> ARCore will run on millions of devices, starting today with the Pixel and Samsung’s S8, running 7.0 Nougat and above. We’re targeting 100 million devices at the end of the preview.
That's...not great? For comparison, ARKit on iOS is going to support 400 million devices at launch (very rough numbers: ARKit runs on any new iPhone Apple's released over the past two years - iPhone 6S/SE/7 - and they sell over 200 million a year). Hardware fragmentation is a tough problem to solve.
The exact same thing stuck out to me as well. Mostly because they had to start the article with this sentence:
> With more than two billion active devices, Android is the largest mobile platform in the world.
So they'll get it onto 5% of active devices, or one in every twenty. Not great, perhaps not even good.
I've not been a fan of android fragmentation for awhile now, and have been surprised that Google hasn't been able to do more to attack the issue. Even when Treble launched, I asked myself... is that all? Rock and hard place I guess.
I actively avoid android devices because of the issue. My last android device was a tablet in 2011.
The purpose of ARCore is for high-end Android devices to be competitive with the iPhone technologically, not to get on 2 billion devices, at least not now.
Therefore it's fine that they get only 5% to start. I assume it's going to turn into 10%, then 20% quickly. For some perspective the iPhone only has about 15% market share globally.
> The purpose of ARCore is for high-end Android devices to be competitive with the iPhone technologically, not to get on 2 billion devices, at least not now.
This article is explicitly spun to promote ARCore as "Android scale" and leads with a sentence pointing out that Android is on two billion devices.
You're right, their ambitions are much lower, but it's fair to point out their actual goals aren't anywhere near the image they are trying to project and they really shouldn't be bragging about it being "Android scale" when it's nothing of the sort.
I've been developing WebGL software for a client for the last few years and my client has had quite good success selling content based on the software and consumers are using the content. The vast majority of browsers fully support WebGL (1).
The biggest performance bottle-neck is the JS itself, specifically stock three.js is quite inefficient. However, that's just a matter of optimising three.js, or using WebGL directly if you need more fine-grained control. You can get a lot done in a browser these days.
I have seen impressive WebGL demos, but for some reason WebGL always makes the GPU sweat more than plain native code driving my fan at full speed, even though it offers less.
Maybe it is a consequence of using JS, I have never researched it.
Ditto.
To add more wood to the fire consider that those 2 billion Android active devices are not “Google’s Android” devices. In China, where the largest bulk of those 2 billion devices are, Google isn’t allowed. Which means that SDKs like this one that are part of or require the Play Services framework are nowhere to be found. I would put Google’s Android active device numbers somewhere around 300 million. Translation - ARCore will show up in 1 out of every 100 users six months downs the line when Samsung, LG and other partners decide to release it. In the meantime, around Sept. 29 of this year Apple will have 100 million+ ARKit ready devices - based on the 80%+ install rate for new iOS versions - when iOS 11 is released. Just saying.
The 2 billion is the number of devices with the Play Store. Android devices in China do not have Google Play Store or Services on them. So devices in China do not factor into this number.
As a user, how does fragmentation matter to you? It shouldn't affect you directly. The place where it matters is that you'll get less developers making software for you, but honestly, in the past 3 years, that hasn't really been an issue on Android. Except one or two big games like Mario, it's been fairly equal to iOS in terms of app support.
The developers support isn't just a minor quibble, especially when it comes to quality of life factors like browser versions (e.g. LG, HTC, Samsung, etc. have all shipped Chrome builds with subtle differences). This also applies to optimizations for CPUs, graphics hardware, etc. — the app may be available on both but it almost certainly runs better on iOS.
The other problem is that fragmentation increases the odds of a vendor walking away from a device since the per-model sales are divided more ways. Trying to pick a device which will get updates basically involves sales forecasting which wouldn't be necessary if the platform was open.
Fragmentation impacts which device you might want to buy. The biggest argument for Android is the wide choice of devices and features on those devices, but if your phone never gets an update, you're left behind. It's possible to get a really nice and inexpensive Android phone, but what's the risk of never getting an update?
