AOSP Maintainer Quits

In a post on Google+, Jean-Baptiste Quéru, long-time maintainer of the Android Open Source Project, has said he'll no longer be working on it. "There's no point being the maintainer of an Operating System that can't boot to the home screen on its flagship device for lack of GPU support, especially when I'm getting the blame for something that I don't have authority to fix myself and that I had anticipated and escalated more than 6 months ahead." Quéru is referring to the recently-released Nexus 7 revision, for which Google has not provided factory images of Android 4.3. This seems to be because GPU maker Qualcomm is refusing to release the blobs necessary to boot the device.

Comments

Google can fix it with a hammer.

[Anonymous Coward]

The best way to solve this problem is for Google to announce that they will not to use any parts that don't include open source drivers. The blobs will be released real quick.

Re:Google can fix it with a hammer.

[h4rr4r]

That is exactly what google should do.
If your drivers are not in the mainline kernel, your parts do not go into nexus devices.

Re:Google can fix it with a hammer.

[Anonymous Coward]

Why?

Why would google do that?

Google doesn't give a damm about open source really. They care about profits.

Customers don't give a damm about open source either. Just a tiny tiny % of geeks care.

They are not going to screw over profits and customers for a tiny tiny % of users... that's just stupid.

Re:Google can fix it with a hammer.

[Anonymous Coward]

You guys are idiots. Google is not your friend. They don't put much energy into the open source part of Android. All the business-relevant parts are proprietary closed source software. Google could not care less if you can't run a fully open source Android version on their hardware. They haven't quite killed the competition yet, but rest assured that Google will cut you off if they feel they can get away with it. The age of the PC compatibles is over. They're being replaced by terminals that access SaaS.

Re:Google can fix it with a hammer.

[tlhIngan]

Except Google will find itself without a Nexus device to sell. Especially since Google has started toning down their Nexus line and starting offering "Google Edition" phones which are stock Android phones.

Because you think companies like Samsung, HTC, LG, etc care that the drivers are open or not? They sign the NDAs and get access to partial source code they need to create their devices.

As for using obsolete fabs and such - it's still expensive. Masks still cost around $100,000 each, and you need 10 or more of them still for a modern chip, so a tapeout run still costs a couple of million dollars.

FPGAs can be used, but when I used them, the dev systems used FPGAs cost $30K each, and the entire system ran at 10MHz. Oh, and you needed 4 FPGAs to simulate a subset of the chip. (That said, if you have 10 hours or so, Android DOES boot...).

The big problem still is the 3D stuff - all highly patented - implementing an open core will basically violate piles of patents, including many dating all the way back to when companies like S3 existed.

Of course, you can run Android in pure 2D mode, as 2D graphics are mostly patent free, but performance stinks. At the very least, a plain old framebuffer with no hardware acceleration can be implemented using open and free drivers.

Re:Google can fix it with a hammer.

[spire3661]

People dont give a shit how a structurally sound a bridge is constructed either, only a tiny tiny % of its users do.. Just because only a few know enough to care doesnt change the argument. Very often it is the unpopular ideas that are correct.

Re:Google can fix it with a hammer.

[kllrnohj]

Then nobody would ship on Nexus and everyone else would carry on as normal. Google doesn't move many devices, they aren't a major player here.

Also open source drivers is not going to happen, but that's not even what's being asked for. Qualcomm isn't allowing Google to redistribute the *binary blobs*

Re:Google can fix it with a hammer.

[Pieroxy]

Wow... That was some analogy and some conclusion. Let's dive in!

People dont give a shit how a structurally sound a bridge is constructed either, only a tiny tiny % of its users do. Just because only a few know enough to care doesnt change the argument

Comparing a phone with a bridge is at best disingenuous. There are public health safety issues with bridges that are obvious to many I guess. The disadvantage for users is obvious. Phones? What disadvantage can you see for a random user by using a "closed" phone? Right. None.

Very often it is the unpopular ideas that are correct.

Well, now you have to define 'correct'. And you will quickly see that there are as many definitions as there are people trying to define it. Google doesn't care about your definition of correct, nor should it. And you don't care about theirs. Just, remember, don't throw around yours as if it is a universal version of 'correct'.

Re:Google can fix it with a hammer.

[spire3661]

There are public safety issues at play in a device you carry with you at all times,devices that will live in our walls, our appliances, everything. Open Source is the cement, the steel beams that will hold up future information society, dont be so dismissive of its true importance. How are we ever going have a galactic computer if we allow art to determine what can and cant be done on a computer? We should get it right and not sell out for movies.

Re:Google can fix it with a hammer.

[Anonymous Coward]

Comparing a phone with a bridge is at best disingenuous. There are public health safety issues with bridges that are obvious to many I guess.

Begging the question.

The disadvantage for users is obvious. Phones? What disadvantage can you see for a random user by using a "closed" phone? Right. None.

Overgeneralization.

Re:Google can fix it with a hammer.

[Anonymous Coward]

There are public safety issues at play in a device you carry with you at all times,devices that will live in our walls, our appliances, everything.

Ok then all mobile phone software should be closed source to prevent people from messing with something that can potentially harm them.

Re:Google can fix it with a hammer.

[mungewell]

People dont give a shit how a structurally sound a bridge is constructed either, only a tiny tiny % of its users do..

Yes, but it's those tiny tiny % who either choose to work at Google or support the community around these 'cool' devices. If those people walk away the devices are doomed to failure, regardless of the 'common man'.

Re:Google can fix it with a hammer.

[Anonymous Coward]

According to wikipedia there's a Nexus 7 with a tegra SoC and another with a QC SoC. I've read that nVidia is actually supplying an open source driver for tegra and Rob Clark is making great strides to get a working driver for adreno GPUs (for proper linux). Is that too different to get working on Android?

Re:Google can fix it with a hammer.

[hairyfeet]

Uhhh...and why EXACTLY would Google do that, especially when doing so would cripple their devices as NOBODY is making a top tier GPU that isn't patented up the ying yang and I would argue that even making one is impossible thanks to patents?

In this case frankly there is nothing anybody can do, there is NO top tier mobile GPU being produced that isn't patented and cross licensed up the ying yang, from texture compression on up is ALL patented 6 ways to Sunday and all of Google's money can't fix that.

Re:Google can fix it with a hammer.

[madprof]

Or they'll end up (eventually) with an OS like Apple's.

Re:Google can fix it with a hammer.

[Zalbik]

People dont give a shit how a structurally sound a bridge is constructed either, only a tiny tiny % of its users do

Bullshit. I don't know anyone who doesn't care whether a bridge is structurally sound. I know a number of users that assume the bridges they drive over are structurally sound however.

And WTF does this have to do with anything anyway?
Companies that build bridges care whether those bridges are structurally sound, as they can face litigation and huge damages if the are not.

I'm guessing Google doesn't care whether they use open-source as they don't see any cost-benefit to doing so.

Re:Google can fix it with a hammer.

[gl4ss]

I think Android needs Qualcomm more than the other way around.

how so? there's other chip design sources. they went with qualcomm because it's 'muuuuuuuuurican. I kid you not.

qualcomm is consistently dicky, with dicks for leadership so why bother with them? all they're good at is lobbying their products and consistently failing to provide sw side in a fashion companies and users would like. does their hw even do what it claims to do in hw?

Re:Google can fix it with a hammer.

[Anonymous Coward]

Uhh... that probably won't work. For most of its life, the Android kernel wasn't mainline

Re:Google can fix it with a hammer.

[dgatwood]

Alternatively, all mobile phone software should be open source because it makes third-party security audits possible.

Re:Google can fix it with a hammer.

[MeBadMagic]

You sir, just FAILED the "Tom Joad Test".

http://freedominourtime.blogspot.com/2009/08/plague-of-punitive-populism.html

Myself, I would say:

Ok, then all mobile phone software should be OPEN source to prevent people from messing with something that can potentially harm them. If I have the source, I can take responsibility for checking potentially harmful code as apposed to giving responsibility for my safety to somebody else.

Open Source == Accountability

Closed Source == Power

Power corrupts. Absolute power corrupts absolutely.

B-)

Re:Google can fix it with a hammer.

[Anonymous Coward]

Alternatively, all mobile phone software should be open source because it makes third-party security audits possible.

Third party security audits don't require open source.

Re:Google can fix it with a hammer.

[TechyImmigrant]

Then they should use intel parts.

Re:Google can fix it with a hammer.

[Anonymous Coward]

Open in all the wrong places!

Re:Google can fix it with a hammer.

[Nerdfest]

Patents do not preclude open source, and really, they work with it quite nicely. To get a patent, you need to publish your ideas rather than rely on trade secrets. Really, copyright law applies more to the source code, but it is still protected.

Re:Google can fix it with a hammer.

[Anonymous Coward]

Not really, Google would be severely limited in their part selection because of that announcement and be at a disadvantage to other electronic makers who will take cheaper better and faster closed source parts.

Re:Google can fix it with a hammer.

[ultranova]

Customers don't give a damm about open source either. Just a tiny tiny % of geeks care.

Customers do give a damn about the availability of apps, though. You can get Angry Birds on any platform, thus it's not a selling point; but the presence of niche applications is, and those are too low-profit for companies to bother. That leaves geeks as suppliers, and as you said: a % of geeks care about open source.

