I would like to take the time to establish a little roadmap for myself and the other technical folks who might be giving this a read.

Now that a lot of general housekeeping has been done for MonasTech, I would like to focus on producing new "styles" of systems. Although I really think tiling window managers are the way to go, I understand that's not what most people want.

I will be designing most of my systems around Wayland instead of X. Specifically, I'd like to focus on wlroots compositors for reasons that are primarily related to compatibility. The follow compositors are the ones I intend to explore and provide solid support for.

Primary:

• Sway
  • It's just good. It's what I daily drive.
• Labwc
  • The most "suckless" floating WM in my opinion.
  • XML-based configuration that will work excellently with Guile Scheme.
• Wayfire
  • Seems to be the "other" stable compositor that everyone uses.
  • Maximalistic effects are very cool.

Secondary:

• Niri
  • I think this could be more approachable to Windows users, but that's just a hunch. Seems like a solid blend between WM-style and floating-style compositors. I like it, but I need to do more experimentation.
• Hyprland
  • The kiddos seem to like this one. Seems like Sway but with a different ecosystem.

The first order of business is to produce example systems that I can confidently dogfood on my own machines. I am putting effort into organizing a very lightweight "base" system style that allows for most customization to be done on the home environment level. That way you could have several different users with completely different configurations on the same machine. It also makes updates a lot cleaner, only requiring a relog at the worst.

I'll create more posts once I get something done for each.

On the backend side, there are a few important things that I need to do.

• Create a proper update service
• Setup a proper build farm with substitutes
  • Substitutes are already available at substitutes.monastech.xyz but it's limited to what I manually choose to build as of right now.
  • I think setting up a build machine interface for incredibly low-end devices is a good idea.
• Organize the aformentioned "base" system
  • This really just means figuring out what is the best platform to build off of. It's looking like it's going to be greetd with something similar to tuigreet because I do not like the idea of running another compositor solely for the login manager. However, I understand that users will be users and they will probably want something pretty like QtGreet. I'll figure it out.