With yesterday’s Flutter Live event and the stable release of Flutter, one of the primary ways to create Fuchsia apps, Google is one step closer to possibly unveiling their in-development operating system. Another unexpected step is coming, in the form of the official Android Emulator from Android Studio gaining the ability to boot Fuchsia’s Zircon kernel.
While Google can be quite fickle, I feel every step forward for Fuchsia is a step towards the grave for Android/Linux.
i wonder if Android was the high point of Linux and it’s going downhill from here.
android/fuchsia on mobile, and kubernetes/docker on datacenter. both are distrupting how software on linux is supposed to be developed and packaged. and there are likely more examples like that.
No. Docker is still very much tied to how Linux works (Linux ABI, Linux kernel namespaces, etc.), and all the cloud stuff is Linux internally. Even AWS Lambda uses a KVM based hypervisor underneath.
Not the variant that makes use of Windows containers and Hyper-V.
How common is it to use windows docker containers. That has to be a real niche use case, no? Or is that basically the under pinning of azure?
They are relatively new, since Windows Server 2016.
The niche case would be enterprises that are Windows shops, running IIS and such.
So… much… wrong….
Well, ok, not “wrong” as much as “oversimplified”.
Linux is a kernel. It manages memory, processes, and IO. It’s not an operating system. Red Hat, Debian, Arch, Android– these are all operating systems. They all have different methods of deploying packages.
Docker is a special use-case that runs software in a container on top of a linux kernel (approximately) which is why I can run Red Hat, Debian, Arch and even theoretically, Android, in a linux container, regardless of the underlying distribution.
While I agree that containers (docker) and sandboxes (android/fuschia) are different ways of deploying applications, the underlying linux kernel remains the same (or pretty close to it).
Docker also works on top of Windows containers.
Docker on Windows generally uses docker-machine (or Docker for Windows, which is the same thing with less flexibility…) to run Linux in a VM, and then uses that VM for containers. There is a Docker which containerizes IIS and the like, but that’s not what people generally are using Docker for Windows for.
I would not call Android the “Highpoint of Linux” for end users possibly but the Highpoint of Linux will always remain Servers and super Computers two areas that will only grow more over time.
Personally I am warry about his kernel change, and will put on my Tin foil hat for a moment and say I feel that google is moving to a different kernel for there mobile O.S. to get more control and prevent things like the alternative roms from being made that remove the Google applications and the related Google creepiness.
My hope is that the current Linux/Android Development is moved to a consortium ran by the different phone makers. it would be exciting to possibly see three different mobile operating systems out there.
On the consumer space, yes Android was the high-point because it did away with all the cruft that is typically placed atop a Linux kernel, much of it developed with the flawed “Bazaar” process (X.org, ALSA, PulseAudio, the QT vs GTK split and the useless wrappers that come with it, dependency-based packaged managers incompatible with each other etc etc), and combined with tons of funding from Google and a centralized development process it was developed into something good.
But Linux is holding Android it back and Google knows it. For example, GPU drivers have to be welded to the kernel. In Fuscia, drivers come as shared libraries and run in user-space. If this works, it will be amazing.
BTW, the MIT license of Fuscia won’t change anything with regards to fragmentation, in any direction. Android is already permissively-licensed, and when it comes to Linux all the OEMs have to do is dump a useless kernel with lots of binary blobs to a server. Google can keep the existing GMS licensing model intact.
Edited 2018-12-09 20:49 UTC
“If this works, it will be amazing”
If this works, companies will release less and less open source code, bugs won’t be fixed because it works for the needed cases and not anymore.
Google controls the development of Android and maybe this fuchsia thing, then he won’t accept any patches or features that aren’t in his agenda of policies.
He will control everything for his sole benefit.
Linux at least many companies control it and anyone can add features if they make it do things better. Imagine the situation with a single company controlling everything.. It will be so sad.
I don’t care about some binary blob-ed source code. I care about being able to take the old drivers and run them on a new version of the OS, just like I can run Windows 7 drivers on Windows 10. Bugs aren’t getting fixed anyway due to them existing inside the binary blobs (aka where the real business value is). Project Treble is a bandaid. The kernel shouldn’t be left behind because drivers are old, the kernel has to be decoupled from the drivers.
Edited 2018-12-10 20:07 UTC
You are kidding right? Linus alone controls everything. He does not just accept anything that has been submitted. WTH
You’re kidding, right? You don’t really believe Linus alone controls everything?
Google has a weird and schizophrenic relationship with Linux and Android as they do with all open source.
They want everything to be “open” but they want control of it…
And Fuchsia seems no different.
Being the originator gives them a special measure of control over the platform, free from the constraints of a Linux. But it is licenced MIT, which gives them even less control over what others do with it.
I’m sure Fuchsia will get some level of support because its Google giving away a free thing.
But if they keeps it MIT/BSD style where the hardware vendors can literally do whatever they want, Google will find that the situation with fragmentation and unmaintained hardware they are just now addressing in Android all the worse.
OTOH, If Google exercises tight control over this new platform, it won’t take OEM’s (especially large ones) very long to figure out they are better off with Android.
I could be wrong, but I see more ways for Fuchsia to hurt Google than help it, unless they keep it very niche. I find myself hoping it grows up healthy enough to develop a community and be more than a toy (all my Linux bigotry notwithstanding).
IMO but surely MIT/BSD licensing does not change anything as long as Google sets identical requirements for certification. Under GPL you still can modify as much as you’d like anyway. Once forked and wouldn’t meet the requirements, it cannot be sold as Android or whatever it’d be called for Fuchsia.
Edited 2018-12-09 20:15 UTC
It should be nice to be able to get a hold of a decent mobile operating system that doesn’t require a snapdragon 865 just to be able to do the same thing I have been doing since the single core 624mhz days. Browse the web, listen to music, remote desktop, navigation, gaming and watch movies.
However, it would be nice if it operates more like a full blown operating system…..
like in the old 624mhz days.
So I am hopeful this goes in the right direction, ill will wait and see. Because, in the end. The only thing that counts for the dumb consumer is if it can run an app
Linux does run on older slow ARMs. The Sharp Zaurus for example. It was quite snappy. The benefits of low resolution displays and bitmapped fonts.