went to all the trouble of switching CPU architectures just to get on a better side of that problem.

2 is a good option, give users security fixes, if you promised to deliver new features to them then do so but make sure you did not made the OS and core apps worse.

Or just make high-powered features optional and don't fuck up your core OS. That seems like the obvious choice. I've had several experiences with OS "upgrades" making a system much slower even though I wasn't using any of the new features. That's poor design if not actually malicious, and it doesn't have to be that way.

4) Don’t ship updates to phones in the EU

And maybe get sued because of unfixed vulnerabilities.

Then ship security updates independently of bells and whistles updates that make old phones dog slow?

Then you get something like spectre and the fix is making things slower.

There are thousands of use-after-free (or generally more pedestrian) bugs for each spectre-like bug. Let's not make perfect the enemy of good.

Then you kindly ask Intel to reimburse the costs. Security > performance.

4) notify users that updates may slow down their model of phone.

(4) you open everything and let the community keep old devices significant numbers of people still use current and secure.

I don't think the problem you're addressing (new features being slow) is the problem the regulator had.

I believe the problem was that the updates made existing functionality slower. e.g. the same Facebook app you had before the update scrolls slower afterwards or the old launcher is replaced with one that's more jittery, without letting you use the old one.

Adding new features is fine, just don't break the old ones.

Except if part of the new features is a rewrite of display output that takes advantage of new hardware features / assumptions in ram characteristics, then that can be both a new feature and slow down older phones at the same time.

Given the actual changes apple+samsung made it's hard to argue this was in any way necessary - they chose to prioritize making the old phones look more like the new ones, which makes sense from their point of view because they want to encourage upgrades - so maximizing (visual) consistency at the cost of sluggishness is much better than maximizing quality for old phones at the cost of making phone differences more apparent.

And in any case: they're free not to offer feature upgrades to old phones; I just doubt that's a wise choice in the market. If they really can't provide a few updates and reasonable performance - good riddance, their successors surely will.

Then you don't include that in the build of the update for the older models it doesn't apply to, or you include both approaches to output and choose which is appropriate at runtime.

Could they have saved themselves by giving users an informed choice:

"this update will harm performance but improves security and gives access to new features; your choice" (imagined release note on update clickthrough)

Option (4): give the user a choice to opt into the new features.

A checkbox isn't that hard.

> A checkbox isn't that hard

A naive assumption. In fact, each checkbox exponentially increases the number of possible runtime environments you have to maintain and test.

Alternatively, a choice of whether to install the update.

"Installing this update will give you access to feature X, but will slow down your phone."

For comparison, my AT&T Samsung Galaxy S6 forces me to install updates, which I hate. Wasn't my choice to get this particular phone, wouldn't do it again.

Assuming some users choose to install, and some don't, doesn't change the fact that updates will become incredibly expensive support-wise. Exponetially expensive even.

This will mean: almost never updates at all.

Well, I have to admit: my mother will be happy about it.

Not if checking the obscure and hidden "Don't throttle me!" box comes with a warning that it is unsupported. Then you don't have to do any serious maintenance and testing.

Reply on news.ycombinator.com