3 ms·

> ChromeOS has better security and performance than Ubuntu [...] I'm going to need a citation on that, especially performance. Doubly so if Crostini is put int

by Topfi 5mo ago

> ChromeOS has better security and performance than Ubuntu [...] I'm going to need a citation on that, especially performance. Doubly so if Crostini is put into the mix. > [...] updates things like peripheral firmware that Ubuntu isn't even aware of. Like what? WiFi cards, etc.? Isn't that generally in kernel already? What kind of updates do you think are not done by Ubuntu or another Linux distro? Last I tried ChromeOS was on the Pixel Slate way back when. A buggy, unstable, clearly not properly tested, unperformed mess that I would not wish upon my enemies. Glad to see it has improved to usable now, but that it is better than any other Linux distros, I can't say how considering even being on par with e.g. Fedora would have been a miracle not to long ago. Happy to admit that purely on the UI/UX, ChromeOS is very solid in my opinion, arguably and subjectively the most consistent and user friendly designed desktop environment I know. Far more consistent than anything MSFT or Apple have provided in quite some time, everything looks like it should, placement is easy to grasp and reliable with a clear identity. Consistency wise, only Gnome can hold a candle to the strictness with which the ChromeOS team execute their vision, though there is the clear divergence in the Gnome team pushing new UX innovations and concepts even if they are controversial and may need to time to learn, whilst the ChromeOS team seems purely focused on the most clearly easy to master approach one can take.

jeffbee 5mo ago
> I'm going to need a citation on that, especially performance Multiple reasons. ChromeOS ships an optimized, platform-specific kernel, built using LLVM with LTO and AutoFDO. No other distro even attempts this. The only one that has even considered it is CachyOS that offers optional LTO, or Gentoo, where you can DIY LTO, but neither supports FDO. Another reason is that Chrome GPU acceleration actually works on ChromeOS. IPU webcams work, too. On Fedora, Arch, and others you'll be patching and rebuilding kernels to get IPU. > Like what? WiFi cards, etc.? Isn't that generally in kernel already? This is another aspect where the ecosystem is the advantage, not the technical details. Chromebook makers are required to furnish firmware updates. ChromeOS will update (silently, without user intervention or notice) everything in a Chromebook: SSD controllers, battery management, radios, touchpad, USB PD controllers, the Titan security chip, the CPU, whatever. This is very different from the situation on random Linux+hardware combinations where the only source for many of these updates, if they are available at all, would be to reboot to Windows.
- Topfi 5mo ago
  > ChromeOS ships an optimized, platform-specific kernel, built using LLVM with LTO and AutoFDO. Ok. How significant is the difference they gain from that? If this yields such major gains, there must be benchmarks showcasing it. At the same time, there must be reasons why something isn't widely adopted if it can provided tangible upsides. Would be very surprised if Clear Linux (rip) and similar in spirit distros didn't go far beyond those optimisations, if they can yield measurable benefits. Even then though, there are measurable performance tradeoffs for anything running via Crostini which I know for a fact any compile time optimisations won't make up. > [...] where the ecosystem is the advantage, not the technical details. [...] SSD controllers, battery management, radios, touchpad, USB PD controllers, the Titan security chip, the CPU, whatever. I just checked and I think you are confused. ChromeOS uses fwupd [0] for those things, literally the same toolset and even sources (LVFS) to e.g. Ubuntu [1]. There is no difference in ecosystem, there is no advantage for ChromeOS here. I have to also point out that these are not "silently, without user intervention or notice", Google says so themselves [2] (except for UEFI/firmware but that was the only one you excluded in that list). Fortunately too, you wouldn't want ChromeOS (or any OS really) to do such major changes silently for many good reasons. The "technical details" are important here. They are the same, they are not automatic, they can't be superior one way or the other. It is really neat that these solutions are so robust and reliable users of ChromeOS can start to think they must be some special secret sauce, when in fact they are just FOSS solutions we have had for a while. Heck, even the verification/testing isn't unique to ChromeOS. > [...] random Linux+hardware combinations where the only source for many of these updates, if they are available at all, would be to reboot to Windows. This does both Chrome OS and the FOSS projects it is built around a disservice and is not true. And not just because I can tell more than one instance where using Windows on a newly released laptop during the early Renoir days yielded driver issues which were unresolvable because Windows Update found it necessary to force a faulty AMD driver onto my system every time I provided a network connection, even after I manually tried to suppress that specific update and had already installed the proper driver, all while Renoir support in the then current Linux kernel was flawless out of the box along with Wifi, touch screen, etc. It is great if everything feels polished and I feel the UX is great on ChromeOS, which may lead someone to think it is better than alternatives even where it can't be. But in regard to updates, how could they be, they are literally using the same solution with the ChromeOS team being happy to give credit and admit such. [0] https://developers.google.com/chromeos/peripherals/fwupd-guide-for-peripherals https://developers.google.com/chromeos/peripherals/fwupd-gui... [1] https://documentation.ubuntu.com/project/SRU/reference/exception-firmware-updates/ https://documentation.ubuntu.com/project/SRU/reference/excep... [2] https://developers.google.com/chromeos/peripherals/fwupd-guide-for-peripherals#q_do_fwupd_updates_happen_automatically https://developers.google.com/chromeos/peripherals/fwupd-gui...
omnifischer 5mo ago
Have you tried configuring secure boot - with - every single protection on Ubuntu /any distro? Just google: Mathew Garrett On The State Of Boot Security implement everything and comeback. Maybe slate with android? If you disable android in ChromeOS it a dream to use. Everything just works. And remember normal people (i.e outside USA) already have Android phone. They just login and use. All passkeys etc. synced and running. All these are a pain in fedora etc.