Unlike desktop, mobile actually has real competition, so can Google afford to give up an advantage just because a primadonna chip maker feels like acting up?

Re:Google can fix it with a hammer.

[Anonymous Coward]

Third party security audits don't require open source.

OK, I'll have my friend hire me to perform a security audit of the code.

Oh, right, you mean third party security audits don't require open source when they're performed by specific third parties hand-picked by the audited company.

Re:Google can fix it with a hammer.

[tlhIngan]

That is exactly what google should do.
If your drivers are not in the mainline kernel, your parts do not go into nexus devices.

Two results:

1) Nexus devices run like crap because open drivers suck and are unable to use hardware.

2) Nexus are the most open devices, but damn does battery life suck.

The first is because you don't need the GPU, but Android performance stinks without acceleration. And the stuff open-source has access too tends to be fairly limited. so performance is never as good as it can be.

The second is because there are two truly "open" SoC manufacturers out there. First is Freescale, though the GPUs are PowerVR and those are NDA'd to heck and back. The other is... Intel, who actually has a completely open graphics stack.

Perhaps that's what Google should do - work with Intel to put the x86 into the next Nexus device. Then you have a completely open stack, save the phone firmware (not a problem for tablets). Of course, I don't think Intel's ARM emulator is open, so we can toss that out - but it'll encourage app devs to build for x86.... win-win-win!

I'm sure you can live with phones and tablets that get half the battery life for now, right? Because it's open!

Re:Google can fix it with a hammer.

[sjames]

The ability to track that user wherever they go even when the phone pretends to be off is never made public. The hole that allows the bad guys to do that as easily as the good guys (if there are any good guys left in that game) goes undetected. or worse, that little corner case that prevents 911 from working goes unnoticed until several people die.

perhaps it prevents the app that transmits recorded video to a safe place in real time from ever coming in to being.

Re:Google can fix it with a hammer.

[Anonymous Coward]

patent last for 20 years - so you want to use only hardware from 1993? I dont think that people will enjoy their "flagship" tablet with a i486dx2 and a Tseng ET4000 "gpu"...

Re:Google can fix it with a hammer.

[Anonymous Coward]

Just because they're hand-picked doesn't mean they're not as good. No company is going to hire an auditor that simply passes them, and no auditing company will risk their reputation by giving a company a passing grade for an insecure system.

iOS still has less vulnerabilities than Android, so I'm not sure how you can back up your raging fanboyism.

Re:Google can fix it with a hammer.

[MightyYar]

Well, I think he's right... whatever authority you use to force people to open up their code could just as easily be used to open up said code to audits without actually open sourcing it. You can use said authority to pick whatever third-party you would like.

Re:Google can fix it with a hammer.

[Anonymous Coward]

The ability to track that user wherever they go even when the phone pretends to be off is never made public.

Damn, that certainly sounds like Google's MO! Too bad many parts of Android are still closed-source by Google, so I guess we'll never know.

Re:Google can fix it with a hammer.

[_Sprocket_]

Customers don't give a damn if there is an API. Just a tiny tiny % of geeks care. But that tiny tiny % are developers. And customers like what developers create.

Customers don't know how the magic black boxes work. But they sure benefit from the magic created when those who do know can do their thing.

Also - for a company who "doesn't give a damn about open source", they sure do a lot of it.

Re:Google can fix it with a hammer.

[Anonymous Coward]

No company is going to hire an auditor that simply passes them, and no auditing company will risk their reputation by giving a company a passing grade for an insecure system.

If there was ever a perfect example of wilful stupidity needed, this is it.

Re:Google can fix it with a hammer.

[Anonymous Coward]

They might.

The problem in this instance is that they used an American manufacturer so they could market the device to Americans. If they chose components from regions that are less precious about locking down their hardware, they won't have the problem.

Re:Google can fix it with a hammer.

[Anonymous Coward]

Not the first time Qualcomm has crippled a device due to drivers or licenses or some other crap. Probably not the last time either.

Re:Google can fix it with a hammer.

[mjwx]

People dont give a shit how a structurally sound a bridge is constructed either, only a tiny tiny % of its users do.. Just because only a few know enough to care doesnt change the argument. Very often it is the unpopular ideas that are correct.

This.

People dont give a shit how safe a bridge is until it collapses, then they care a lot. People dont give a shit until it affects them, but by then it's too late.

People are dicks like that.

So it really is up to that tiny, tiny % of geeks or structural engineers to effect change.

Re:Google can fix it with a hammer.

[erroneus]

Google is a for-profit enterprise but it is a geek and techy operated enterprise and it would be stupid to presume none of them influence decisions made by the company. And anything that threatens Google's geek and techy fan base is a serious threat to Google.

If Qualcomm doesn't come around quickly, it will be the last time Qualcomm and Google are even so cozy in the future.

Re:Google can fix it with a hammer.

[Antique Geekmeister]

Given the existence of the Google Summer of Code projects, the ongoing publication by Google of Java patches, and the contribution of Google employees to fascinating projects on github.com, quite a large percentage of Google employees both use and publish open source and freeware. Numerous business partners and collaborators work with it extensively, especially when they see me publishing my patches or updated code and see that they benefit from my ongoing involvement. And they are willing to pay my company more because our projects are available, as source, so that work can be evolved or continued even if one of our developers changes employment.

I've certainly helped engineers try to reverse engineer software without source 10 years later, and it is _barbaric_. The last time, I fortunately found that the developer had actually cannibalized software I'd written decades ago to build the application. It was a reminder of why a GPL can be so much more powerful than an Apache or BSD license: the developer had not chosen to publish their modifications to their clients, for various legal and workflow reasons.

Re:Google can fix it with a hammer.

[hairyfeet]

Sigh...interesting? Really? Does nobody here know how this stuff works? okay let me break it down, you see NOBODY makes a GPU by themselves, NOBODY, its ALL cross licensed, again going back to S3 texture compression and going up through the modern like tesselation, its ALL controlled by one of a half dozen or so GPU makers and they ALL require an NDA.

NOW do you understand why those patents preclude FOSS? Because if I own a patent on X and the ONLY WAY I will let you use X is to sign an NDA you can't then FOSS it can you? Even the two players that are FOSS friendly, AMD and Intel, are NOT FOSS friendly to each other, fo example there are parts of the AMD GPU that they can't give the FOSS guys because Intel owns the rights and won't give them permission.

But guys like me tried to warn everybody, that unlike X86 where the CPU can do the heavy lifting you HAVE to have a decent GPU to share the load on mobile and its all cross licensed and patented and proprietary as hell, did anybody listen? Nope, so enjoy your future, a future of appstores and code about as useful as the code supplied with tiVos because ALL that shit is proprietary as hell.

Re:Google can fix it with a hammer.

[BigFootApe]

Atom chips with their completely open PowerVR graphics chips?

Oh, wait...

Re:Google can fix it with a hammer.

[Anonymous Coward]

Ummm how can Qualcomm tell Google what to do, didn't Google buy them?

Re:Google can fix it with a hammer.

[Anonymous Coward]

I don't think you're thinking of patents. The whole point in patents is that everything that is patented is made public in the patent document. There's no point in saying "I'll only let you use my patent if you sign an NDA" because licensing a patent doesn't actually give you anything secret, it just means they won't sue you for using the information that's public. NDAs come in when you want to use some funky GPU chip and want to know how to actually write code for the damn thing. The registers and layout and design and stuff are secret and you have to sign NDAs to get the documentation. This precludes OSS, but is not patents.

Re:Google can fix it with a hammer.

[Darinbob]

Google should do that. However they won't because they're Google.

Re:Google can fix it with a hammer.

[gunzy83]

Just because they're hand-picked doesn't mean they're not as good. No company is going to hire an auditor that simply passes them, and no auditing company will risk their reputation by giving a company a passing grade for an insecure system.

iOS still has less vulnerabilities than Android, so I'm not sure how you can back up your raging fanboyism.

Please link your evidence of less vulnerabilities for iOS or you are just showing your raging fanboyism.

Re:Google can fix it with a hammer.

[dudpixel]

iOS still has less known vulnerabilities than Android, so I'm not sure how you can back up your raging fanboyism.

FTFY. The point may still be just as valid.

With respect to the argument between closed vs open source, my clarification is entirely relevant.

Re:Google can fix it with a hammer.

[Anonymous Coward]

where the hell are you getting $30k FPGAs that can only run at 10MHz?

Re:Google can fix it with a hammer.

[PNutts]

Just because they're hand-picked doesn't mean they're not as good. No company is going to hire an auditor that simply passes them, and no auditing company will risk their reputation by giving a company a passing grade for an insecure system.

iOS still has less vulnerabilities than Android, so I'm not sure how you can back up your raging fanboyism.

Please link your evidence of less vulnerabilities for iOS or you are just showing your raging fanboyism.

Here you go: Malware infestation running amok on Android

And I'm pretty sure the fact that it's a pro-Apple / iOS website is a coincidence.

Re:Google can fix it with a hammer.

[Tyr07]

Public safety issues? Your post is a little confusing, are you saying that closed source is for safety issues?
I agree with regulation and security measures like requiring users to specifically allow unsigned content to run, but uh...

When's the last time anyone complained about an open source hammer?

It sounds silly, but imagine if they could ensure you only hammered approved nails?
Part of having freedom is that flip side, where people are free to make poor choices. For some, it might be a bad idea using software, for otherwise, it's a benefit.

