[hi-v-rocknroll]

No idea. It seems to have gone off in the direction of "shiny and new" without deeply considering more practical architectural decisions like isolating system vs. userland, configuration management, or manageability. It's unfortunate because there are zillions of Linux distros, and most of them (numerically) are terrible, insanely impractical, user hostile, and/or poorly documented. If you're going to create a distro, at least have a use-case justification for it where other distros or other OSes ([wd]o|ca)n't meet the needs of them adequately.

[e3bc54b2]

> It seems to have gone off in the direction of "shiny and new"

Ikey's last foray into distro-building also left same impression on me.

> without deeply considering more practical architectural decisions like isolating system vs. userland, configuration management, or manageability

I dunno, FWIW, Solus was very well regarded in UX for a while. It had the best steam integration, and the whole thing felt fairly well thought out for something built and maintained by half dozen people.

> If you're going to create a distro, at least have a use-case justification for it

That is unfair. The developers felt there is a gap between fully imperative state-modifying mudball and fully declarative purity-land, which is true, and decided to try their hand at it. How well they did is yet to be seen.

[hi-v-rocknroll]

> That is unfair. The developers felt there is a gap between fully imperative state-modifying mudball and fully declarative purity-land, which is true, and decided to try their hand at it. How well they did is yet to be seen.

What I meant was: the thesis has to be crystal clear rather than building into the aether "organically" by hope and wishes rather than practical need. This is a similar failure of coding away on a "startup" without sales only to discover no one needs or wants to buy the result.

Time and time again, Linux distros create messy, commingled system with user dependencies that increase system entropy and footgun stability. Even Alpine isn't immutable or really self-contained. The major, novel departure in this area was Core OS.

As an example under-served category in the marketplace: If one wanted to create a more resilient server OS, I would evaluate bringing together aspects of the following:

- Qubes - app isolation

- RHEL - kernel

- Rocky/Alma - governance

- Core OS - system isolation

- NixOS - package management

- FreeBSD - development, integration, and QA as a system

- (N/A) - fleet management (missing in 99.999% of Linux distros)