Roughly a year after launching on the original Nest Hub, Google is making the Fuchsia operating system available for the Nest Hub Max.
For over five years now, Google has been quietly toiling away on Fuchsia, an operating system intended to replace and/or compete with Linux. While many Google fans were hoping that Fuchsia’s launch would be a splashy one, like that of Android in 2008, the real launch was nearly as quiet as the development itself.
The slow, steady march to replace every operating system on consumer Google devices with Fuchsia continues.
Things to keep an eye on when comparing it to GNU/Linux and Android. How many years of official support will the device get and what happens when that is over. How many versions of (future) Fuchsia OS will the device support. Will the device allow root access. Without relying on exploits. Security in regards to CVEs. Will there be public device drivers available. Will you be able to install custom version of Fuchsia OS on it. Will you be able to install software on it in a normal way. That is without 3rd party being involved. How can you collaborate and improve Fuchsia OS long term. For it to support your specific needs and hardware. Lets see.
Geck,
I am in agreement with that stuff too, however at the same time there are a lot of linux/android devices that don’t offer those things. Some people have gotten their devices bricked for trying including the gen 1 linux/android nest controllers.
https://forum.xda-developers.com/t/google-home-nest-hub-flash-method-exposed-and-tool-revealed.4284689/
If I had any authority at google, I would do everything in my power to make sure that future Fuchsia devices did not repeat the mistakes of the past that make devices so stupidly difficult to mod and support. IMHO this would be a huge selling point for the modding community. “Designed for Fuchsia” devices could guaranty robust standards for the modding community right out of the box. I know it’s probably a pipe dream, but I’m so sick of how difficult and non-standard things have been on the linux/arm side of the equation. It would not be hard for them to define mod community friendly standards for Fuchsia certification if google leadership actually wanted to.
sukru: if you can somehow convince your employer to do this, you’d be doing the world a huge favor, haha.
Alfman,
Organizationally we are like separate companies. I have no contacts there 🙂
But I have seen their workplace. They are really hard working, and the work environment “speaks” the team is dedicated to hardware and OS hacking.
I agree. I hope they will continue an open approach to development.
@Alfman
I will stick with GNU/Linux and their way of doing things. As i find it technically and practically superior in all key areas i care about. For people that don’t see it like that and feel somebody else can do a better job. Or do have a grudge against GNU/Linux. A feeling they are being held back. I am OK with that. For such people to go down a different path. As long as you will get what you are after with Fuchsia OS. I am OK with that.
If at some point people will feel cheated then likely GNU/Linux will still be there. Nothing is for ever but said that we aren’t going anywhere.
Geck,
I know that you will and if linux & android are working well for you then more power to you. However it does feel like you are using double standards at times.
It’s not so much that I want Fuchsia over linux, I’m already familiar with linux and I’d like it to work out every time but it always comes back to the same problems. In particular I hate being tethered to non-mainline & obsolete manufacturer kernels with bad support. If Fuchsia were able to deliver a better FOSS experience over linux then I’d have to consider switching. It wouldn’t be because I have a grudge against linux, but because of the pragmatic reality that linux on non-x86 has has a hell of a lot of hurdles and I’m just tired of that.
@Alfman
Without FOSS drivers you are not getting a FOSS experience. On top of that if you have no control whatsoever over the device. It’s like running Windows. But even Windows allows you to install different versions on hardware. Correction. It used to allow that.
Geck,
That is exactly the double standard I’m talking about. If manufacturers are providing source code then yay both linux and Fuschia can benefit. Otherwise though the closed drivers are a problem for Fuschia and linux alike. However in this case it’s objectively worse for linux! The drivers needed to use the hardware are tightly coupled to a specific linux kernel. When manufacturers aren’t providing source code we NEED drivers to be loosely coupled in order to maximize the FOSS experience for the rest of the kernel. This isn’t an opinion so much as it is a logically conclusive fact. Ignoring this inconvenient fact comes at the expense of more restrictions for the user base, which is completely contrary to the goals of FOSS.
