On Wed, Dec 02, 2015 at 12:27:14PM +0100, Norman Feske wrote:
Hello,
shortly after almost every Genode release, I see people wondering why we have not released a new live system since many years. There are even a few who conclude that the project is not going to get anywhere because there is no easy way to try it out.
Because this topic is brought up every now and then, I'd like to share my opinion about it:
https://github.com/genodelabs/genode/issues/1000#issuecomment-161260312
I hope that you find my line of thoughts reasonable. If you have ideas or thoughts to add, please don't hesitate to share them.
Hey there,
I see Genode much the same as you do- but I still think it could benefit with a unified way to build and configure modules using a declarative language, somewhat like NixOS or Guix. Though Genode wouldn't be as complex given its nested structure.
Having some scripts to help people build images with a certain configuration would be immensively valuable for general work, and sharing and showing off what Genode can do in different circumstances.
I'd certainly like to have a bucket of Genode bricks with some programmable glue. Though perhaps this is dreaming, I mention this because currently I'm setting up my long-term GNU/Linux configuration and find myself battling the idea of not having an isolated system and not being able to thread applications together. Firewalls, interfaces, routing... I'm surprised what I'm writing is holding together, and I'm sure there's ways it'll leak. I'm not a smart person, but this is the kind of 'what-if' situation you get with implicit system state.
Cheers Norman