Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

5 Years of Pop _OS

blog.system76.com

Post reply on HN
Print view View original

91–100 of 187 posts

Re: 5 Years of Pop _OS

#91

Post by colordrops »

prox wrote: post #89
colordrops wrote: post #36

Earlier quoted context omitted.

If you look at my comment history, you'll see I'm a bit of a broken record, but my records don't break anymore after switching to NixOS. It was precisely the breakage after every upgrade that got me to try it. With declarative config-based reproduceable deployment that can be rolled-back, the broken OS doesn't need to be a thing anymore.

Noob here. Why is Linux breakage even a thing? Is it complete rewrites, new drivers, a design decision?

It's built by a conglomeration of users across various projects, with differing levels of quality and support. There is only so much a distro maintainer can do to curate and test and avoid breakage. The permutations are too many for any manual process to handle, so you've got breakages.

Re: 5 Years of Pop _OS

#92

Post by colordrops »

ZoomZoomZoom wrote: post #85
colordrops wrote: post #36

Earlier quoted context omitted.

If you look at my comment history, you'll see I'm a bit of a broken record, but my records don't break anymore after switching to NixOS. It was precisely the breakage after every upgrade that got me to try it. With declarative config-based reproduceable deployment that can be rolled-back, the broken OS doesn't need to be a thing anymore.

Try a buggy kernel and breaking your network in one `nixos-rebuild switch`, garbage-collecting older working generations and trying to fix your config back into working state. Great fun. Other than that, NixOS is a godsend.

Why would you garbage-collect older working generations without testing whether the latest works. You'd have to go out of your way to do that, and it goes against the entire purpose of it.

Re: 5 Years of Pop _OS

#93

Post by npteljes »

thisiswrongggg wrote: post #80
nu11ptr wrote: post #6

I really like Pop Os. I've always liked Debian descended distros, and prior to this I was using Ubuntu. While I like it, I always found a lot of rough edges esp. when it came to proprietary drivers, etc. where things didn't always work without some tweaking. Pop Os on the other hand seems to "just work" - it is like a polished version of Ubuntu with a nice selection of apps in the Pop Shop. Theme is pretty nice as we…

I'm curious. With all these debian-based distros, why people don't just use debian? I mean it has to be sturdier than the downstream distros.

One of the things, for me, was that Debian's releases were confusing. Ubuntu had a clear schedule, aligned to a calendar, with a clear out of life policy.

The second thing was that I couldn't figure out my way from debian.org, to the relevant ISO to download. There's just so many options, scattered explanations, I felt overwhelmed by the site and couldn't figure it out reliably.

The third thing is the out of box experience. Every major Debian descendant has this mentioned in the articles that compare them. Ubuntu is like a refined, user friendly Debian. Mint further polishes Ubuntu. Elementary is designed for newbies and to have excellent OOTB experience. And so on.

Fourth thing, that Debian Stable seems to be behind in version numbers, and Debian Testing seems to be not stable enough. Maybe these are not true, but they haven't inspired confidence in me either.

Fifth, Debian came with some proprietary things off by default, so no WIFI on my laptop, no MP3, bluetooth, other minor inconveniences. I love open source, but I'm not a purist, and I wanted to make use of my computer first, and improve on it later.

After Ubuntu pushing their current in-house bullshit again (snaps this time), and Mint behaving weird on my main computer for some reason (maybe the multiple release upgrades), I took the plunge into Debian after some research, and ended up on using Debian Testing with KDE. So far, the experience has been good, but this really feels like a Linux enthusiast's distro, where I had to look up things on the Internet, because the installed just haven't explained anything - me, with 10+ years of Linux experience, ending up with installing a system that won't boot into the graphical desktop, despite selecting the KDE option in the installer!

Re: 5 Years of Pop _OS

#94

Post by julienfr112 »

I left Ubuntu for pop_os for the 22.04 release, on a Thinkpad p14s. Whent well, except mouse crashed sometime when suspend (left click anavailable !) And video crashing (geometric pattern all over the display). I'm not sure I made the good choice.

Re: 5 Years of Pop _OS

#95

Post by ZoomZoomZoom »

colordrops wrote: post #92
ZoomZoomZoom wrote: post #85

Earlier quoted context omitted.

Try a buggy kernel and breaking your network in one `nixos-rebuild switch`, garbage-collecting older working generations and trying to fix your config back into working state. Great fun. Other than that, NixOS is a godsend.

Why would you garbage-collect older working generations without testing whether the latest works. You'd have to go out of your way to do that, and it goes against the entire purpose of it.

Not every bug manifests itself right after reboot. You can easily have a bunch of generations you just think are ok.

The point is entirely different: if you update the channel-list you need the network for any changes to the configuration and I don't know of a way to go back to an earlier channel state.