Have to let people make that call for themselves, part of living, like everything else.

Re:Google can fix it with a hammer.

[Anonymous Coward]

iOS 7 - what a pathetic response to Android. You'll fall further behind when 5.0 is released in 2 months.

Re:Google can fix it with a hammer.

[smash]

Go for it. I'm sure the tiny precent of android devices google sell will just get tinier. Samsung will quite happily fill the gap Google leave with any supplier they give that ultimatum to.

Re:Google can fix it with a hammer.

[Anonymous Coward]

So, this is not about source code it's about redistributing the binary blobs - which are still obscured due to code.

Re: Google can fix it with a hammer.

[Anonymous Coward]

Can everyone please stop using the word 'less' when they should be using the word 'fewer'. It's horrible!

Re:Google can fix it with a hammer.

[Anonymous Coward]

It's also a well known fact (within certain circles at least) that iOS is laden with malware. The absolute worst of it cannot ever be uninstalled either, such as CarrierIQ, without illegally rooting your phone (which is itself an exploitation of a flaw, i.e. malware). Every iOS device in existence is keylogged. Every. Single. One. And you literally cannot fix it without breaking the law.

It's a grotesque feeling to see the things some people can acquire on literally anybody they want to that uses an iPhone. I have friends with certain powers in certain industries that do just that as a dayjob. Luckily I use a custom rom that has been vetted by the community on my Android device, and his company has absolutely zero information on me.

Re:Google can fix it with a hammer.

[Dog-Cow]

You are a retarded piece of shit. Anyone who travels over a bridge cares how sound it is. The exception would be children and the like who simply don't know enough to care.

Re:Google can fix it with a hammer.

[21mhz]

That leaves geeks as suppliers, and as you said: a % of geeks care about open source.

Do you have any evidence to confirm that open source geeks write any significant share of applications on the Play Store?

Re:Google can fix it with a hammer.

[Anonymous Coward]

This sounds very well, until you realize that we do in fact have open source 3d drivers. They are lagging behind due to missing specs, not due to patents.

Re:Google can fix it with a hammer.

[Anonymous Coward]

you are assuming that open source software is free software which it is not.

Re:Google can fix it with a hammer.

[Anonymous Coward]

No, completely wrong.

People *do* give a shit, but they trust and assume competence on the part of the engineers until given a reason not to.

It's only the structural engineers that would be interested enough to investigate it and spot the problem beforehand.

Really not the same thing as

People dont give a shit how safe a bridge is until it collapses, then they care a lot.

Re:Google can fix it with a hammer.

[Tastecicles]

if I'm expected to use a bridge I expect it to be structurally sound! THAT DOES NOT MEAN I need to know what kind of fucking cement it's made with!

Re:Google can fix it with a hammer.

[Tastecicles]

...I wonder why he posted that AC?

Re:Google can fix it with a hammer.

[Tastecicles]

let's think:

iOS: unknown quantity. Apple won't tell us what kernel they're using. They won't tell us how many vulnerabilities they're tracking.
Android: GNU/Linux kernel. Known quantity, we know pretty much exactly how many vulnerabilities there are, because each and every one is tracked and peer-reviewed on a regular basis.

So no, you cannot make any kind of comparison as to which is the more secure system, because security through obscurity is NO SECURITY AT ALL therefore ANDROID WINS.

Re:Google can fix it with a hammer.

[Tastecicles]

kudos for hitting that AC button right after you misused "you're".

Re:Google can fix it with a hammer.

[Tastecicles]

two words: Tacoma Narrows.

Re:Google can fix it with a hammer.

[Anonymous Coward]

Why did you have to call spire3661 "a retarded piece of shit"? So you disagree with their statement, how does that make them "a retarded piece of shit"?

Re:Google can fix it with a hammer.

[Anonymous Coward]

Actually, it is simpler than that. Android is designed to be insecure. Think about it: carriers do not want user control. Users want it and buy based partially on the idea they can get it. That means you have what amount to curated vulnerabilities in android devices. Hell, google basically joked about it when an old exploit rooted the glass explorer.

bottom line: android is vulnerable as a market position strategy... Why else would so many phones ship with known zero day exploits? Just remember, that hole is not something that goes away when you don't use it.

Re:Google can fix it with a hammer.

[TechyImmigrant]

No, the other ones.

Re:Google can fix it with a hammer.

[RMingin]

I support this, and would pay a price premium for Nexus to come to mean such an open device.

Open hardware means you can compile and install any version of AOSP/Android you like, and enables third party ROMs and modding.

Closed hardware means the Nexus 10 GPU memory leak took months to get addressed.

Closed hardware means that when Google declines to release recovery images of Android for your device, it becomes a doorstop at some later date.

Re:Google can fix it with a hammer.

[Anonymous Coward]

The OP took a literary shortcut, but the principle stands: people don't give a shit whether bolt #287 on the bridge they're using is grade 8 or grade 5 UNLESS they know that it makes the difference between a structurally sound bridge and an unsound one. People do not understand the implications of closed source any more than they understand the esoteric principles of engineering that could have a dramatic affect on their lives, so they don't care. They do care once it affects them negatively. Geeks care all the time just like the structural engineer cares about those bolts all the time.

Re:Google can fix it with a hammer.

[AC-x]

Here you go: Malware infestation running amok on Android

And what percentage of those are due to OS vulnerabilities rather than users simply installing malware disguised as apps downloaded from untrustworthy websites?

Or do you count being able to install arbitrary apps as an OS vulnerability?

Re:Google can fix it with a hammer.

[Anonymous Coward]

In this case frankly there is nothing anybody can do, there is NO top tier mobile GPU being produced that isn't patented and cross licensed up the ying yang, from texture compression on up is ALL patented 6 ways to Sunday and all of Google's money can't fix that.

Patents should only apply to their implementation (hardware)
Driver interfaces should not be patented or patentable (software)
Binary blobs should be on hardware and act like what they really are which is firmware.

Re:Google can fix it with a hammer.

[Anonymous Coward]

iOS: unknown quantity. Apple won't tell us what kernel they're using.

Bullshit. You don't know what you're talking about.

They won't tell us how many vulnerabilities they're tracking.

Google does?

Android: GNU/Linux kernel. Known quantity, we know pretty much exactly how many vulnerabilities there are, because each and every one is tracked and peer-reviewed on a regular basis.

My, you're naive.

So no, you cannot make any kind of comparison as to which is the more secure system, because security through obscurity is NO SECURITY AT ALL therefore ANDROID WINS.

And you think logical fallacies built on faulty premises are sound arguments, too!

Obscurity doesn't have anything to do with "isn't open source". It's about whether a system is so rare in the field that it's thought that it might be sorta pseudo-secure because nobody is bothering to find vulnerabilities. It might not actually be secure, which is where the "is no security at all" phrase comes from -- if you're counting on security through obscurity you're doing it wrong.

The problem with your premise: iOS is not obscure, you dunce!

Another faulty premise: even if the sole criterion anybody used to judge obscurity was "is or isn't open source", Apple publishes source for the OS X branch of Darwin. That'd be the same kernel used in iOS. They're not diverging the two either, so you can pretty much read iOS kernel source excepting device drivers.

The problem with your logic: Even if you accept that security through obscurity isn't real security, this DOES NOT ACTUALLY IMPLY that an obscure system is necessarily insecure! The only valid point to the old cliche is that it's not safe to assume that an obscure system is secure just because there aren't a lot of known attacks on it.

So howzabout you take your own advice and start to actually think? You're not doing a good job of it.

Re:Google can fix it with a hammer.

[Anonymous Coward]

It's not that the individual FPGAs can only run at 10 MHz. He's talking about prototyping ASICs using FPGAs. The very biggest FPGAs out there are typically made in very low volumes and the main application is ASIC prototyping, hence the $30K per unit pricetag (though the current size champ, the Xilinx XC7V2000, is cheaper than that right now).

The FPGAs can in principle run much faster than 10 MHz, but if you're attempting to do cycle-accurate simulation of an ASIC too big to fit into a single FPGA, that's when you run into problems. You must partition the logic across two or more FPGAs, but typically the number of interconnects required will vastly exceed the number of pins the FPGA actually has. (Because in the real ASIC version of the chip, these aren't external pins, they're internal nets, so you can have tens of thousands of them.) To make the prototyping system work, you have to multiplex FPGA pins, i.e. use each one as a serial transmitter or receiver to send the values of N ASIC wires over a single FPGA pin. You need to transmit N different values for each tick of the emulated ASIC clock, so the pin mux clock has to be at least N times faster.

I do this for a living. In one design I worked on, the amount of interconnect between partitioned blocks was high, and we were forced to use an 8:1 multiplexing ratio. The FPGA prototyping board we used couldn't handle more than about 100 MHz I/O speed using the original muxing approach. And although we were multiplexing 8:1, the actual clock ratio had to be 12:1 for some technical reasons I'm not going into here. So we ended up running the emulated ASIC clock at less than 10 MHz.

Later in the project, I put a lot of effort into an advanced mux design (self-calibrating muxes which compensated for variable trace delay, essentially a mini-SERDES per I/O pin, implemented in FPGA fabric). I was able to push I/O rates to about 240 MHz, and reduce the clock ratio to 8:1, meaning we could run ASIC logic as fast as 30 MHz. But that was somewhat aggressive. It's pretty normal for ASIC emulation of very large chips to be limited to less than 10 MHz. Particularly if you're relying on automated tools to do the partitioning and muxing. (we weren't, we were going so far as to reorganize the structure of the design a bit for the FPGA version, which gives you more opportunities to optimize FPGA performance.)