While it’s completely fine for you to have a preference for linux despite all of this, I think this sort of self-serving hypocrisy in linux circles is unfortunately hurting FOSS users more than you care to admit. While Fuchsia hasn’t done enough to earn my trust yet, the unwillingness of the linux community to recognize and address it’s own shortcomings is an indicator we really shouldn’t be putting all our eggs in the linux basket. The linux monoculture has placed ego over user needs, the cumulative effects of which have been quite detrimental for FOSS.
@Alfman
Rest assured that i will ask you from time to time. Things like on how many “Zircon kernels” were you able to upgrade on this new Nest device. On how things like that are working on your side. Discussing security should be fun too. Will it satisfy all this high expectations. From as you say FOSS person point of view. Or if it will basically be just another brick for people that couldn’t care less. From all the key points we discussed in the past. Then Google can do with it you know what. In “worst case” scenario a library will likely be made and we will run GNU/Linux on top of Fuchsia OS blobs. Obviously minority of us. Or something like that. But i still feel that ARM on mobile, desktop and server space will need to produce FOSS device drivers in the future. To stay relevant. And when we will have that. Then we can make an unbiased and in-depth analysis on what is better and what should be the standard. Before that we can’t make any final conclusions. And i am sure that we will get to there. To have both options available in an flagship package. That is with Google or without them.
Geck,
I wouldn’t object to that, but don’t forget that linux is in a decades long stalemate that might never end. We need a viable alternative to the linux FOSS monopoly to give users some hope that the future of FOSS can be more than false promises adnauseam.
@Alfman
That is your opinion. In my opinion GNU/Linux enabled decades of innovations, services and products we otherwise wouldn’t have. Including Google. And i don’t see this changing. With or without Google. If you really want to call out somebody. Then call out Google. On why they haven’t produced a flagship mobile device with FOSS device drivers. As they could if they wanted to.
Geck,
No Geck, unfortunately closed linux kernels are not my opinion. They are a matter of fact for millions and maybe even billions of linux devices.. You choose to ignore it out of fondness for linux, but it’s all factually true and it continues to be a serious problem for FOSS under linux.
I agree that plenty of innovation gets built on linux but that is in no way a rebuttal to user gripes when the hardware is so tightly coupled to specific manufacturer kernels. And if you don’t see linux ever changing, then that’s all the more reason we should have an alternative to it.
Most device manufacturers aren’t manufacturing the silicon themselves, so I think the world’s qualcoms are more directly responsible for providing binary drivers & code under strict NDAs. I don’t like this state of affairs any more than you do, but pointing the finger at everyone else has done crap to put FOSS in the hands of users who’ve wanted it for decades. If linux is either unwilling or unable to change then it may be time to put another batter to plate who is willing and able to do everything possible to maximize FOSS benefits for the community.
@Alfman
Indeed. Billions of devices (each year?). That information alone should tell you GNU/Linux is really not an issue here. But it’s a solution. Device manufacturers. For them it’s all open source anyway. They can do whatever they want with it. It’s hence in the end a choice and not technical limitation. As for you personally. You will never be able to do more with blobs. If you want to do more. Then you need FOSS device drivers. There is just no way around that. Anyway, good luck to the ones that feel they can do a better job at it. In addition we will get to a flagship mobile device with FOSS device drivers. With or without Google. Google doesn’t seem to show any real interest for achieving such goal anyway. Somebody else is hence the solution we are waiting for. Not Google.
Geck,
But that’s such a dumb argument…just because something is widely used does not mean there is no problem. Let me flip the tables just to make a point: you’ve been complaining about google chrome becoming the monoculture of web browsers. But FYI chrome is on billions of devices. That information alone should tell you that google chrome is really not an issue here.
