@civodul 2025: all guix services running in ocap containers with no distinction between "system" level services. directory and file arrangements granted only via explicit capabilities and arranged based on what's logical to that particular process
@cwebber@civodul sounds like a mess, but now that i think about it, this is probably the only way we're gonna get android-style safety on linux in a user-friendly way without having to flatpak every app?
@civodul Let's be clear that Debian's situation is an entirely internal clusterfuck. It will make a nice case study someday in botched project management and toxic interpersonal dynamics, but it has no implications re whether abolishing the / vs /usr distinction is a good idea in the abstract.
@zwol@civodul I consider Debian users to have a great pragmatism but Ive always found Guix/Nix/Arch/Alpine coders or system crafters to be more interesting.
Beyond the conservative tooling availability that Debian offers I feel that the lack of risk taking and personal responsibility towards their operating system impedes their personal knowledge, capabilities, and development.
@indieterminacy@civodul I couldn't disagree more, actually. Conservatism in the development of the lowest levels of an operating system is extremely valuable as it enables people to experiment at higher levels without fear of having the foundations pulled out from under them. I used Debian exclusively for 20+ years for precisely this reason. I'm moving away from it now because of the usrmerge clusterfuck, but that's a deviation from their normal practice and I still hope that it won't change what their normal is.
@indieterminacy@civodul And it's also interesting to observe that if they had carried out this transition in the boring, slow way the dpkg maintainer wanted it done, they would be finished already, and with far fewer flamewars.
@zwol@civodul ... and as somebody who started out using Linux using a small harddrive, I was punished iin the face of limited disk space. With no warning or contextualisation updating would then uninstall X11 facilities as a first point of protection
Hey, I got to overcome knowledge gaps as a consequence but my predilection for startx is a consequence of trauma not from experimentation but rather uncaring settings which could have been mitigated with signposting and forcing choices.
@zwol@civodul My last Debian build died an ugly death. I was compiling a font package tool which enveloped the entire GUI infrastructure, banjazing anything in that purview.
Silly hacker, missing a bit sure.
But anything I tried to save the patient only made it sicker and sicker, what a dark rabbit hole to enter with a surgeon's knife.
If only rollbacks or containerising akin to guix shell was available, I wouldnt have left that slick rig I had lovingly crafted.
@indieterminacy@civodul Yeah, I think the Guix way has a lot of promise as the way of the future -- but Guix explicitly set out to do something New And Different.
@zwol@civodul ... well NixOS set out explicitly to do something New And Different, ... Guix merely approached it better #Guile
Arrgh, I hope people arent feeling bad. OS developers and packagers have to do lots of patient and arduous work and yet its the bits above these layers which get all the shiny toys and then get to complain.
Its all ants and grasshoppers, really. If people arent happy with how their operating system is working they should shop around or step up and fill leadership gaps
@civodul As an organization Debian unfortunately has had serious problems with toxic interpersonal dynamics for a very long time. Before the usrmerge argument there was the systemd argument, and before that there was something else that I (thankfully) can't remember what it was anymore.
It's actually an impressive achievement of theirs that for the most part users of Debian have been able to ignore all of the screaming on the development lists for decades.
Add comment