The future of Linux

pmk@lemmy.sdf.org to Linux@lemmy.ml – 252 points –

I'm not proposing anything here, I'm curious what you all think of the future.

What is your vision for what you want Linux to be?

I often read about wanting a smooth desktop experience like on MacOS, or having all the hardware and applications supported like Windows, or the convenience of Google products (mail, cloud storage, docs), etc.

A few years ago people were talking about convergence of phone/desktop, i.e. you plug your phone into a big screen and keyboard and it's now your desktop computer. That's one vision. ChromeOS has its "everything is in the cloud" vision. Stallman has his vision where no matter what it is, the most important part is that it's free software.

If you could decide the future of personal computing, what would it be?

182

You are viewing a single comment

For Linux in general? If I could decide? Here goes:

  • I would want people to realize that distro maintainers are actually important and Flatpak and co. are not actually as good of a thing as everyone makes them out to be.
  • I would want a full actively developed GNUstep-based desktop environment as the "default" Linux desktop (which apparently was the original intention).
  • I would want Xorg to finally go away.

I love Flatpaks and have embraced them totally on my desktop. They just make sense for sandboxing applications with Flatseal. Distro maintainers also ship software with bad defaults. I want to be able to easily control that.

sandboxing applications

Sandboxing is nice, but is not exclusive to Flatpak. I'm talking about the software distribution model.

Distro maintainers also ship software with bad defaults. I want to be able to easily control that.

So you trade distro packages which you can usually customize pretty well if you need to (e.g. modified Arch packages on the AUR) for Flatpaks where you have no chance at all to do that because the package build script isn't available? And the problem still isn't solved because now the people who can set bad defaults still exist, they're just different people.

Aside from all the things Flatpaks get right, a sandbox framework built right into the design is a major win and in my case, it's one of the big reasons I went with it.

You can also modify the build script for any flatpak's manifest and create packages with flatpak-builder. May not be as easy as PKGBUILDs but it's certainly possible.

The problem of bad software defaults is easily solved with Flatseal. My point is that, it takes a few clicks to deny permissions to Flatpak applications as opposed to sandboxing a traditional app yourself.

Writing sandbox profiles for apparmor or something similar is usually a complex elaborate affair. And even when you do finally manage to get a working profile, it still requires maintenance to keep the sandbox functional as the target software updates. You won't face any of these problems with Flatpak's bubblewrap.

5 more...