Linux Is Not Stable(crowfunder.github.io)
crowfunder.github.io
Linux Is Not Stable
https://crowfunder.github.io/posts/linux-is-not-stable/
7 comments
Hi, I wanted to share a bit of my frustration about Linux usage after whole 2 years of using it as my daily driver. I collected most of the more annoying issues I've encountered, the list is not long, fortunately, but features some especially gruesome issues.
In other words, Linux remains a tinker's toolkit more so than a polished product.
Who knew?
Who knew?
I feel like recently there's been a certain discourse that Linux is a full-fledged replacement for Windows and MacOS, even for non-technical users.
I love and live inside linux, so please go easy on me but...there is some level of product design that an OS needs to be a viable consumer OS. Linux has 100 ways to do any thing, this is 1: paralysis inducing, 2: causes high cognitive load when all you need is to edit/share a spreadsheet, 3: does not create a common experience that you can share with people across jobs. Everyone will do it differently, so you will have an issue communicating how to do task X.
But I see no product people on Linux, I see only engineers wanting maximum Linux. We aren't willing to be more single minded, we want to be nothing like Microsoft (good), but we also want to be nothing like Apple (good in some ways, very bad in others).
Regular users do not need to know what apt is, what a repository is, or any of the 1000 linux things. But those things need to work so consistently well that they could use the OS without ever, and I mean ever, having to know what they are.
Then, I haven't used a linux desktop in a while (tried elementaryOs 2y ago, was a bit lacking), but the desktop environments need to stop looking like some college student's java GUI project.
Finally, I don't know much about the driver/nvidia issues that I hear so much about (that's not where my job takes me), but I don't think we need to solve those before we can get Linux to be a daily desktop driver. I mean let's some up with a list of Linux certified cards and let OEMs pick from those? Maybe this is already done, but if not, we could start there.
But I see no product people on Linux, I see only engineers wanting maximum Linux. We aren't willing to be more single minded, we want to be nothing like Microsoft (good), but we also want to be nothing like Apple (good in some ways, very bad in others).
Regular users do not need to know what apt is, what a repository is, or any of the 1000 linux things. But those things need to work so consistently well that they could use the OS without ever, and I mean ever, having to know what they are.
Then, I haven't used a linux desktop in a while (tried elementaryOs 2y ago, was a bit lacking), but the desktop environments need to stop looking like some college student's java GUI project.
Finally, I don't know much about the driver/nvidia issues that I hear so much about (that's not where my job takes me), but I don't think we need to solve those before we can get Linux to be a daily desktop driver. I mean let's some up with a list of Linux certified cards and let OEMs pick from those? Maybe this is already done, but if not, we could start there.
So it's the year of the Linux desktop --- again.
Surely 2026 will be it!
The issue I see is that if someone comes up with actual polished desktop experience, they will eventually ruin it like rest. And SteamOS is not desktop experience. Even if it is extremely nice store client. They make money from store. Others will make it from adds, analytics and so on... So it will end up ruined.
The issue I see is that if someone comes up with actual polished desktop experience, they will eventually ruin it like rest. And SteamOS is not desktop experience. Even if it is extremely nice store client. They make money from store. Others will make it from adds, analytics and so on... So it will end up ruined.
The biggest operational hazard to any personal computer is being forced to update it by dependencies outside of your control.
The next biggest hazard is giving in to the temptation to update it because you think updating will improve something.
The operational instability of a given system is a distant but significant third place hazard, which is difficult to mitigate because of the lack of clear understanding about what the system is good for and how it should work in the first place.
If you truly don't care to how a personal computer works (or can't know due to some limitation) and you find yourself dependent on other people to deliver an appliance to do something — which by definition you only know how to do because you were shown — then you are ensnared in the technology and issues about stability are mostly about imperfections in a social contract, and rarely about about imperfections in any instance of the technology.
Everybody knows how to what they do with a personal computer because they are motivated by curiosity about its utility.
Linux is useful, and if you find utility motivating, you can learn how to use Linux to get work done. Just like Win and mac.
Along with linux's utility, you also get the complete freedom to purpose the system as you see fit, whereas Win and mac impose sharp limits on your freedom.
As to the value of your freedom and the value of social contracts, that's a complicated domain of tradeoffs.
As history has presented to us all a great case study in these trade offs in the form of personal computers, we find ourselves subject to the enduring observation that individuals who trade liberty for security penultimately suffer the lack of both. But you might not notice if you're very lucky, and the computer industry has given rise to more than its fair share of very lucky people.