A roguelike of Mine

I have been developing a little roguelike during last weeks, named Mines of Elderlore. This game has been created for the 7 day roguelike challenge, the code was only one file. I wanted it to be full GNU/GPL v2, so that it may be of some help for anyone interested.

I started developing it with simple needs in mind: I wanted to create a little roguelike around the new dungeon generator I had just created for Lands of Elderlore, my other roguelike. And I wanted to test how I could implement those simple features:

Permanent levels

Well, not only permanent levels on your computer, but on every computer of every player! So that when you provide (automatically or by player input) a dungeon name, the same levels will always be created.

Python is a great language for this, and it is really easy to store dungeons in a dictionary and save games with the pickle module.

A common risk when creating permanent levels is to reduce the amount of space to explore for the player: once every level crawled, what can he do ? So I decided to create two downing stairs for every level, each one leading to a different -and permanent- level.

There are 9 floors in Mines of Elderlore. That may not seem a lot, but with this 'two-stairs' design, you can have as much as 1 + 2 + 4 + 8 + 16 + 32 + 64 + 128 + 256 = 511 different floors!

Having permanent levels provides a great advantage: player can compare their scores. I did not have time to implement it yet, but I plan to provide an online display of every player scores on landsof.elderlore.com. By sending their keymovie files, not only players will be able to compete with each other, but when published online their movies will inspire other players and challenge them to perform better! Some kind of Massively Mutliplayer Offline Role Playing Game ;-).

Easy gameplay

Roguelikes can have a steep learning curve for new comers, and to my point of view every effort to ease their first roguelike experiment should be sherished.

First idea is to display symbols with two ascii characters. So that the player can see at a glance what weapon he is wearing, what level is the monster in front of him, and so on. Simple and effective. And with fonts twice as high as large, symbols are displayed perfectly squared!

Then I tried to add weapon special attacks based on the player movements rather than additional keys or over complicated RPG skills. Charging is maybe the simplest example: by moving in the same direction during several rounds, you increase your 'charge' amount, and if you manage to hit a monster at the end, you will inflict increased damages.

I think this enhanced gameplay based on player movement can really be developped very far; you can count on me to investigate further on this !

A recent post on rec.games.roguelike.development made me realize that I am really focusing too much on frills instead of developing important features, such as gameplay ones. Why that ? There are several reasons that I will enumerate and see how I could change my evil habbits.

The eye candy syndrome

Lands of Elderlore has graphics, so a part of its development requires adding eye candy features. The challenge here is to set the limit between pure decor and gameplay impact.

Of course, it is satisfying to have a nice GUI or round buttons with pictures, or nice alpha channeled shadows. And having graphic effects will definitively attrack new players to the game - specially RPG players that don't want to try ascii roguelikes.

But the goal of Lands of Elderlore is also to immerse players in medieval worlds, push roleplay as far as possible, and provide a feeling of gigantism and freedom through procedural generated content.

The blogger case

Speech less, code more ! I like to discover new web technology, try website engines, learn new ways of publishing on the web. But I don't have so much spare time, and each hour passed blogging or mounting web services around the game is an hour with no new feature and no new bug erased. I really need to code more. Hopefully, now that this website is launched, I should be able to focus on the code feist.

The selfish developer

As I am approching release 0.1 - the first truly playable release -, I feel strangely more and more reluctant to publish my precious. Right now, I'm the only developer, and I wish to stay in that state, even if the game is under GNU /GPL. I've been giving a lot of time and energy in that game, and I guess that somehow I am a little anxious about the feedback from the first real players that will appear after that release.

More than that, once this playable release published, the game will not only be mine, but will belong to the community. I've never published an open source code like that before - well, except TomeTik, but that was just a hack from Tome -, and perhaps I am just over reacting. But I guess by reading other roguelike dev blogs that I am not the only one having a hard time to publish the first real release of my code.

The shortest path

After all those considerations, why do I still want to release ? My main concern is that my spare time will probably be reduced in a couple of monthes, with the arrival of my second little hobbit.

And, well, I'd really like to know what the gamers think of Lands of Elderlore. Taking care of one's ego is never a bad thing !

So here we are, here is the short list of features I still need to add in the game before 0.1:

• Create a short list of weapons and armors, with flavors giving different strength to them
• Add sellers in cities, with the OcempGui window for buying and selling
• Add basic dungeons under world cave tiles
• Add simple monsters in mountains and dungeons
• Add close combat to animals and monsters
• Deal player death and heirs

Once these feature are added, I swear it by the name of the dark lord of the forgotten code, I will release Lands of Elderlore 0.1 !

So be it.