by Topfi 5mo ago

jeffbee 5mo ago

> I'm going to need a citation on that, especially performance Multiple reasons. ChromeOS ships an optimized, platform-specific kernel, built using LLVM with LTO and AutoFDO. No other distro even attempts this. The only one that has even considered it is CachyOS that offers optional LTO, or Gentoo, where you can DIY LTO, but neither supports FDO. Another reason is that Chrome GPU acceleration actually works on ChromeOS. IPU webcams work, too. On Fedora, Arch, and others you'll be patching and rebuilding kernels to get IPU. > Like what? WiFi cards, etc.? Isn't that generally in kernel already? This is another aspect where the ecosystem is the advantage, not the technical details. Chromebook makers are required to furnish firmware updates. ChromeOS will update (silently, without user intervention or notice) everything in a Chromebook: SSD controllers, battery management, radios, touchpad, USB PD controllers, the Titan security chip, the CPU, whatever. This is very different from the situation on random Linux+hardware combinations where the only source for many of these updates, if they are available at all, would be to reboot to Windows.

Topfi 5mo ago
> ChromeOS ships an optimized, platform-specific kernel, built using LLVM with LTO and AutoFDO. Ok. How significant is the difference they gain from that? If this yields such major gains, there must be benchmarks showcasing it. At the same time, there must be reasons why something isn't widely adopted if it can provided tangible upsides. Would be very surprised if Clear Linux (rip) and similar in spirit distros didn't go far beyond those optimisations, if they can yield measurable benefits. Even then though, there are measurable performance tradeoffs for anything running via Crostini which I know for a fact any compile time optimisations won't make up. > [...] where the ecosystem is the advantage, not the technical details. [...] SSD controllers, battery management, radios, touchpad, USB PD controllers, the Titan security chip, the CPU, whatever. I just checked and I think you are confused. ChromeOS uses fwupd [0] for those things, literally the same toolset and even sources (LVFS) to e.g. Ubuntu [1]. There is no difference in ecosystem, there is no advantage for ChromeOS here. I have to also point out that these are not "silently, without user intervention or notice", Google says so themselves [2] (except for UEFI/firmware but that was the only one you excluded in that list). Fortunately too, you wouldn't want ChromeOS (or any OS really) to do such major changes silently for many good reasons. The "technical details" are important here. They are the same, they are not automatic, they can't be superior one way or the other. It is really neat that these solutions are so robust and reliable users of ChromeOS can start to think they must be some special secret sauce, when in fact they are just FOSS solutions we have had for a while. Heck, even the verification/testing isn't unique to ChromeOS. > [...] random Linux+hardware combinations where the only source for many of these updates, if they are available at all, would be to reboot to Windows. This does both Chrome OS and the FOSS projects it is built around a disservice and is not true. And not just because I can tell more than one instance where using Windows on a newly released laptop during the early Renoir days yielded driver issues which were unresolvable because Windows Update found it necessary to force a faulty AMD driver onto my system every time I provided a network connection, even after I manually tried to suppress that specific update and had already installed the proper driver, all while Renoir support in the then current Linux kernel was flawless out of the box along with Wifi, touch screen, etc. It is great if everything feels polished and I feel the UX is great on ChromeOS, which may lead someone to think it is better than alternatives even where it can't be. But in regard to updates, how could they be, they are literally using the same solution with the ChromeOS team being happy to give credit and admit such. [0] https://developers.google.com/chromeos/peripherals/fwupd-guide-for-peripherals https://developers.google.com/chromeos/peripherals/fwupd-gui... [1] https://documentation.ubuntu.com/project/SRU/reference/exception-firmware-updates/ https://documentation.ubuntu.com/project/SRU/reference/excep... [2] https://developers.google.com/chromeos/peripherals/fwupd-guide-for-peripherals#q_do_fwupd_updates_happen_automatically https://developers.google.com/chromeos/peripherals/fwupd-gui...

omnifischer 5mo ago

Have you tried configuring secure boot - with - every single protection on Ubuntu /any distro? Just google: Mathew Garrett On The State Of Boot Security implement everything and comeback. Maybe slate with android? If you disable android in ChromeOS it a dream to use. Everything just works. And remember normal people (i.e outside USA) already have Android phone. They just login and use. All passkeys etc. synced and running. All these are a pain in fedora etc.