Google is replacing some Android apps for Chromebooks with Progressive Web Apps (PWAs). A PWA is essentially a webpage that looks and feels like a traditional app.
This will certainly be good news for many Chromebook owners. In some cases, PWAs are faster and more functional than their Android counterparts. PWAs also take up less storage and require less juice to run.
When PWAs are a better option than Android applications, you know you’re scraping the bottom of the barrel. I really don’t understand why Google doesn’t just turn Chrome OS into a more traditional desktop Linux distribution – they’ll get better applications, better tooling, and better performance than shoehorning Android applications into Chrome or pretending a website is an application.
ChromeOS is based on Gentoo and has a custom GUI layer. That was a key advantage when ChromeOS was basically an immutable flat image built from source, but now it means they’d have to do a bunch of work including providing an X server and huge binary package mirrors everywhere to provide graphical applications, and either tell their users who want newer application versions “deal with it lol” or keep up with the breakneck treadmill pace of building from source that the Arch maintainers do. Given how Google has a history of killing projects that suddenly spike in effort required, they probably want to avoid that.
Don’t ever let them fool you. There is one and only one reason why companies push for “web” based apps: they only have to maintain *one* codebase.
Damnshock,
The way you started, I thought you were building up to something more ominous, haha.
The article says it’s because of this:
It doesn’t surprise me that apps made for android don’t work as well on chromebooks.
Yes of course this is it. With google chrome using web apps, they have to do near zero toolkit development or maintenance.
“pretending a website is an application” – the distinction between a website and an app has blended tremendously already over the years. With web technology one can make mobile apps (e.g. using Cordova), and PWAs are no exception. Really, do you feel you’re on a website when using “apps” like Google Docs in a browser?
I completely agree. Websites are a great way to distribute applications. The developer completely controls the upgrade process so everyone is always on the latest version of the application. Lots of processing can already happen on the client and only the secure/datasensitive parts normally run on a server. There is a whole bunch more of positives (and negatives) that can be written about webapps, but by far the bulk of all NEW development in the last 15-20 years has been done as a web-application.
On my computer I use Slack (built on friggin’ jQuery) all day, Notion, even VS Code is built on web tech, etc.. They are all still a standalone apps. No one cares what the app is built with. No one. The way you interact with it determines whether it’s an “app” or a “web page”.
“I really don’t understand why Google doesn’t just turn Chrome OS into a more traditional desktop Linux distribution”
Because that would mean delivering Linux applications, turning the Chromebook into a more traditional netbook/laptop, and competing with Ubuntu, etc.
Unless you plan to go after the desktop and server space, it doesn’t make any sense for Chrome OS to be anything other than a platform that can share Android apps with phones / tablets, or PWA apps with other platforms.
Whether Chrome OS / Chromebooks make any sense at all is another question.
Why don’t they turn the Chromebook into a traditional Linux computer? Because Linux netbooks did so very well the last time. In all seriousness, they won’t do that because it would be far too much effort to maintain and support, not to mention the loss of marketshare once the education institutes lose their nice Google built-in MDM.
Because of security,
“Linux for Chromebooks: Secure Development (Google I/O ’19)”
https://www.youtube.com/watch?v=pRlh8LX4kQI
From “everything is a file'” to “everything is a web page” ……
I’d love to see a proper Linux Desktop, with real developer tools, and a GUI subsystem that isn’t DOG SLOW (X11 on PopOS!), or quirky as heck (Wayland – which has a secondary problem of having an opinionated maintenance team). I’m not sure if the Linux desktop enthusiasts can’t see the sluggishness, or just don’t care. ChromeOS doesn’t have these problem though, the gui is light and snappy – so it would be nice to have a more full app ecosystem. And it’d be great if they’d release standalone builds, so I could play around on my own hardware…
(BTW, Linux is not slow – properly ported games on my nVidia based machine run FASTER and SMOOTHER on Linux – it’s just the damned GUI.)
Wayland is an api specification, not an implementation. So are you talking about the spec team having odd opinions or the reference implemntation? or Mutter?