I really hope you roll your eyes at that! I know as a linux evangelical you are hooked into defending linux even to a fault, which is your prerogative. But the FOSS community really is suffering and after all these years I for one am growing sick of false promises. Nobody really knows where the next leadership for FOSS will come from, but after decades of stalling it’s clear that saying the course isn’t going to solve the problems with linux. And that’s the problem, you are consistently unable to offer any solutions other than pointing fingers and waiting indefinitely. We need to take active & concrete steps to maximize FOSS utility today! Everything else is an excuse!!
@Alfman
Your opinion and claim basically comes down to GNU/Linux is bad as you decided it is bad. The main grudge you hold against it is due to you not being able to upgrade a kernel on an ARM device. But that is not due to technical limitations. It’s due to choice of the device manufacturer. And now hence we all should let Google force upon us an operating system that promotes and encourages blobs by design. And after you hope you will be able to upgrade Zircon on it. Well, buy this latest Nest device and tell me if you can upgrade Zircon on it. After we can talk some more.
Geck,
No it does not and I’m disappointed that you would choose to misrepresent my position so badly.
We’ve spoken at length about this and it’s clear you are thinking dogmatically rather than pragmatically. We can’t control other actors, and I know it sucks, but we can control ourselves!!! To this end Linux is not doing everything possible to maximize FOSS utility for users today. That is 100% linux’s fault and nobody else! If you open your eyes and read the room you have to admit that there are many of us being harmed by linux’s position to keep drivers tethered to specific kernels. The entirety of your plan consists on waiting for other external actors to do the right thing by you, but you don’t appreciate how badly your plan condemns linux users to a future where FOSS remains non-viable for so many of us on our own hardware.
Unfortunately it’s not just google, but the vaste majority of product manufacturers. They all exist under the same NDAs. We got very lucky with x86. Linux grew of age when direct hardware access was prevalent on x86 and there’s a critical mass of enterprise deployments to keep it well supported there, which is good. But new architectures are not following x86’s evolutionary path; proprietary drivers and manufacturer dependencies are the norm and there’s no signs it will change, which is extremely bad for your plan to keep waiting things out.
I hope that you are right, the FOSS community is in desperate need of a platform that works better than linux/android do. However for the record I’ve never put Fucshia nor Zircon (or google) on a pedestal, I only highlight them because they’re potentially more viable than others at making a dent in the linux monoculture keeping us trapped in this situation.
Geck,
Some corrections…
Cast devices use the aptly named “Cast OS” which is based on Chromium, not Android (with some exceptions):
https://chromium.googlesource.com/chromium/src/+/refs/heads/main/build/config/chromecast_build.gni
Chrome devices usually have really long support cycles (I have seen 7+ years). And when that runs out it might be possible to install the “Flex” version (albeit with less features):
https://www.androidpolice.com/install-chrome-os-flex-chromebook-explainer/
I don’t know the business details, but I believe the migration to Fushia is mainly for driver support and security reasons. Of course we don’t actually know how long it will be around. Nevertheless coming from the Chrome team, I would expect it to be really long. (This is my personal opinion).
Thanks for the explanation. Basically yes. Lets see if Google will offer more then 7 years of official support then for such devices. And what will happen once they stop providing support. If there will be alternatives. Maybe GNU/Linux? And we will get a more clear picture on how the situation with CVEs will be. It’s not like we have much control over it. And currently we are more or less having theoretical discussions due to lack of such devices and operating system on the market.
Geck,
Fuchsia has already been rolled out to (some?) devices a year ago:
https://www.techhive.com/article/579622/how-to-check-if-your-google-nest-hub-is-running-fuchsia.html
This is the smaller version (Nest Home), not this other one referred in the article (Home Max). They are mostly the similar, but this one has a camera and other minor differences.
Opinion: This would be very different if Linux provided a stable driver ABI, or chipset manufacturers would support their ARM products on Linux for more than a few short years.
@sukru
Yes we have discussed this device already in the past on this site. But it looks like nobody has it to give some remarks in regard to what i wrote initially. All just mention “stable ABI” and claim that is on why it is better. Beyond that no clues whatsoever on what is actually going on with this device. And no real comparison with the GNU/Linux counterpart on all key points.