In another design I worked on, emulating a very different kind of chip, the amount of interconnect between blocks was small enough that no FPGA pin multiplexing was required. The challenge on that project was tweaking FPGA place and route to deliver high speeds -- logic designed for ASICs can be very difficult to run fast in an FPGA. We were able to hit ~150 MHz for some of the logic in the design, and about half that for the embedded ARM CPU.

Re:Google can fix it with a hammer.

[shutdown -p now]

Toning down? They've just released a next-gen Nexus tablet that's pretty much the best Android device on the market today.

Re:Google can fix it with a hammer.

[dfghjk]

Modded Insightful even though it is incredibly wrong and contrived. Nice job, /.

100% of users care about how structurally sound a bridge is. Open source Android is nothing like a bridge.

Re:Google can fix it with a hammer.

[ppanon]

Yeah? I think you need to refresh you memory about Arthur Andersen. Reading the Wikipedia article, it's sad how far the firm had strayed from the co-founder's principles.

Re:Google can fix it with a hammer.

[Eunuchswear]

Please link your evidence of less vulnerabilities for iOS or you are just showing your raging fanboyism.

Fewer, FFS

Re:Google can fix it with a hammer.

[Andy Dodd]

Except then, they just won't have a single SoC vendor willing to deal with them.

Re:Google can fix it with a hammer.

[smash]

How do you lock your house?

Let anyone forget...

[CAIMLAS]

Lest anyone forget, or for lack of never knowing, that this reason is likely only the tip of the iceberg.

It's not to discount it as a significant factor, but anyone who's quit from a position knows it's not just one thing, usually, there are several - lack of pay/low pay, poor work structure, poor work environment, demeaning personalities, etc.

Getting endless gripes and complaints about lack of support for something as popular and 'open' as the Nexus 7 when they've got no ability to fix the situation - but should, by Google's own marketing claims - has got to be pretty disheartening on its own, but I'm certain it's not the only thing.

Re:Let anyone forget...

[symbolset]

"The task is not possible" can be a pretty compelling argument for giving it up.

Re:Let anyone forget...

[jeffmflanagan]

It sounds like obnoxious people were blaming him for the problem, which is completely beyond his control. Walking away seems like a good decision. It's up to Google to fix this.

Re:Let anyone forget...

[Trepidity]

It's also worth noting that this was his full-time day job at Google (possibly more than full-time, if it went as this kind of project often goes). That's sometimes a good situation, because you're getting paid rather than putting in unpaid nights/weekends on the project. But sometime it can actually be worse, and more stressful, because it's your real job and you have to work on it daily. At least if you get burned out on a volunteer open source project, you can ignore it for a bit, step back from the mailing list and bug tracker for a little while things settle down, and then come back to it later with some fresh energy. But if it's your actual day job that is harder to do, unless you have an exceptionally flexible boss.

Re:Let anyone forget...

[scream at the sky]

It can also be one hell of a motivator to prove someone wrong.

Re:Let anyone forget...

[symbolset]

When someone else says "the task is not possible" that is a challenge. When you yourself find the task is not possible it is time to try something else.

Re:Let anyone forget...

[Anonymous Coward]

Anyone who's been following JBQ on Twitter or in his build group knows the guy has been overworked for years now, release after release. Burnout was inevitable, and I'm surprised and impressed he hung on this long.

Re:Let anyone forget...

[CAIMLAS]

When you're trying to do x (develop), and people keep asking you to do y (aka make a political/legal change), it's highly frustrating and disillusioning because you've got no time for x anymore. "You don't want me to develop? Fine, fuck you, do it yourself."

Re:Let anyone forget...

[symbolset]

I do the impossible every day. That is actually the scope of my "special projects" mission. But I am not disposed to tell you how I do it.

Sad, but not unexpected...

[Anonymous Coward]

I don't blame him for his frustrations. It's time for Google to step up their game a bit when it comes to guaranteeing proper support of AOSP. I put the blame squarely on Google for this mess, which is of their own creation.

Well, I guess that settles that

[NoNonAlphaCharsHere]

Up until this news, I was seriously considering buying one.

Re:Well, I guess that settles that

[h4rr4r]

Join the club. I guess I will have to stick with the original one a little longer.

I would pay extra for a device with all the drivers in the mainline kernel. So far that seems impossible in the tablet/phone area.

Re:Well, I guess that settles that

[wierd_w]

Agreed. I'd be willing to pay considerably more, and would consider it a major point of sale feature. It would mean I wouldn't be at the mercy of the device maker for firmware updates, at the very least. (A vanilla build of android from source is practically garanteed to work if all device drivers are in mainline kernel. If push came to shove, I could roll my own damned update.)

At this point I seriously wonder why there aren't people clamoring to produce fully open hardware SoC solutions for this market. Even lower powered devices on obsolete fab processes would be very desirable given the lockouts presented by the major players. A shiny toy is worthless if you can't actually use it.

The only thing I can come up with for why this hasn't happened is the employment of thermonuclear patent portfolios. Again, refusal to hold a patent bomb would further influence my purchase choice. Combined, i'd be willing to pay over 200$ more. (But I must have BOTH features. Mainline support, and peace of mind for not supporting the patent madness.)

Seriously. Show me a device that does both of those things, and can actually fit in a pocket, and I will buy it.

Re:Well, I guess that settles that

[NoNonAlphaCharsHere]

If "fit in a pocket" is in your criteria, then the Nexus 7 isn't for you; you want a phone form-factor. However, if you're looking for the perfect tablet form factor, then 7" is the right size to fit in your hand without danger of dropping it.

Re:Well, I guess that settles that

[uvajed_ekil]

Some of us don't have tiny girly pockets, in which case the Nexus 7 is indeed borderline pocket-sized.

Re:Well, I guess that settles that

[c 0xFF]

My Nexus 7 fits perfectly fine in my pockets, thank you. Even with a case.

Not that I'm going to roam around town 24x7 that way, of course. But I can slip it into my back pocket when I need to chase after my toddler. It also fits into my suit coat pocket -- a bit heavy, of course, but convenient for business purposes.

7" tablets are "pocket-able," but not "pocket-sized." And that's perfect for me.

Re:Well, I guess that settles that

[Anonymous Coward]

can confirm, i wear regular jeans and have an average size penis, it fits in my pocket fine, not that i would walk around all day with it in my pocket, but if i need my hands for some reason, it fits fine in the front pocket. So if your a hipster with skinny jeans and a tiny penis it might not fit, also if you're black and were regular jeans, and if your black and were skinny jeans, well i'm not sure what to say to that, maybe, I'm sorry?

Re:Well, I guess that settles that

[Anonymous Coward]

Pretty sure girly pockets are bigger than man pockets. Cuz guys don't need to carry around as much useless crap. :P

Re:Well, I guess that settles that

[wierd_w]

This is an easy mistake to make.

The reality is that pockets on women's clothing are often tiny, or outright missing. This is why women carry purses, and why they constantly ask male friends/lovers to hold things for them. This is because female clothing is tailored to accentuate the curvature of the female anatomy, and pockets distract from this effect. Take note the next time you go clothing shopping with the girlfriend. While she's in the changing room, just look casually at the clothing there in the lady's section. You will note that slack pants almost never have pockets of any kind, and female jeans tend to have oddly shaped or diminutive pockets. Skinny jeans especially.

Clothing designers (cough) design clothes (for skeletons) this way on purpose. They expect that the woman will have a matching handbag to match "her outfit." This is *why* women have 50 purses, and insist that they need to buy a new one every time they get a new outfit.

Seriously, you'd have trouble getting an ipod mini into the pockets on women's skinny jeans. Forget about getting a 7in tablet in there.

(Having 2 sisters sucked big time while growing up.)

Re:Well, I guess that settles that

[Anonymous Coward]

I did not realize my old navy cargo shorts were girly pants, but ok.

Re:Well, I guess that settles that

[NatasRevol]

Can you sit down with it in your pocket?

You know, without breaking it, or your penis?

Re:Well, I guess that settles that

[MeBadMagic]

Ever tried to get in your girlfriends pants through the pockets? (Yes, I know this is Slashdot)

DENIED!

lol

B-)

Re:Well, I guess that settles that

[Anonymous Coward]

I did not realize my old navy cargo shorts were girly pants, but ok.

These guys must be wearing trench coats or something. Come on! you put the nexus in your pocket? Can you bend or even crouch with that behemoth in there?
You must be kidding me.

Re:Well, I guess that settles that

[GoogleShill]

Old Navy cargo shorts are the worst! The normal side pockets are too small to hold anything without the risk of them falling out. My phone and keys are always ending up on the ground unless I put them in the lower "cargo" pockets, and those are just a pain, and without buttoning them down stuff falls out of them too!

Re:Well, I guess that settles that

[MightyYar]

You must be a big dude... I'm 6'3", and in my pair of Dickies jeans (not even remotely "skinny" jeans), the tablet sticks about 1/3 of the way out of the back pockets and about 1/5 of the way out of the front. In the front pocket, it is a very tight fit. In either event, there is not going to be much sitting! I've definitely "slipped it in my back pocket when I need to chase after my toddler", but to call it pocket-able is a bit much.