Linux GUI’s seem fast to me on relatively modern, but not super powered hardware ( i5 kobe lake U series nucs). I cant’ tell the difference between windows 10 speed on same hardware.
I have a yoga 730 – i7-8550U with nVidia 1050. I can’t even see the option in PopOS! (which is labeled “wayland” – I don’t care if it’s in some way just an spec, that’s what the option is called) when nVidia is enabled. Without enabling the Wayland, it’s slow. Noticeably sluggish. Firefox is even worse, because all the hardware accelleration is disabled. And even if I enable hardware support through about:config, it’s still slow – and it spins my fans up like a mother, but it’s not just Firefox. Again, I don’t know if it’s that Linux enthusiasts just get used it, don’t notice it, or don’t care, but I’m a graphics heavy user, and I can tell you – it’s slow (even text input has a noticeable lag – so it’s not like it’s not noticable even in the terminal). It’s much slower than Windows, and slower than macOS (which is every so slightly slower than Windows, probably because of the extra buffering it does). Twitch styles shooters are my favorite type of games, so maybe I’m just sensitive to it.
When Wayland is enabled, its fast, but then I have various problems with tearing and other types of issues. But at least with the Wayland option enabled, it gives me gestures, so that’s another plus.
The Nvidia GPU is most likely the problem. You might try an Intel iGPU or AMD GPU with the FOSS driver, and see that makes a difference.
If gaming or CUDA are use cases, sure Nvidia, but for desktops, something with FOSS drivers (Intel, AMD) is preferrable.
And there’s the rub. Computer users don’t give a crap about ideology. They want to buy something, install the driver, and have it just bloody work. The sooner Linux devs understand that and make it a stable environment for *developers* rather than ideologs, they might just then see their Linux desktop gain traction. But who am I kidding? The ones who understand this… they work at Google!
Right, which is why Intel and AMD GPUs work out of the box on mainstream distros. Nvidia is the difficult one. People would do the work for them, if they would let them, but they don’t want to play.
Linux is already a great-ish environment for developers. Great-ish because most distros are static distros, and I think the FreeBSD model where the third-party software is separate from the base is better. Anyway, the Linux kernel devs freeze the user facing ABI/API stuff to a fault, but they do break the kernel’s internal API/ABI when needed, which is where Nvidia gets into trouble since they don’t have an in-tree driver like Intel and AMD do. Without an in-tree driver, the kernel devs can’t fix regressions.
https://en.wikipedia.org/wiki/Application_binary_interface#/media/File:Linux_kernel_interfaces.svg
Technically savvy people can make it work. However, Linux is not a great environment for regular people who want to do stuff. It’s all about software, and Windows and MacOS have a lock on that. Apple would have died without Adobe and MS Office, and without commercial software, Linux is doomed to be a niche desktop OS.
Nvidia had lots of problems with changes in Windows Vista, by the way. The Nvidia drivers were responsible for 28% of Vista crashes. https://www.engadget.com/2008-03-27-nvidia-drivers-responsible-for-nearly-30-of-vista-crashes-in-20.html
Speaking of Windows, people are on an upgrade treadmill with MS. People do have to manage their software and hardware with each new Windows release, especially commercial software, or risk getting stuck on a specific version. Window’s backwards compatibility isn’t as great as people make it out to be.
The reason Google doesn’t want to give users a real Linux Desktop is because the traditional Linux Desktop is a mess. X.org is a bug-ridden mess (particularly around multi-screen support, video v-sync and font scaling) and a slow one at that. The different toolkits (GTK and KDE) are a mess particularly in the way they work with each other, with random bugs showing. The repositories system is a mess. Then there is the problem that lots of the libraries that paper over the nastiness that is X11 are fricking GPL, which leaves proprietary developers to use SDL and having to draw their own frickin’ buttons and drop-down menus or buy a licence for Qt. Meanwhile the Android libraries for creating UI elements are free.
Literally every successful or semi-succesful Unix-like OS had to do away with the traditional X.org desktop stack. Mac OS X, Android, Chrome OS, OpenELEC and Enigma come to mind.