Long time no post :). I got a new job some time ago and haven't had too much time nor enthusiasm for anything involving a mouse, a keyboard and a monitor. Nonetheless, I'm beginning to make plans for a small rewrite. Before you make a facepalm, please let me elaborate.

Umbrarum Regnum engine is so far the best thing I have ever produced. It's completely object-oriented and at least in the beginning it was pretty easy to manage. The problems started arising when I wanted to add different displays, for instance, a character sheet, a menu screen, the map view, etc. They all have their own inner loops and their own keybindings. These seem only small issues, but the truth is that they are annoying and there's helluva lot of repeated code. That's getting too unefficient.

So, what's the plan? I think I should use the experience I gained by working with Jice and especially Merc, who was able to open my eyes to a more efficient way of software design. I think the engine should run small modules, each one associated with a "view". The module sequence would be controlled via a system of fallbacks, either fixed (the character creation module falls back, or is followed by the main map view module when the game begins) or dynamic (imagine you press ESC while playing and you get a screen with Resume, Options, Save and Quit items; Resume falls back to the previous active module, Options launches the configuration module, but then falls back to the menu again, etc.). The main engine contains the inner loop and the modules are updated in each loop iteration pretty much the same way they work right now.

Additionally, this solution would allow me to use global keybindings. The engine would execute its own keybindings check (for instance, Alt+F4, Alt+Enter, Printscreen would be global keybindings, as would be the check whether the window has been closed) along with a virtual keybindings method, owned by the current module (the 'l'ook command needn't work in the main menu, for instance).

90% of the code would remain unaltered, so I would be able to get back on track rather soon, and with a more flexible and versatile engine (which would be a double benefit, since I still plan to release Umbra as a generic engine for libtcod game developers). The only question is when I'll be able to do it. Let's hope it's sooner than later.