Re:Well, I guess that settles that

[BrokenHalo]

Ever tried to get in your girlfriends pants through the pockets?

You're supposed to RTFM.

$ man woman
No manual entry for woman

Re:Well, I guess that settles that

[Anonymous Coward]

Parachute pants.

Re:Well, I guess that settles that

[Anonymous Coward]

I'm about 5.5 ft. None of my pants (I don't wear tight jeans, or generally jeans) had pockets that the Nexus 7 wouldn't fit in. I generally wear slacks, or shorts. I'm not particularly fat or skinny. (Probably a bit over weight, but not by too much.)

I had exactly one pair which had any problems, which would quite often hit the power button on the way in/out, as the opening was a bit small, and I didn't want to sit down. So 1 pair vs about 11 pairs.

Re:Well, I guess that settles that

[Anonymous Coward]

The "fashion" here in Europe now at least is skinny jeans for men too.

I can't really fathom it because last time I tried anything even approaching skintight it basically nearly involved the seams cutting my balls and penis off so I can only assume that men who do wear these trousers have no balls and small penises.

Re:Well, I guess that settles that

[Anonymous Coward]

Here's a free fashion tip. When wearing skinny jeans (as a man), you're not supposed to wear them all the way up. If so, they will indeed squeeze your balls. Instead, wear them about three quarters up your butt, and use a belt. That way you'll both look fashionable, and still have spermatozoa that's alive when skinny jeans go out of fashion! HTH!

Re:Well, I guess that settles that

[Tastecicles]

absent my cargo pants, which have velcro straps on the lower pockets (tip: lose the button, use velcro!), I wear a gilet. That thing goes with *everything* and I can carry a tripod (shameless plug for the greatest tripod on Earth!) in it as well!

Re:Well, I guess that settles that

[h4rr4r]

Here is a free fashion tip:
If you are male you should not be wearing skinny jeans.

Re:Well, I guess that settles that

[c 0xFF]

I'm 6' even. It sticks out of my pockets, too, but that's my point exactly.

You're not going to be sitting down with this in your pockets. But the 7" form factor means you have a place to put it when the need arises, unlike a 10" tablet.

BTW ... I think Apple missed this key feature with their iPad Mini. It's over a half inch wider than the Nexus 7, making it that much harder to pocket. Especially with any sort of case.

It's Qualcomm's decision to make

[Anonymous Coward]

From AOSP's own site:
"Android is about freedom and choice. The purpose of Android is promote openness in the mobile world, and we don't believe it's possible to predict or dictate all the uses to which people will want to put our software. So, while we encourage everyone to make devices that are open and modifiable, we don't believe it is our place to force them to do so."

So why cry so hard when Qualcomm went with one of the choices you gave them?

Re:It's Qualcomm's decision to make

[h4rr4r]

Sure, just like it is my choice to never buy a device with one of their SoCs.

Re:It's Qualcomm's decision to make

[Trepidity]

It'd be one thing if this was in a third-party android device; nobody is insisting that Google must require every Android device to have open drivers, too. But this is Google's flagship device that's supposed to show off their platform. If they really "encourage everyone to make devices that are open and modifiable", they could lead by example by making sure that's true of their own device!

Re:It's Qualcomm's decision to make

[JesseMcDonald]

It'd be one thing if this was in a third-party android device; nobody is insisting that Google must require every Android device to have open drivers, too.

No one is insisting that Google police all the other Android device manufacturers, no. However, anyone shipping an Android device with closed-source kernel drivers is not following the terms of the GPL, which means they have no license to redistribute the Linux kernel. Legally speaking, no matter what Google requires, it's a choice between open drivers and copyright infringement.

Re:It's Qualcomm's decision to make

[Anonymous Coward]

However, anyone shipping an Android device with closed-source kernel drivers is not following the terms of the GPL, which means they have no license to redistribute the Linux kernel.

Well then good thing the Linux kernel isn't licensed under the GPL. It's licensed under a modified GPL allowing for binary drivers.

Re:It's Qualcomm's decision to make

[spire3661]

I just hate how the entire open source-ness of google's stance revolves around commercial video. Its disgusting ot see how much of computing is being sold out for copyright. Abolishment is the only possible rational course from here on out.;

Re:It's Qualcomm's decision to make

[Anonymous Coward]

Yawn...what's this...someone complaining about kernel modules being proprietary again...here's an old analysis: http://linuxhelp.150m.com/politics/kmodsGPL.htm

Re:It's Qualcomm's decision to make

[phantomfive]

Well then good thing the Linux kernel isn't licensed under the GPL. It's licensed under a modified GPL allowing for binary drivers.

That's not true at all. The license is clarified to indicate that user-land programs aren't licensed under the GPL. The license modification is to make it GPL 2.0 only, not 2.0 or greater.

See here for more info. The only question is whether the driver is a derivative work or not, which is a complicated legal term which didn't have software in consideration when it was (poorly) defined.

Re:It's Qualcomm's decision to make

[JesseMcDonald]

Well then good thing the Linux kernel isn't licensed under the GPL. It's licensed under a modified GPL allowing for binary drivers.

Stop spreading misinformation. There is no exception for binary drivers. There is a clarification that the kernel copyright "does *not* cover user programs that use kernel services by normal system calls - this is merely considered normal use of the kernel, and does *not* fall under the heading of 'derived work'." User programs; not drivers. Otherwise it's stock GPLv2.

Here's the actual license so you can see for yourself.

A few companies like nVidia get around this by never distributing the drivers with the kernel. In nVidia's case, they use the same driver for Windows and Linux, so they can also argue that there is nothing Linux-specific about the part they're distributing. Even so, many see this as a grey area. The Android case is completely different, both because these are Linux-specific drivers and because they are being distributed with the Linux kernel on the same media as part of a complete operating system. This is just as much a violation of the license as distributing a closed-source program which depends on a GPL library.

Re:It's Qualcomm's decision to make

[Anonymous Coward]

Not true. It's straight GPLv2. The kernel devs just interpret the GPL in a way that allows for binary drivers in limited circumstances.

Re:It's Qualcomm's decision to make

[JesseMcDonald]

The only question is whether the driver is a derivative work or not, which is a complicated legal term which didn't have software in consideration when it was (poorly) defined.

That is a complicated question, but in this case the real question is simpler: is the combination of the kernel and the driver on a running system a derivative work of the kernel? The two are distributed together, which makes it no different than a closed-source program distributed with a GPL library dependency. (Think of the kernel as the "library"—it doesn't matter which part is larger.) If binary modules do not violate the GPL, then GPL and LGPL are effectively equivalent for libraries, which would be a significant change.

It is much less likely that something like the nVidia binary driver, which is carefully not distributed with the kernel, or even specific to Linux, would be considered a derivative work. However, as you say, that area of copyright is not clearly specified for software.

Re:It's Qualcomm's decision to make

[Raenex]

The Android case is completely different, both because these are Linux-specific drivers and because they are being distributed with the Linux kernel on the same media as part of a complete operating system. This is just as much a violation of the license as distributing a closed-source program which depends on a GPL library.

I agree, but who is going to call them on it? By that I mean what copyright owner of GPL'd code. I think the Linus pragmatism of not caring about GPL principles has dominated the entire culture around the Linux kernel, so as long as the kernel source is released, they turn a blind eye on blobs that get distributed with Android.

Re:It's Qualcomm's decision to make

[phantomfive]

I'm not sure the distribution technique matters, since it is Google doing the distribution, not Qualcomm.

Re:It's Qualcomm's decision to make

[Xest]

But what if it's a choice between leading by example or producing a device that's a viable competitor to the iPad Mini because alternate hardware options just aren't competitive?

There seems little point leading by example if there's no one to sell an uncompetitive device to and that's really the dilemma I guess because otherwise there's a danger that people will associate "open source" with "uncompetitive device" if the only open source devices on the market aren't competitive.

It's really the hardware manufacturer's faults for not supporting open source as a standard option for device manufacturers as it means the market for open source supporting hardware isn't as competitive and hence as efficient as the closed source hardware market due to less choice.

Linux... Free for everyone!

[Anonymous Coward]

... as long as you have our secret proprietary boot loader!

Re:Linux... Free for everyone!

[Anonymous Coward]

That's why you need GNU/Linux, for with free software you have freedom!

lol hey remember when they sold linux by the pound? lol

and so i says to the guy, i says, i says, "excuse me but is that your nose or is it the telecom industry?"

in soviet russia, the telecom industry's nose insults YOU!

Much Noise, No Change

[Archangel Michael]

Quitting AOSP doesn't solve the problem. It makes it worse.

What we can do, is start a campaign of "name and shame", that starts asking key questions of Qualcomm spokesdrones, why?

And here is the real bits that should concern people, I rather doubt there is anything all that special about the blobs of code needed, or even the underlying hardware. Further, given the Copyright and Patents that SHOULD be protecting the "intellectual property" of Qualcomm, there is NOT A SINGLE REASON to release the code.

Even if the Lawyers want to be involved, how about writing a waver for AOSP so they can include the blobs needed, or the APIs to code themselves what is needed (probably showing up the crappy programmers at Qualcomm) etc etc etc.

