The new demo should address common Allegro questions in a clear and commented manner.

I agree. I have one right off the bat. Drawing a bitmap with partial transparency.

I think a demo should be written in simple C with an effort to illustrate how to use Allegro's best features in a logical manner. Great gameplay and graphics should be a secondary goal.

The focus should be on:

• Clear, easy to read, understandable source code!

• Getting input from joystick, keyboard, and mouse

• Loading some external media from a datafile

• Playing background MIDI and sound effects

• Using a double buffer and blitting sprites (perhaps with rotation, scaling and translucency - if easily done)

• User controlled object moving in simple x/y directions

• Simple collision detection between sprites

• Highscore list (saving / reading to disk)

• Loading / Saving settings (config files)

Stuff like 3D, GUIs, etc, just make the thing messy for a beginner to understand. I'd go with a single resolution (with high-color) to avoid the confusion that comes with "customization". I'd even stay away from potential pitfalls like linked lists and clever programming tricks. There's also no need to get into complex math - I'd keep it as simple as possible.

A demo that simply illustrates the "power" of Allegro is a waste of space in the source package. I'd hate to see even more space in the ZIP be spent on a demo.dat that is rarely used. If someone wants to see what Allegro can do, then they ought to just download a full fledged game.

All in all, the "keep it simple" philosophy.

The more I think about it, the harder it seems. Games aren't simple.

A game I thought of doing would be kind of like missile command, but polar. You have a satellite on a planet that you control with the left and right keys (your orbit direction). The mouse is where the satellite shoots its lasers and rockets. Rockets constantly track the mouse position. Missiles come in from all directions. The outposts on the planet constantly rebuild themselves (but slowly).

But... while I'm still going to make it. Making it simple (and pretty!) seems almost impossible. I'll need a list for bullets, a list for rockets, a list for (enemy) missiles, and so on. It'll be pretty complex. So, I guess I'll have to find something simpler, and keep "Planetary Defender" as more of a "what you can do with Allegro" thing. As it'll have a lot of spiffy blending and lighting (the lasers glow).

A platform game has a tile system, collision detection, AI, general logic, and scrolling (to just name a few!). That's pretty big in itself.

That's why I'm having trouble of thinking of something "simple" that hasn't been abused already.

Hmm.. I was thinking of another Dungeons kind of game (Dandy dungeons, the inspiration behind Gauntlet). Basically it's gauntlet with boulderdash style tile movement for everything. It's fun, and would be easier to program (for the game complexity) than other genre's.

I did a version on the Amiga.

Why not use Alex4?
None of us have produced something as polished before, now have we?

That code is bad. Great game, but man that code is hard to read.

Your entry, perhaps?

I would improve an older game of mine but all my games are written in C++. I really cant understand why people still write games in C, but I guess thats neither here nor there..

I dont have time to start a new project but I would be more than willing to review code written by other people and also write comments.

Just write a demo game in C or C++ and submit - then Evert (or whoever) can select one or more of the games as demo(s).

How many of you have studied the demo game code to learn something? I usually go for the example things.

That's exactly the point. A good demo would be something you can use as base of your own games. You definitely shouldn't do that with the current one.