Geck,
I agree. Even though Fuchsia is open source, we don’t actually know much about it. For Linux, I can go to https://almalinux.org/ or any other distribution, and try it out today.
To try out Fuchsia, I have to go though these instructions: https://fuchsia.dev/fuchsia-src/development/hardware/intel_nuc. (Not that I did not try, but it failed on some step, and I don’t actually have as much free time as before).
Time will tell.
But wrt. “stable ABI”: That is a real concern. Android has been bitten badly because of it. Even with large efforts like Project Treble (https://android-developers.googleblog.com/2017/05/here-comes-treble-modular-base-for.html), the issue still lingers. Would using another existing OS, like FreeBSD, be better? Maybe, we will never know.
@sukru
Let say Google would fully embrace GNU/Linux upstream and maintained device drivers for (their) flagship mobile phones. There would be no such issues they are trying to resolve in the first place. In my opinion Google is just using GNU/Linux in the wrong way. And in my opinion they should at least try to produce one such device. And lets see after if the world will stop. Or what exactly will happen. As somebody will do it and if it won’t be them. Then somebody else will. Other manufacturers would in my opinion follow soon. Seeing all the benefits and tons of money to be made.
@sukru
I haven’t tested it just yet and it does look a bit complex. But you could play a bit with dahliaOS. If you are interested in trying out something that is Zircon based.
Geck,
Don’t you think it would have taken off years ago if there was tons of money to be made? We who care strongly about FOSS are a tiny niche of the overall market, less than 1%. While I’m fine with you lobbying for that to happen, you and others have already been preaching the same thing for eons and frankly these passive “let’s keep waiting” strategies are not that likely to pay off. Even if you believe there’s a small chance, the refusal to do anything else is still holding back other avenues of progress that could be accomplished. The FOSS community has suffered long enough over inaction.
Geck,
We need to put things in the historical context. If Android was designed today things might have been different.
About 15 years ago, it had virtually no market share (only a single phone model on T-Mobile), and was competing with Windows Mobile and Nokia/Symbian, and to a point iPhone. Phones were “once and done” devices, and almost none of them received firmware updates.
And it was not only Android that was bit by lack of manufacturer support. TI exiting OMAP business for example, left X11 out in the cold as well: https://www.phoronix.com/scan.php?page=news_item&px=X.Org-Driver-Shambles-2022
Alfman,
I agree, there is almost no market for “fully open source” hardware. If we add Raspberry PI, Altera, System 76 and others together, they won’t even make 1% of the entire computing industry. That is the sad truth.
Linux on x86 prospered thanks to either hardware manuals being more common (Texas Instruments for example had very detailed information on the NE2000 cards), or being more accessible to hacking.
With ARM, things are different. Again, TI OMAP was different with a 3000+ page manual: https://www.ti.com/lit/ug/spruf98y/spruf98y.pdf . But even that lost support by open source community.
@Alfman and @sukru
It looks like that even Raspberry Pi will get a working FOSS 3D GPU device driver before Google on any of their flagship mobile devices. And i agree that without GNU/Linux there would literary be no Google. So Google get your act together. I don’t care if you tinker with some ideas from 80s in your warehouses. Just do your job. Produce a FOSS flagship mobile device. Or face the consequences of standing in the way of that.
Geck,
The thing is removing one entity from history means that all the billions of dollars invested in linux would have gone towards other similar projects. The BSDs could be more prevalent or heck maybe even hurd would be mature, haha.
Geck, Alfman,
Open sourcing “your own hardware” is not always a straightforward task. Even open sourcing your own software can be a problem: https://github.com/id-Software/DOOM (id software could not release DOS DOOM source code due to third party sound library).
I have no idea what the business deals include with the hardware partners. However given Android was unable to convince them to keep open source drivers, even though there is an actual measurable need for it, my guess would be on the negative side of the spectrum.
I might of course be completely wrong, and the next version (Android 13?) comes with fully open sourced driver and graphics stack, and Samsung releases source code for all their firmware.