There are plenty of ways around this issue, but if Qualcomm won't play nice, then it is time to start playing hardball. Believe me, a very loud "name and shame" Campaign would work. Here is just a one suggestion.

1) Android Apps detect if the device is running a Qualcomm chip (of any kind) and simply puts a blurb up that says "Your device is using Qualcomm Chips. Qualcomm doesn't fully support Android Open Source Projects, so please consider as part of your next Android Phone/Device one that doesn't use Qualcomm chips. Thank you.

Re:Much Noise, No Change

[BreezeDM]

I was thinking of getting a new phone with a Qualcomm Snapdragon 800, but since I heard about this I am no longer interested. What chip makers/producers are friendly to AOSP?

Re:Much Noise, No Change

[Anonymous Coward]

Quitting AOSP doesn't solve the problem.

How does it not solve the problem? The problem is he's getting blamed for stuff out side of his control. He quits, no more blame, problem solved.

Re:Much Noise, No Change

[Anonymous Coward]

Quitting AOSP doesn't solve the problem.

How does it not solve the problem? The problem is he's getting blamed for stuff out side of his control. He quits, no more blame, problem solved.

Yep, and AOSP is free to get a developer who sees things their way, and the proprietary BS with phones rolls on. This isn't even about Nexus devices, since as a whole they receive AOSP-based releases very quickly. This becomes a "problem" when you want to compile Android 4.3 for an older device that will never see official support for it, and find that the drivers are precompiled with no source available, and you are stuck either writing your own, or using the older version of the OS.

Re:Much Noise, No Change

[Anonymous Coward]

and the proprietary BS with phones rolls on.

That doesn't sound like a problem to me.

Re:Much Noise, No Change

[gnoshi]

Texas Instruments seems good, based on their history with the Galaxy Nexus.
Samsung isn't necessarily friendly to AOSP, but they do make their own SoC (Exynos) so at least Samsung isn't hamstrung by another manufacturer when making drivers. They also have a pretty good history for long-term updates.
It is worth noting that Qualcomm also makes the SoC for the Nexus 4, which has been updated to 4.3. Still, they hung the HTC One S out to dry, and there is this example too.

Not to encourage crime or anything but it would be awesome if the drivers source for some Qualcomm devices was leaked. Or if Google (and other manufacturers) actually created adequate legal contracts with the hardware providers for their Nexus devices so this shit didn't happen.

Re:Much Noise, No Change

[Threni]

> Samsung isn't necessarily friendly to AOSP, but they do make their own SoC (Exynos) so at least
> Samsung isn't hamstrung by another manufacturer when making drivers. They also have a pretty
> good history for long-term updates.

You're fucking joking. Samsung made noises about "the community" but then did nothing to help Cyanogenmod, to the extent that pretty much all the developers gave up on porting CM to the Exynos versions of the S3. The Exynos is a case in point - Samsung refuse point blank in handing over any docs/source, which means the Exynos version of CM sucks compared to the Qualcomm version.

In terms of updates - it's now August 2013; Android 4.2 source was released last October, 4.3 recently, so guess which version of Android this "flagship" product is running? That's right - 4.1. They've promised 4.2 several times and each time they miss it. As soon as it turns up people are going to go "Huh? 4.2?".

Re:Much Noise, No Change

[phantomfive]

What we can do, is start a campaign of "name and shame", that starts asking key questions of Qualcomm spokesdrones, why?

That has worked so well with NVidia.......

Re:Much Noise, No Change

[Billly Gates]

> Samsung isn't necessarily friendly to AOSP, but they do make their own SoC (Exynos) so at least
> Samsung isn't hamstrung by another manufacturer when making drivers. They also have a pretty
> good history for long-term updates.

You're fucking joking. Samsung made noises about "the community" but then did nothing to help Cyanogenmod, to the extent that pretty much all the developers gave up on porting CM to the Exynos versions of the S3. The Exynos is a case in point - Samsung refuse point blank in handing over any docs/source, which means the Exynos version of CM sucks compared to the Qualcomm version.

In terms of updates - it's now August 2013; Android 4.2 source was released last October, 4.3 recently, so guess which version of Android this "flagship" product is running? That's right - 4.1. They've promised 4.2 several times and each time they miss it. As soon as it turns up people are going to go "Huh? 4.2?".

How do you know it was not Qualcomm who did this?

Remember the hoopla about getting Java GPL with a classpoint excpetion? Adobe were the assholes with some of its graphics rendering and fonts which made up a full .6% of the source code held the rest hostage.

Qualcomm and ARM have patents and part of the agreement is you have to make it closed source so competitors do not steal our ideas BS or whatever Samsung had to sign in order to make its own chip.

We do not know. But what I do know this problem is dramatic in the hardware world and why we can't have great drivers for things like ATI cards. ATI wants to help but they outsourced and contracted some of the work to others who are being assholes when it comes to copyright.

Why aren't Nvidia drivers opensourced? You can thank the MPAA and h.264 consortium for that! You want hardware accelerated video? You make em closed source etc.

Re:Much Noise, No Change

[Billly Gates]

Hollywood and the h.264 consortium is who blocked the release of opensource drivers for Nvidia.

Which is why I was one of the opponents of it being part of HTML 5 when so many slashdotters were going on and on how much better it was bla bla. Side effect is the patent holder can tell you how to run the company.

It would not surprise me if Qualcomm and ARM have agreements with others like Samsung to make them closed source if you want to use are brilliant ideas etc.

Re:Much Noise, No Change

[Anonymous Coward]

Maybe if Linus gives Qualcomm the middle finger, things will change.

Re:Much Noise, No Change

[Anonymous Coward]

Maybe if Linus gives Qualcomm the middle finger, things will change.

To BSD

Re:Much Noise, No Change

[Anonymous Coward]

Well, the position of getting shit for something that is out of your control is open, if you think that you can do a better job.

Re:Much Noise, No Change

[slew]

Texas Instruments seems good...

Except for the small problem that last year Texas instruments quit making SOCs for tablets and phones...

Re:Much Noise, No Change

[Anonymous Coward]

Yep, and AOSP is free to get a developer who sees things their way, and the proprietary BS with phones rolls on.

Or equally likely to get a developer who is more committed to actually doing something about opening up drivers.

Re:Much Noise, No Change

[exomondo]

There are plenty of ways around this issue, but if Qualcomm won't play nice, then it is time to start playing hardball. Believe me, a very loud "name and shame" Campaign would work.

Why bother? If openness is so advantageous then people will buy open devices for their advantages and companies will realise they need to fall in line, so better to spend your time actually doing something with the existing open devices to prove your point.

Instead of trying to bully and shame those corporations (which won't work anyway) how about showing them why what you want is so good, in the end 'it's open' isn't an argument or a feature, it's just a means to a tangible benefit that you cannot get on closed devices. So whatever that tangible benefit is, provide it on existing open devices and - assuming it's as good as you believe it to be - people will follow.

Re:Much Noise, No Change

[Anonymous Coward]

Name and shame? Oh no.. some PR guy gets hated on by 4 people bitching about blobs. No one gives a shit. The new Nexus 7 will still be the best selling android tablet and you know what? only a handful of people know about this issue and even less care. They have shit to do, like play with their new tablet.

Re:Much Noise, No Change

[Anonymous Coward]

While you're at the naming and shaming stage, you should also add Google to your list of companies. After all, the Android brand name is a trademark that Google owns, and Google could, at their sole discretion, choose to not allow this brand name to be placed on said product. By allowing their trademark to be prominently displayed on said product, Google has tacitly accepted the actions of Qualcomm.

Replicant

[Anonymous Coward]

Just a reminder that the Replicant project is trying to make a completely free and open source version of the Android software stack, including the parts that interface with the hardware.

Re:Replicant

[viperidaenz]

.... that works on hardly any hardware, because it lacks drivers.

been there, done that

[ClassicASP]

I've been in that whole situation myself. A developer anticipates and escalates problems many months in advance and is just ignored and dismissed of, but then when it all goes wrong as predicted, they still get blamed. Sucks, but it happens.

By the Way...

[Anonymous Coward]

The GPU in Intel's upcoming Baytrail tablet SoC already has 100% GPL mainline Linux drivers in at least the 3.10 kernel... just sayin'

Re:By the Way...

[Anonymous Coward]

... and a lot of apps in the Market rely on native ARM code.
Google never should have allowed native code into apps outside of their control until they had their LLVM implementation (renderscript) working.

Re:By the Way...

[tibit]

Presumably if it's not self-modifying code, you can just statically translate it to x86 or whatever on loading and be done.

Re:Qualcom?

[Anonymous Coward]

Interesting...so you assume that NSA's tracking software is part of the phone OS? I assumed that it would be easier to grab the traffic from the towers (or better yet, on the backbones that the towers tie in to. Either way, I think your best bet to not be monitored is to get rid of your phone(s) entirely. That's the direction that I'm going in...as soon as I have a job that doesn't require one, that is...

Andy Rubin's comment....

[Anonymous Coward]

I remember an Andy Rubin tweet bashing Apple about about open source complete with a one-liner to check out and build Android.
Apparently you can no longer do that.

It is time to wake up

[trifish]

and realize that your geeky wet dream of fully open source OS on mobile phones is just that -- a dream.

Samsung, practically the only true major player behind Android, sells more phones to the dumb sheep than Apple does. They don't care about geeks a bit.

Re:It is time to wake up

[Anonymous Coward]

The geeks are the dumb sheep.

