None of these reasons have anything to do with user experience -- no wonder none of these reasons have earned Linux any market share outside of picky developers.
> Astronomically lower probability of catching a virus.
Correct, but when was the last time you got a virus on Windows? (Obviously it's different for Regular Joe User, I'm talking about the HN audience).
> No bullshit forced updates when the OS feels like it.
In practice, my Windows install always updates in the middle of the night when I'm not using it. It closes all of my apps and only reopens the web browser. But, it's not so bad. It doesn't bother me. I can however see how it might annoy some people. But I think the effects of this are exaggerated.
> No phoning home with you being unable to launch an app when your manufacturer's server is slow.
Windows phones home, but to be fair, this post is about Linux vs Windows, and being unable to launch an app is a MacOS problem, not Windows.
> No shuffling about trying to install drivers for a thousand components before your computer is usable.
This is true in some situations. For instance, if you install Windows from scratch you'll have to wait while it downloads drivers. More of an automated process, no shuffling about. The alternative on Linux is to roll the dice and hope your drivers are baked into the kernel, if not you're back to (very manual) shuffling about.
> In practice, my Windows install always updates in the middle of the night when I'm not using it. It closes all of my apps and only reopens the web browser. But, it's not so bad. It doesn't bother me. I can however see how it might annoy some people. But I think the effects of this are exaggerated.
Apple has definitely figured out the "reopen apps when you reboot your computer" feature. I wonder why it's taken Microsoft so long to do the same.
Even if Microsoft were to provide some special API for this, tons of legacy programs won't be using it, and most new ones will also ignore it.
Under Microsoft Windows, application installers can register some handler to run on startup, and can implement this themselves: the program can check whether an instance of the program was interrupted by reboot, and if so, start it up in a special way whereby it is told to recover the state from the most recently saved parameters. Those could include volatile state like position of windows, object selections and whatever.
There is really nothing for Microsoft to do there other than maybe lead by example; implement some sort of best practice in a few notable Microsoft programs, document the practice and encourage developers to do same.
I've been using Windows as a daily driver for over 20 years and have never had a virus or malware. For a technical savvy user it's really easy to avoid by just not doing stupid things.
> Astronomically lower probability of catching a virus.
Yes because nobody uses consumer desktop Linux. With higher market share the malware will come.
> No bullshit forced updates when the OS feels like it.
True, but that also means you're going to end up with a lot more botnet nodes out there because people will never update. Or they'll never update and still complain things are not working.
> No phoning home with you being unable to launch an app when your manufacturer's server is slow.
I hear you, but this doesn't happen enough for anybody to care. People would be more fet up by Facebook downtime.
> No shuffling about trying to install drivers for a thousand components before your computer is usable.
Correct me if I'm wrong, but Linux drivers still can require endless faffing about. The biggie is that distros won't distribute nonfree drivers by default, for ideological reasons. The average person doesn't understand that they need to go enable some repo in their package manager, only that their machine isn't working.
> True, but that also means you're going to end up with a lot more botnet nodes out there because people will never update. Or they'll never update and still complain things are not working.
Or people update when it's convenient for them vs when it's convenient for the OS.
...which is nice from a user control perspective, but completely ignores gp's argument. You can either have user control or everyone up to date. You can't have both.
> > Astronomically lower probability of catching a virus.
> Yes because nobody uses consumer desktop Linux. With higher market share the malware will come.
Linux has a higher server market share already, and its virus problem there is still not as bad as Windows'.
> > No bullshit forced updates when the OS feels like it.
> True, but that also means you're going to end up with a lot more botnet nodes out there because people will never update. Or they'll never update and still complain things are not working.
Today, Windows has forced updates and Linux doesn't, and Windows hosts are more likely to be part of a botnet than Linux hosts.
> distros won't distribute nonfree drivers by default, for ideological reasons.
True but misleading. For example, Ubuntu doesn't install nonfree drivers by default, but all you have to do to install them is check the checkbox when the installer asks you. Way easier than dealing with drivers on Windows.
> Linux has a higher server market share already, and its virus problem there is still not as bad as Windows'.
Server use cases are vastly different than Desktop use cases. I have administered a lot of Windows servers in my time and not a one of them has ever had a virus. Workstations on the other hand...
In fact, the latest versions of Proton are so good, the only games I haven't been able to get to work with regularity are the ones that require EA Origin, and honestly, I've had similar problems on Windows. In fact, sometimes the Proton version of the game works better than the Linux binary. So, if gaming is holding people up from switching (which is a common excuse), that's not as much an obstacle these days.
2. Any time there is an update, it's an opportunity for something to break. If a user has no choice when a computer updates, they risk interrupting important, time-sensitive work. So, a user must ask themselves, are they more worried about their machine possibly used in a botnet, or are they more worried about it rebooting right in the middle of a meeting or video call? Or taking 20-30 minutes to update after an unexpected reboot, such as a power failure? I would rather have the choice, and deal with problems caused by updates only I want to, when I have the time. Also, many zero-days are for software that's over a year old. As long as a user has done an update in the past year, which is a reasonable expectation, the risk of compromised security is much lower.
3. Facebook downtime and the incapability to launch any third-party application at all are two very different problems. Further, Gatekeeper is not the only source is potential problems. Windows Defender has been known to quarantine DLLs and executables that are perfectly harmless, but fixing problems caused by an overzealous update to (and application of) Windows Defender definitions is often beyond the capabilities of the average user.
4. The Linux kernel absolutely supports non-free drivers. Inclusion of non-free driver "blobs" is a common argument.
In fact, I would like to interject for a moment. What you're referring to as "ideological Linux", is, in fact, GNU/Linux, or as I've recently taken to calling it, GNU plus Linux...
>"The Linux kernel absolutely supports non-free drivers. Inclusion of non-free driver "blobs" is a common argument."
Linux kernel has no stable ABI for drivers as a matter of ideology / strategy, we can debate it's merits but you certainky can't claim it suppports them.
And when there is no driver available the only option you have is to buy a new device. None of my 4(4!) Sound devices work with a current linux distribution. Not the Realtek Chip on my x570 board, not my logitech webcam, not the speakers in my monitor via hdmi and not my USB Sound Card from Creative.
Not even the onboard Realtek Lan controller works, i had to buy a usb one.
Oh and, i had to plug the cord from all sound devices after a reboot into Windows, because none worked anymore.
This situation is even worse than in the late nineties/early 2000 ish.
Yeah, sadly. The Chips I have are supported, but the device ids they have aren't supported by the kernel, i would have to check it out from source, add those ids and compile it on my own. At least i could find that for the network adapter, i guess its the same for the sound devices, cause they just use normal parts from realtek and the like.
Ive just bought those parts because they had the best performance and features for their price point, i could literally spend over a thousand euros more to only get a little bit more features or performance. I kinda had hopes that nearly everyone who would build a rig these days would buy those and because of that someone would have made them linux compatible in the last two years.
> * No bullshit forced updates when the OS feels like it.
Eh... kinda, actually. I can recall many times wanting to try out an application only to discover that my LTS distro which was gasp 4 years old didn't have it in the repo, so I was forced to upgrade in order to get it without setting up a build environment and recompiling.
If AppImage were more widely embraced this sort of thing wouldn't need to happen.