You don't have to avoid Android devices, just buy Google ones. Most of the others are filled with crapware and bad customizations anyway, no idea why any savvy HN reader would want those.
Which is easy to say, if you’re on silicon valley wages.
But a student trying to develop for Android can’t afford devices that start (!) at 908.35 USD. (Price of the Pixel 5" with 32GB storage, in Germany, right now).
And Google just dropped support for my phone, the Nexus 5X (which I bought for 324 USD, on sale, just before they stopped selling it)
I never developed for iOS, because a platform that has in my home market (the EU) a marketshare similar to what Windows Phone 7 had in the US (below 15%) is useless. Especially when development starts upwards of $3000.
I’m comparing between developing on Google devices, and non-Google devices.
A new mac Mini with 8GB RAM is 819€ in Germany. A new iPhone 7 32GB is 759€. That's 1578€, or $1880. Add a one-year developer subscription and you're still below $2000. Your claim is just plain wrong.
The developer subscription isn’t just 99$. You need to pay with a credit card, the cheapest credit card available in Germany has an annual fee of 39.99€, unless you move > $1300 a year via a CC. (I don’t have a CC, neither do 90.1% of Germans)
So, you have $150 annual fee just for the developer account. Plus an additional massive cost every I time I need to replace the device.
The starting costs are significantly higher; maybe not 3k, maybe just 2k, but that’s the same anyway: Completely unaffordable.
> the cheapest credit card available in Germany has an annual fee of 39.99€, unless you move > $1300 a year via a CC
Only if you ignore the free credit cards that are independent of a current account from Advanzia, Santander Consumer, Barclaycard, Hanseatic, Payback, Targobank, ICS, and some I forgot.
Then there are also a lot of unconditionally free current accounts that offer a card (credit or debit) that can be used to pay that subscription, such as those from DKB, Comdirect, ING-Diba, Consorsbank, Wüstenrot, Norisbank, Fidor, N26, and so on.
Those offers are not hard to find, some of them are available for over a decade or more now. Even pretty much every consumer bank you can name in Germany has such a card on offer for less than 40€ yearly. That includes small cooperative banks. If yours does not you should consider switching (or opening a second account), competition is fierce.
Santander, Barclay, Hanseatic, Targo, DKB, Comdirect, ING-Diba, Consors, Wüstenrot, Noris, and N26 require that you have at least 600€ each month of movement via the bank account, or at least 1200€/year via the card.
I’ve tried getting a card from all of them. All of them refused, as I’d not use them enough.
> If yours does not
I’m a customer at 2 major national german banks (postbank and commerzbank), and a local one, none of them (or their subsidiaries, such as comdirect) offer this.
I just want a credit card, I’ll use it at most once a year, for at most 100 bucks. That is not available from any bank.
The only offer that exists even similar in cost is the old MyWirecard offer, and they closed my account due to not using it enough.
All of the ones I listed have at least one free offer without 600€ requirement, not sure you got that idea from. I am a customer at some of them without using the account at all and don't pay any fees. Some of the banks are quite picky and refuse customers for no apparent reason but none of them I know even asks how much money you plan to spend with it so that couldn't be a reason. If all of those banks have refused you there has to be something seriously wrong with your credit report, I would get a free § 34 report from Schufa and Creditreform to check if everything is correct if I were you.
Postbank has a Visa or MasterCard for 29€ yearly. Comdirect has a free Visa Debit card. I have it and have not used that account for over two years now apart from downloading the statements. Nothing happened so far and I haven't paid anything.
In your case the easiest solution would probably to order the British prepaid card from Revolut which is available for a one-time 8€ fee and I don't think they refuse anybody. Just note that they are not regulated as a bank but as electronic money so I wouldn't leave large sums of money for a long time on the card (but that's not something you want to do anyway).
> If all of those banks have refused you there has to be something seriously wrong with your credit report, I would get a free § 34 report from Schufa and Creditreform to check if everything is correct if I were you.