Re:It is time to wake up

[Scutter]

You had me until you brought out the "dumb sheep" trope. You would be more effective in persuading people if you could leave out the hyperbole and tired, cliched insults.

Re:It is time to wake up

[Anonymous Coward]

mmm yes. Absolutely correct. It's all about the drones these days. Sheeple is soooo 2008.

Re:It is time to wake up

[phorm]

I believe that Samsung also includes the appropriate blobs necessary to boot their devices.

One odd thing...

[fuzzyfuzzyfungus]

I'm not surprised that Qualcomm are being dicks about driver source(though I would assume that they have some haha-nominally-GPL-compliant shim for interacting with the Linux kernel, like Nvidia does); but the lack of a factory image seems very weird indeed.

Do they somehow think that anybody who wants to steal their precious secrets (and has the resources to actually be a threat), is going to be stopped by the need to buy a $200 consumer electronics widget and crack it open? If the device is shipping, the driver binaries and firmware blobs are shipping with it, in millions of units. They aren't going to stay secret long against anybody who cares.

Re:One odd thing...

[Anonymous Coward]

This was ALWAYS the argument about VHS piracy back in the day with MacroVision. Anyone who had the equipment to make 1000's of VHS copies usually had the equipment that stripped off the macrovision protection as bad signals in the waveform. Where the home user can't piggyback another VCR if they tried.

Re:One odd thing...

[Microlith]

I would assume that they have some haha-nominally-GPL-compliant shim for interacting with the Linux kernel

They do, and they've tried pushing it upstream. It was rightfully rejected because without the closed source blob that Qualcomm controls tightly it's useless.

Re:One odd thing...

[onkelonkel]

If memory serves, back in the day Radio Shack used to sell a $20 "Video Image Stabilizer" that was marketed as a way to clean up "home videos", but was understood by everyone to be a quick easy macrovision filter.

Re:Qualcom?

[the_B0fh]

And you think those are open sourced...? Have you even tried to install a rom on one of them? How do you even find which rom to use...?

Just wondering...

[carou]

Is there an upper limit to the number of times we can sarcastically quote "Open Always Wins!" after news articles like this one, before it stops being funny?
I know we haven't reached it yet, I'm just asking for information.

Re:Just wondering...

[treeves]

I think it's related to GIMPS. No, not the open source image editing software; the Great Internet Mersenne Prime Search, i.e no limit.

Paging the old school

[Applekid]

I guess it's time for the old piracy groups to step up and steal the factory images that the production house is flashing onto those devices.

Re:Qualcom?

[Anonymous Coward]

Open source like SE Linux?

That the NSA created.

we are all to blame

[Anonymous Coward]

All of us that buy, or let relatives buy Android devices that use restrictive and closed blobs.

Customers and only customers have the power to make the change.

Every time we encourage someone else to buy a non-Android device make sure to publicise the fact on open social media like Farcebook or Twatter on the corporate tags and pages to raise awareness in the less technology literate and make some voices heard.

It is one thing for companies to see bottom-lines get affected, but they also need to understand why. Don't just boycott, let the world know why too!

Re:we are all to blame

[petermgreen]

And what alternatives can you suggest for sourcing phones and phone-like tablets that are more open.

Re:we are all to blame

[Anonymous Coward]

> And what alternatives can you suggest for sourcing phones and phone-like tablets that are more open.

This: http://wiki.openmoko.org/wiki/Main_Page

Google PRISM

[Anonymous Coward]

" I assumed that it would be easier to grab the traffic from the towers"

Its part of PRISM, to which Google signed up. So I want Android, but not the Google pack on top of it. The Google data is grabbed by the NSA, the OS, Android is fine, its a good open source OS, but sadly usually comes with that Google pack on top of it.

So I will be switching to an Android device without the Google extensions. There's plenty of choices there, lots of Korean tablets, Taiwanese tablets are Android but not Google. So I won't have problems. Qualcomm be damned.

Re:Google PRISM

[Nerdfest]

Stop spreading that FUD. Companies were *forced* to provide data, with the possible exception of Microsoft from what I've read. If you want to stay out of PRISM, don't use any US services, including your ISPs.

Re:Google PRISM

[Anonymous Coward]

It's not FUD at all. Your ISP can't track your (EDH) TLS conversations, but Google can track all of your email, web searches and pretty much any URL you hit in your browser thanks to google-analytics. Sure, the NSA may have forced Google to provide that data, but it doesn't make it any less important to switch to another provider that isn't on the NSA radar.

It's all pretty simple technology. If you don't understand it, you should turn in your geek card now.

Re:Google PRISM

[Anonymous Coward]

Or ISPs that have some of their traffic switched through the US. Or Europe. Or Japan. But then, the non-US ISPs that avoid North American/European hops are EXACTLY the links that the NSA is the most interested in. So, you're pretty much screwed no matter what.

Re:Google PRISM

[stiggle]

You need to expand it a bit.

If you want to stay out of PRISM don't use or contact any US services, or services from the Five Eyes (FVEY - USA, UK, Canada, Australia, New Zealand), or services provided by companies with US parents, or comms links routed through any of the FVEY allies.

Or, you can accept you're going to end up in PRISM somewhere and carry on using those companies and services feeding the data in, but with the knowledge that they will be passing on the info.

Re:Qualcom?

[Anonymous Coward]

As if the NSA hasn't worked out all the proper agreements with our buddies in Taiwan and South Korea.

Android is deprecated

[frinkster]

Read between the lines.

Queru is gone. Rubin is gone. The Chromecast, whose original and main purpose was to get Android devices connected to external displays, ran Android in prototype builds but was released with Chrome OS. Look who runs the Android group now... The head of the Chrome OS group, who is still the head of the Chrome OS group.

I'd give it no more than 2 years before the Nexus & Motorola products are released with Chrome OS and Android is 3rd-party device only with all Google services removed.

Face it, Google just isn't getting what they wanted out of the platform.

Re:Android is deprecated

[Applekid]

Face it, Google just isn't getting what they wanted out of the platform.

Or they got exactly what they wanted: market penetration. The majority of happy Android users will have no problem upgrading to a closed Chrome Mobile as long as they get keep their apps (which will then be emulated in an Android VM, a VM within a VM if you will). And Google dropping old, smelly, and open Android means they won't keep their apps on future Android devices.

If I could go back in time and tell myself 5 years younger that Google, not Microsoft, was going to lead the next wave of Embrace Extend Extinguish, I'd have laughed in my own face.

Re:Android is deprecated

[foma84]

Android is 3rd-party device only with all Google services removed
OH GOD, that would be sooo awesome. You actually got me exited with this thought.

Re:Android is deprecated

[Rich0]

Face it, Google just isn't getting what they wanted out of the platform.

Or they got exactly what they wanted: market penetration. The majority of happy Android users will have no problem upgrading to a closed Chrome Mobile as long as they get keep their apps (which will then be emulated in an Android VM, a VM within a VM if you will). And Google dropping old, smelly, and open Android means they won't keep their apps on future Android devices.

ChromeOS isn't closed - well, no more than Android is. ChromeOS is actually based on Gentoo, believe it or not, so if anything the foundation is even more open.

However, in general everything in ChromeOS is web-based, and web-based apps in general don't have touchscreen UIs. I'm not sure that we'll ever see full Android-ChromeOS convergence. If we do, the result will be a platform that is actually much closer to the traditional Linux distro.

Re:Android is deprecated

[saleenS281]

Face it, Google just isn't getting what they wanted out of the platform.

Or they got exactly what they wanted: market penetration. The majority of happy Android users will have no problem upgrading to a closed Chrome Mobile as long as they get keep their apps (which will then be emulated in an Android VM, a VM within a VM if you will). And Google dropping old, smelly, and open Android means they won't keep their apps on future Android devices.

ChromeOS isn't closed - well, no more than Android is. ChromeOS is actually based on Gentoo, believe it or not, so if anything the foundation is even more open.

However, in general everything in ChromeOS is web-based, and web-based apps in general don't have touchscreen UIs. I'm not sure that we'll ever see full Android-ChromeOS convergence. If we do, the result will be a platform that is actually much closer to the traditional Linux distro.

Wait, what? The chromebook pixel is fully touchscreen enabled, and ChromeOS fully supports touchscreen, including all of Google's native apps. Tell me this is some attempt at a bad joke?

Re:Android is deprecated

[Rich0]

ChromeOS isn't closed - well, no more than Android is. ChromeOS is actually based on Gentoo, believe it or not, so if anything the foundation is even more open.

However, in general everything in ChromeOS is web-based, and web-based apps in general don't have touchscreen UIs. I'm not sure that we'll ever see full Android-ChromeOS convergence. If we do, the result will be a platform that is actually much closer to the traditional Linux distro.

Wait, what? The chromebook pixel is fully touchscreen enabled, and ChromeOS fully supports touchscreen, including all of Google's native apps. Tell me this is some attempt at a bad joke?

I didn't say that ChromeOS wasn't touch-screen enabled, only that web-based apps in general don't have touchscreen UIs. Google's native apps supporting touchscreen was actually news to me, but when I'm browsing websites on Android from time to time I still run into sites that depend on mouseover for something, and that doesn't work with touchscreen (though it would work fine with the trackpad/ball if your device has one and the OS supplies a cursor).

