Short update:
School assignments plus a bout of gastric flu hindered work. The big roadblock now is the random level generation, but I managed to get rooms out. Right now I am forsaking elegant architecture and making it work first before refactoring…
Interesting project. I haven’t played NetHack, but I played a lot of ADOM ( www.adom.de ) from gamma 10 all the way to 1.1.1. The true strength of the game wasn’t in its engine (though this was very carefully thought out), the range of monsters/items/weapons (huge), or the puzzles, but in the richness of gameplay paths and the feeling of genuine freedom. The more diverse the range of overlapping ‘quest’ options at any given time, the better. Hitting things for the sake of hitting things gets boring after a while, no matter how snazzy the weapon/enemy/experience level.
It would be very easy to write a roguelike that was no more than a thin veneer over hitting things.
I did start writing a roguelike about 10 years ago, but didn’t get very far. I had dungeon generation, basic movement, and a very cool lighting / line of sight engine. ASCII text representations of 16M-colour raytracing, with proper shadows and everything, worked very well, and actually wasn’t that hard to code up. (at the time I was writing in ARM BASIC on an Acorn RISC PC. These days I’m a professional C++ developer.)
I would personally advocate a strong OO approach, which will require a lot of design up front, but I realise that’s not how hobby projects work 
The “Debate religion with XYZ” quest is going to be challenging 
I look forward to seeing the fruits of your labour, and offer my services for testing.
L.
It’d be amusing to have stats like “Grammar”, “Spelling”, “Snark”, “Eloquence”, and “Cite”. Don’t forget to have a mission to bring pie when the player comes back to any home base area.
I can offer to program in TurboPascal using Jackson structured programming. :eek:
What? 