My credit report is perfect, and I’ve had no issues with that – but the problem is that I, as student, currently have zero income, and zero spending.
> Comdirect has a free Visa Debit card
Only in combination with the Girokonto, which is only provided if you have any monthly income. Same with Hanseatic. Check their AGB, or call them.
> Postbank has a Visa or MasterCard for 29€ yearly
Same story with this, I’m a customer at Postbank, have a Sparbuch in the 5 digits, but I can’t get a free Visa or MasterCard until I’ve > 670€/month income, and I won’t pay a single cent for a credit card.
> In your case the easiest solution would probably to order the British prepaid card from Revolut which is available for a one-time 8€ fee and I don't think they refuse anybody. Just note that they are not regulated as a bank but as electronic money so I wouldn't leave large sums of money for a long time on the card (but that's not something you want to do anyway).
I’ve had one with them, too, they closed the account because it wasn’t used often enough.
Oh quit the nonsense! I'm paying nowhere near 40€ for my credit card in Germany (and have no minimum transfer volume requirement). Also, ⅓ of Germans have a credit card: https://www.bundesbank.de/Redaktion/DE/Pressemitteilungen/BB..., and it's not like this is the only occasion where one might be useful.
Not to mention that all that gives you the ability to, you know, sell apps on the App Store and recoup your expenses. $2000 is a very reasonable entry price for something you could be making your living with. It's expensive for a hobby, granted, but it's not like the stuff you buy is single-purpose. And it keeps resale value amazingly well.
You can really stop it with the made-up numbers now. This is ridiculous.
I bought a Nexus 5 (that was supposed to give you the Google experience) after my iphone 4 and was the worst phone that I ever had.
I barely managed to not throw it against the wall for several times.
After less than 2 years I sold it and bought an iPhone 6s that I still have and I use happily without any "phonicidal" thoughts.
I wasn't comparing Google Phones to iPhones, but to the rest of the fragmented Android market. iPhones don't come with bloatware and crapware either. But what issues did your Nexus 5 give you?
Not the parent comment, but my Nexus 5 power button got stuck and bootlooped my phone.
Replaced the motherboard myself, a year later, same issue.
Bummed about it because otherwise it's a good phone at a great price. Only thing I complain about other than that is how terrible the camera is, but for a cheap phone that's a few years old, it's unfair to compare to today's devices.
That doesn't make it very appealing for app developers, though, if the size of their base is dramatically limited to just people with both the latest and the greatest devices.
I'm not sure what's unappealing about targeting users that can afford to always have the latest and greatest. That's effectively exactly what you're doing if you choose to Target iOS over Android.
Considering Android apps make 15% of the money iOS apps make, perhaps it is for the best to only target devices that happen to be owned by people who spend money.
If android apps make 15% of what iPhone apps do despite having a much larger market share, what would be the point of targeting a smaller fraction of that? Even if those users are all willing to pay you're probably not going to get higher than that 15%.
The "high-end" at market on iOS is bad enough, would you want to try and fight that battle on android?
Today's "a few" is tomorrow's "everyone", this has been proven before (the original iPhone, car technology, etc.). You take the time to perfect what you build with a smaller audience and more powerful technology, and eventually there will be more people to sell it to.
The question is if you're trying to earn your living doing that can you wait long enough for "a few" to become "everyone"? There are a lot of businesses that fail because they're too early.
I think it runs on every A9 (iPhone 6S) powered phone and tablet or newer. So more or less every iOS device sold since 2015/2016.
Yeah and fragmentation in Android versions is pretty bad. Also the quality of the underlying hardware might be not as consistent as in iOS devices. It relies on a lot of sensors to get it right, not just raw CPU power and a nice camera.
iOS has had a ton of effort put into its audio stack - it was an iPod replacement as well as a phone. Remember when the app was still called iPod?
iPod users wouldn't have taken the iPhone as seriously if Apple hadn't put in the work to make audio a first-class citizen. Android doesn't have that pedigree/baggage so it's never been as important.