I'm not really saying it can't be done, just that so far nobody really does it that way, perhaps a few Google apps aside.

Re:Android is deprecated

[Anonymous Coward]

Exactly his entire job was just PR, just like this closing embassies thing; it is all just PR. After all Google is an advertising company they understand PR. The billionaires of Google don't give a damn about a free and open operating system but some of the people working for them inside and outside the company do, not to mention the people who freely give them quality debugged software. This job allowed Google to spread warm fuzzy feelings in a much better way than any amount of advertising could do.

Re:Android is deprecated

[frinkster]

Or they got exactly what they wanted: market penetration. The majority of happy Android users will have no problem upgrading to a closed Chrome Mobile as long as they get keep their apps...

I don't think that closed vs open is what Google cares about. What they are seeing is that handset makers are replacing Google services with their own and that means that Google is paying for the OS development but not getting anything in return. Google thought that forcing the handset manufacturers to keep Google services on the phone in exchange for access to the very latest versions of Android would be more than enough - every iPhone user can't wait to get the latest version of iOS. But it turns out the the handset makers are doing perfectly fine selling phones with old versions of Android and the users don't seem to care.

The future Nexus (which probably wont survive) and Motorola devices will be running Chrome OS and will have no possible way to rid themselves of Google services. They will probably also be able to keep their apps in exactly the same manner as you describe. It's all the other handset makers that Google plans to kill. They have bet the farm on Android. In a few years, the smartphone landscape is going to be Apple with iOS, Nokia with WP, Motorola with Chrome OS, and... HTC/Samsung/others with a completely outdated Android that gets zero support from Google.

I'm not in the US, and don't want Google

[Anonymous Coward]

"Stop spreading that FUD. Companies were *forced* to provide data,...If you want to stay out of PRISM, don't use any US services, including your ISPs."

I'm not in the US, I have sympathy for companies forced to provide the data, but that's not my problem. My problem is they provided/provide the data to the NSA.

Since I like Android but not the Google pack on it (it tries to grab my data too often for Google and in the process grabs it for the NSA), I want to switch to an Android tablet without the Google pack. Since Qualcomm don't support open source base OS, Qualcomm will go as part of that. Well meh, their choice.

I suspect you think I'm an Microsoft PR person, but Skype was the first thing to be binned.

Re:I'm not in the US, and don't want Google

[BrokenHalo]

...I like Android but not the Google pack on it (it tries to grab my data too often for Google and in the process grabs it for the NSA)

Fair enough. In which case, you can broaden your choice a bit by picking any device you like the look of and flashing it with CyanogenMod.

Or alternatively, you could root any other phone and replace/modify the hosts file to exclude all Google servers, just like you can on any other Linux box.

But personally, I'm not going to lose any sleep over Google's relationship with the NSA (ignoring for the moment the fact that I don't live in the US). Sure, the NSA's spying is obnoxious and unacceptable for heaps of reasons, but there's nothing you can do about it. The NSA doesn't *need* Google to track you. It has access to your pipes. You could (I suppose) run your own VPN and rent a terminating server in some country that doesn't take part in the intelligence network (if you can find and trust one) if you're that paranoid, but even that isn't bombproof.

I think the best way to neutralise the NSA would be for *everybody* to drop keywords into *every single* email, SMS, forum post, and crapflood their system:
Jihad, kill infidel, bomb Satan America, martyr for holy Islam, 72 virgins...

You get the picture. Now excuse me, I seem to have a couple of guys in suits banging on the do

NO CARRIER

Re:Its not fragmentation!!!111one

[Anonymous Coward]

Just like Google, fandroids are all about censorship. Modding people down for expressing a different opinion is the Google way!

Re:android sucks anyway

[Anonymous Coward]

Dude, you can't express a dissenting opinion here. The fandroids have mod points!

Cellular data firmware

[tepples]

Then you have a completely open stack, save the phone firmware (not a problem for tablets).

Without cellular data firmware, which I assume to be as closed as phone firmware, how does a tablet connect to the Internet while its user is riding a bus? The only workaround I can think of is an external hotspot device such as a MiFi.

Re:Cellular data firmware

[MiG82au]

I push one button on the drop down menu of my phone to set up a wifi hotspot for my tablet.

Re:Cellular data firmware

[Anonymous Coward]

wooosh, etc

(In that case you're using the closed cellular data firmare of your phone instead of your tablet)

Fat binaries

[tepples]

a lot of apps in the Market rely on native ARM code

That's a problem only if A. these applications are written in assembly language (highly unlikely), or B. the publishers of major NDK applications have announced their refusal to build fat (ARM/x86) binaries.

Re:Fat binaries

[Anonymous Coward]

I don't know about other people but I always build NDK apps for x86 as well as ARM, because the x86 Android emulator runs an estimated 1,000,000,000 times faster than the ARM emulator, making development less of a headache.

Kindle Fire

[tepples]

Android is 3rd-party device only with all Google services removed

OH GOD, that would be sooo awesome. You actually got me exited with this thought.

Then perhaps you can "exit" the Google ecosystem by buying a Kindle Fire tablet.

The enemy is everywhere, and it is YOU.

[hackus]

Looking at the submissions lately for AOSP and finding out just how pissed off I am about recent events including this one, which just made my day.

Something to consider while you dine this evening:

1) Increasingly Handsets and anything that shows video is being locked out.
2) This post is just one example, but I can cite others if you can't google about the whole sickening GPU/DSP issues in the industry which just keep getting worse with everything that is LINUX.
3) The convergence in my mind, that it just so happens that governments are ape shit over knowing everything you do. Further, if I may point out, the huge contracts cellular providers are getting behind the scenes to make this happen from DARPA.

Which to me, makes me wonder if the idea of knowing exactly how the video and camera hardware work is something by design, is not something your local friendly cellular provider wants you to know.

Think of the hardware GPU/DSP on your phone as partitioned as sorta "that room" you never go into while working at your local friendly Time Warner NOC for example.

I mean, wouldn't it just be dandy if the DSP/GPU hardware is BLOB'ed and secret so that the NSA/CIA can turn it on any time, preferable in a manner other software on the phone knows nothing about.

Think about that next time you AOSP a compile and include those nice little BLOB's.

-Hack

Re:The enemy is everywhere, and it is YOU.

[drinkypoo]

Well, suddenly I feel vindicated for buying a rockchip-based USB stick. At least I'll have a path to Linux, as the OSS Mali driver is already good enough to run Q3, and there is already a Linux kernel for RK3188 (although I couldn't get it to build an actual kernel image, it did build objects...)

Re:Qualcom?

[erroneus]

I wouldn't say he assumes it. But we all KNOW it is possible. And that knowledge is bad enough. We all end up making decisions in life without knowing everything we need to know. But there are some things we do know. You know? At the moment, things which have been compromsed by the US government can't be fully trusted. It's just that simple.

Every thing not fully free and open sourced

[Anonymous Coward]

Should be put to death as a general principle.

We don't need this crap.

Proprietary = Toxic

Re:Its not fragmentation!!!111one

[Anonymous Coward]

My my, the Wintard holligans must be frothing at this news. They become so excited because their own platform is such a miserable pathetic failure.

Re:android sucks anyway

[Anonymous Coward]

Be sure not to get a laggy, crappy and crashy windows phone. Worst phone I've ever owned and their app marketplace is filled with junk and counterfeit apps.

Buy chinese!

[Anonymous Coward]

There's thousands of knockoff chinese devices out there. And pretty much all of them are running on RK, MTK, or Freescale chips. Add in some cheap realtek parts and you have as close as you can get to an open source device (Haven't not had the opportunity to purchase any of these I can't vouch for it, but I'd assume a decent number of these devices are running the cheapest peripheral chipsets possible, which are likely either VIA or Realtek companion chips.)

Perhaps someone with more firsthand knowledge can confirm or deny these assumptions.

Re:Buy chinese!

[Tastecicles]

piss the NSA off: use Huawei.

App store review:

[Cid Highwind]

"Would rate ZERO STARS if that was an option! App works, but pesters constantly about koala communist chippies or something. Super SUPER duper annoying. Use SomeOtherApp+ Free instead."

Re:android sucks anyway

[Tastecicles]

I'm glad I still have my V3. Had it since 2003 and not once have I had any complaint to make about it!

Time to decap the chips and publish the back doors

[Anonymous Coward]

Can't do it in the USA - but folks overseas can. This WILL happen - sooner or later the facts will leak out.

Now it starts to make sense to buy only hardware from China?

Another Prisim (or should we say Prison) court action? The separate company for IP may also be a government project - There are several 'private' companies that are in actuality owned by the government.

I mean - think a bit - Skype was not worth $8B - unless the money actually came from a rich uncle.

For the idiots that think that this intelligence isn't/won't be abused: Having access to all the recorded communications of someone, lets bad guys blackmail, threaten people into submission ( and yes - ALL the recorded communications - the meta data would fit on just a few hard drives - you don't need acres for that! ). If you don't believe me, ask someone that lived under the Stasi what happens next.

It is obvious that O is just a teleprompter reader by now, I wonder who is calling the shots - same exact policies of Bush as far as Intel/freedom goes.

Ubuntu Edge?

[Anonymous Coward]

How is the Ubuntu Edge doing in this aspect?
http://www.indiegogo.com/projects/ubuntu-edge

(No, I'm not trying to increase traffic there. I'm just curious.)