Re: 5 Years of Pop _OS

#96

Post by pjmlp »

tlamponi wrote: post #90
pjmlp wrote: post #67

And this is why the Linux desktop is domed to be a VM inside other OSes, or filled with Electron apps. Gtk, Qt, XFCE, GNOME and KDE applications will look out of place in the new desktop, and most likely they will never create a Rust based set of frameworks that can match the above ones on features, development tooling, and extensibility points for desktop developers.

I, and most people I worked with, couldn't care less if the border/title bar looks slightly odd, the same happens on Windows a lot. The issue ain't different WMs or toolkits, but a lock-in effort from Microsoft with HW and SW vendors, they still had to create WSL to actually improve their dev experience though, could not manage that without. Most people ain't technical experienced or interested in IT and just don't c…

I am yet to bother with WSL, it hardly brings something I wasn't already having with VMWare and VirtualBox, and in any case "developer != UNIX".

The cleverness of WSL is that Microsoft came to realise, thanks to Apple, thank many folks only care about having a POSIX userland and don't really care about Linux as such.

Given that current generations mix up Linux with UNIX, it made more business sense to add Linux compatibility than revive SUA.

So now those folks mostly jump between Apple and Microsoft platforms instead of supporting Linux OEMs as they should in first place, genius.

Re: 5 Years of Pop _OS

#97

Post by npteljes »

pjmlp wrote: post #67

And this is why the Linux desktop is domed to be a VM inside other OSes, or filled with Electron apps. Gtk, Qt, XFCE, GNOME and KDE applications will look out of place in the new desktop, and most likely they will never create a Rust based set of frameworks that can match the above ones on features, development tooling, and extensibility points for desktop developers.

It's not like Windows is not a disjointed experience, UI-wise. The Win 98 days are long gone, when all the apps had a nice uniform native Windows UI - and even then, maybe it's just that my glasses are rosy. Maybe macOS fares better in this regard, I have no experience, but they have got the reputation.

To zoom out - the reasons desktop Linux is not adopted on a wider scale have nothing to do with its technological prowess. It's working as well as any other desktop OS would. The reasons are rather political, and especially business-related. With making schools teach Windows and MS Office, giving governments and public offices deals, having vendors bundle it by default to new computers, not enforcing anti-piracy, is the way Microsoft achieved that their software is the desktop standard, not by making the software superior in specialized contexts like a uniform UI, or a better architecture. Their blend of adequate software, vertical integration, ruthless business, and some support for every type of user is what won them their current status.

Re: 5 Years of Pop _OS

#98

Post by colordrops »

rgoulter wrote: post #54
colordrops wrote: post #36

Earlier quoted context omitted.

If you look at my comment history, you'll see I'm a bit of a broken record, but my records don't break anymore after switching to NixOS. It was precisely the breakage after every upgrade that got me to try it. With declarative config-based reproduceable deployment that can be rolled-back, the broken OS doesn't need to be a thing anymore.

I think it's worth adding that NixOS probably will require more effort to use than the "it just works" UX that GP seems to be wanting. Getting familiar with nix can take some time. e.g. figuring out how to write a Nix package if there's not a package already can be difficult. When things work in NixOS, it really does feel very nice to know you have a declared config (so that you don't have to worry about what changes…

You could say the same about Linux in general. It does takes some time to really get comfortable with the complex surface area of the system, but once you do, the benefits are tremendous and without doubt. Same goes for programming languages, driving a car, learning guitar, etc. I don't think people really understand the fundamental shift of moving to declarative config. It's absolutely worth the cost of admission and offers "superpowers" for many of us. Others may not need or desire these powers, and Ubuntu or Windows or MacOS will do the job.

Re: 5 Years of Pop _OS

#99

Post by pjmlp »

npteljes wrote: post #97
pjmlp wrote: post #67

And this is why the Linux desktop is domed to be a VM inside other OSes, or filled with Electron apps. Gtk, Qt, XFCE, GNOME and KDE applications will look out of place in the new desktop, and most likely they will never create a Rust based set of frameworks that can match the above ones on features, development tooling, and extensibility points for desktop developers.

It's not like Windows is not a disjointed experience, UI-wise. The Win 98 days are long gone, when all the apps had a nice uniform native Windows UI - and even then, maybe it's just that my glasses are rosy. Maybe macOS fares better in this regard, I have no experience, but they have got the reputation. To zoom out - the reasons desktop Linux is not adopted on a wider scale have nothing to do with its technological p…

Too much focus on the UI and very little on OS frameworks.

Re: 5 Years of Pop _OS

#100

Post by colordrops »

ZoomZoomZoom wrote: post #95
colordrops wrote: post #92

Earlier quoted context omitted.

Why would you garbage-collect older working generations without testing whether the latest works. You'd have to go out of your way to do that, and it goes against the entire purpose of it.

Not every bug manifests itself right after reboot. You can easily have a bunch of generations you just think are ok. The point is entirely different: if you update the channel-list you need the network for any changes to the configuration and I don't know of a way to go back to an earlier channel state.

Once again, why wouldn't you keep the last few generations around? No network is required to use them.
Post reply on HN