The sad part is that they aren't serious to sort it out.
Treble could have been the solution, yet they are still allowing for OEM customizations and the OEM are the ones expected to keep pushing the updates, as discussed on the ADB Podcast.
Guess how many OEMs will choose to push Treble updates instead of selling new devices.
Software updates alone can't fix it as you can ship all sorts of hardware with Android. Apple has a known combination of hardware across its models. Even subtle differences in imaging sensors would be hard to account for.
It's mind-boggling how Windows Updates has been successful for decades now in the even more fragmented PC world, but Google gets a free pass because... reasons.
No software update in the world is going to help when there are thousands of different sensor configurations in use. Apple only has a handful and can test against all of them with ease.
Simple explanations for complex issues (e.g. getting many vendors and layers of the vertical cooperating) are seldom of much use. Bringing up Microsoft -- they tried the whole fixed hardware thing with Windows Mobile cum Phone and failed catastrophically, always two steps behind -- is not convincing.
Android solved a very different problem than iOS. It has compromises and benefits (that iOS has enjoyed by proxy) in doing so. The OS that we know today is the combined knowledge of many OEM customizations. The hardware of a vicious battle between many vendors. The screens and designs the result of an endless tug of war. And again, Apple has enjoyed the fruits of these.
But saying "If only they..." is seldom from a place of reason.
Apparently no one at Google ever read PC hardware specifications from Microsoft like "PC 97 Hardware Design Guide" or "Hardware Design Guide for Microsoft Windows 95", among a few other hardware guides from them.
Of course the Windows Phone attempted at it failed, because by the time Microsoft tried to apply the same rules as they did on the PC hardware, OEMs already had the Android wild west to play with.
I don't think you can judge Microsoft's approach on update management based on the failure of the platform. The platform failed because there's a monopolist in the space with an 85% market share and huge network effects regarding app development.
Microsoft hit update handling for Windows Mobile spot on. A 2014 Windows Phone today[1] is a safer bet if you value security than any Android on the market right now. May not run the apps you want, but it is going to have the latest security updates.
[1]My Lumia 929 currently is running Windows 14393.1593, released August 8th, 2017. I can safely expect to receive the next update on September 12th.
The platform failed because there's a monopolist in the space with an 85% market share and huge network effects regarding app development.
When Windows Phone 7 was released (this is ignoring that Windows CE was around for about a decade before, giving Microsoft a massive headstart), Android had about 18% of the market. Microsoft leveraged their enormous base of desktop developers, pulling them through various failed initiatives, to get them developing for Windows Phone. Android started effectively from nothing.
In the context of actual history your post is simply surreal.
In the context of the discussion, your post doesn't make any sense. We're talking about how Windows Mobile fixed updates... which happened multiple years after Windows Phone 7. Windows 10 Mobile is the first version that doesn't handle updates through OEMs the way Google does it. (Even the stricter hardware requirements didn't really set in until Windows Phone 8.)
As I understand it, the OEM modifications will be relegated to the vendor/odm partitions. They specifically hint at being able to push new system partitions directly in the future.
2 - Android 7 has currently about 13% acquisition after one year
3 - There is currently no obligation for OEMs to provide updates
4 - Given the failure of security updates, only being delivered to flagship devices, I believe Google will ship them directly when they actually do it.
Basically, one of the issues that's plagued OEMs in the past is the fact that each SoC vendor (e.g. Qualcomm, Samsung, ARM) would not only have a kernel for each SoC, but each OEM device using that SoC would be another branch on that SoC-specific repository. This would then be compounded by the OEMs having their own kernel repositories that pull from the device-specific branches of a SoC repository, and bug fixes from OEMs would have to trickle back upstream to the vendor repository before making them back to other device-specific kernels.
The new proposed model would see a single kernel soruce for a given SoC, and device-specific changes would come in the form of kernel modules, config overrides, and device tree overlays. They are also strongly recommending SoC changes to the kernel be upstreamed to mainline Linux, which would greatly ease porting to newer kernels in the future.
Recommendations are just that, advices that might be followed, or not.
It isn't the first time that Google does recommendations to OEMs.
They also did recommendations for security updates. An advice that OEMs only follow for the customers that paid the big bucks for the flagship devices.
So unless they actually provide requirements instead of recommendations, nothing will change for consumers.
So how come these devices account to "millions of devices" or "targeting 100 million devices"? AFAIK, Pixel devices and S8s combined would be much fewer than 100 million devices. Is this even the correct number?
Yeah, I don't think shipping for all Androids would have been realistic. AR requires pretty tight tuning for the available hardware stack and I guess Googlers just tuned it for their own phone and the most popular flagship for start.
Apple has significantly easier job here - especially since it doesn't need to get OEM buy-in for these kind of things :/
Does this mean that the lower end OEMs may not provide this on new devices at all? Or if they do it themselves that it may be a very low-quality calibration so the feature doesn't work well?
first, 100MM or even 400MM devices is no fucking joke...
but Q is: bandwidth to deliver drivel to the devices... not the apps - but how much bandwidth will ads consume over consumed content.
If I were a VC, I would be purchasing every-single-pipe-in-existence for the future.
We went from pipes-are-pope to content-is-king -- but the fact is that the control of information is in the plumbing ( at the high level ) -- not the holders of attention... they are just firewalls.
Facebook is a firewall.
Google is a firewall.
Reddit is a honeypot.
etc...
you get the analogy.
so...
Who owns the pipe.
ISPs were vilified... FFs the NSA sent the head of Qwest to prison for not spying... the scientologists refuted the NSA on carnivore (because they were already monitoring the networks for clams) -- the .gov went after the guy who detailed all fibers.. -- how many more examples would you like...
> ISPs were vilified... FFs the NSA sent the head of Qwest to prison for not spying... the scientologists refuted the NSA on carnivore (because they were already monitoring the networks for clams) -- the .gov went after the guy who detailed all fibers.. -- how many more examples would you like...
Can you explain your last paragraph a bit? It's a bit telegraphese and I'd like to know to what individual things you are referring. Thanks!
There is way too much to educate ppl in one post on the evolution to the above, but I can go back decades if I have time - and ppl have interest...
Anyway - the NSA was running all over and installing spying boxes... they got caught.
The CEO of Qwest was convicted on "tax fraud" or some such after he had the balls to stand up to the NSA and not install their shit.
Aside: Qwest communications was founded as a splinter from a railroad company who realized they had right-of-way easments along ALL their rail lines -- which was easy to lay fiber alongside, and then build a comms co on that backbone.
Earthlink, which was funded early by key hollywood scientologists, refused to install the boxes as well -- but because they were already spying on all their user traffic to hunt for "clam" people (clam people are those that made fun of scientology due to scientology's belief that humans are evolved from clams - among a great many other weird beliefs... anyway they werent prosecuted, presumably due to holywood big-shot money... etc...)
---
There was a student who was able to look up all the public-works filings for all municipal and federal digging plans/permits/etc and wrote some detailed report on the critical network infra of the US, thus revealing its weaknesses -- the report was nabbed as "national security" as "terrorists" might attack it.. (though they didnt really report on the fiber cuts which caused major upheaval and outages in silicon valley) -- but they shut that kids thesis (i think it was a thesis) down...
The revolving door with the NSA and silicon valley is wide open, obvious and pretty big...
if you want more info. email me... revealing too much on HN is stupid.
Could be a camera or sensor quality/feature issue. Or could be they just haven't built a lens map for it yet, so they don't have fov and distortion metrics.
That's...not great? For comparison, ARKit on iOS is going to support 400 million devices at launch (very rough numbers: ARKit runs on any new iPhone Apple's released over the past two years - iPhone 6S/SE/7 - and they sell over 200 million a year). Hardware fragmentation is a tough problem to solve.