Category Archives: Programming

Lua remake of an Amiga cracktro

Recently, I wanted to test the Plus wrapper of Harfang on a simple project, and I found that I could try to remake an ancient Amiga cracktro I used to enjoy back in the days.

YouTube Preview Image

The starfield is a simple list of 3D points on which a vector is applied each frame, transformed from 3D to 2D and displayed using 2D sprites. The 3D to 2D projection works the usual way :

x2d = (a.x * zoom) / a.z
y2d = (a.y * zoom) / a.z

There’s room for an optimization, as the whole list of sprites could be drawn as a unique batch. If more performance was needed, for several thousands of stars for example, DrawSpriteAuto() would really work faster.

The venerable ProTracker module, available on the Exotica archives, was converted to the XM format and replayed easily in Harfang, using the following instruction :


Let’s Scroll!
The scroll text, finally, was made easy to render thanks to the Plus wrapper, by blitting each glyph to the screen. The original font was grabbed from the original cracktro using a memory snapshot from the WinUAE emulator.

Each glyph is then drawn, using :


This remake is not 100% accurate as the scope of this project was more focused on using the Plus wrapper than being in competition with the most excellent Flashtro project.

Here is the Win64 release :

Here is the source code on Github :

Amiga Memories

This blog remained silent for too long. As I’ve been rather busy lately, I’ll adopt the shortest Input/Output style for this post πŸ™‚



The recent project (purely hobbyist project, that said) that occupied me during the past weeks is centered on the old computer Commodore Amiga.

The input of the video sequence below are (basically) :

– The GameStart3D game engine.
– The MaryTTS text to speech engine.
– A few 3D models borrowed to friends or made by myself.
– The following squirrel script :

The output is the video sequence below :


What I did is more or less an automated talking box, some sort of “Max Headroom” inside an Amiga. This sequence is automatically generated, frame by frame.

More to come … stay tuned πŸ™‚

Astlan on Steam/Greenlight!

Please help me to push ‘Astlan’ into the selection process of Steam/Greenlight. It needs every upvote, favorite or comment from you. The more votes there are, the more a game is likely to be considered for the Steam games catalog!

Click here to visit my page on Steam and to support AstroLander.

Astlan on Greenlight.

I bet it’ll take a million of votes in order to go thru the selection process, but hey … let’s try!

Astro Lander, a development summary

This project started as a simple remake of Lunar Lander.

Lunar Lander relies on a 2D gameplay with a simple use of physics and it exactly what I was looking for. I’m going to explore in this post how it finally evolved into something bigger, iteration after iteration.

A long list of improvements.

The initial change I wanted to implement on my remake was a multi directional scrolling, to offer a larger game area.
I built rapidly a minimalist level made of cubic blocks and started to write the physics and the controls of the Lunar Lander.
I found out that the game play could be accelerated, and then I made the ship to move faster. Rapidly, the challenge evolved from “land softly” to “navigate softly and avoid the walls”.
The goal of the game would be not only to reach a specific point, but also to collect a series of items scattered across each level.

The controls.

I started to focus on the controls of the ship. A Lunar Lander is basically about controlling the ship’s reactors versus the gravity and the inertia.
But I wanted to avoid the usual problems related to the rotation of the ship. If the player is in charge of both the thrust and the roll angle, a situation might occur where the ship is upside down and the thrust direction will sum up with the gravity, thus sending the ship even faster against the nearest wall.
To suppress the frustration inducted by such a control scheme, I wrote a simple routine that always controls the rotation of the ship. The player is free to move in any direction, the physics engine handles all the collision and physical responses, the ship can translation, rotate, bump into walls, but in the end it always remain properly oriented.
In my opinion, this is a good compromise for a physics-driver action game, where the physics engine makes the game looks real but doesn’t spoil the gameplay.

The code of this routine takes only a few lines of code :

function AutoAlign(item)
local _align, _speed

// Get the current orientation
local _rot_z = ItemGetRotation(item).z
// Get the angular speed (if the ship is rotating)
local _ang_v_z = ItemGetAngularVelocity(item).z

// Magic factors applied using an empirical approach …
_align = Clamp(Abs(RadianToDegree(_rot_z)) / 180.0,0.0,1.0)
_align *= 250.0

// … that defines the strength of the torque (angular force),
// applied on the ship to stabilize it
_speed = ItemGetLinearVelocity(item).Len()
_speed = RangeAdjust(_speed, 0.25, 0.5, 0.0, 1.0)
_speed = Clamp(_speed, 0.0, 1.0)
_align *= _speed

// Applies the torque
ItemApplyTorque(item, Vector(0,0,-_rot_z – _ang_v_z).Scale(_align * ItemGetMass(item)))

One of my friends who works in the game industry always told me that I needed to refine the controls before doing anything else on the game. During my past attempts at making a game, I disregarded this advice, and started with the level design. I eventually understood that Romain was right.

To bring the ship in motion, the player will only need to trigger the 2 side thrusters by touching the 2 halves of the screen. Such a basic control is perfectly suitable for smartphones and tablets, allowing you to concentrate on the game without having to constantly check that your fingers are properly positioned on the screen.
With the help of Emmanuel, I was finally able to tweak the core routine that drives the ship and defines the way lateral forces are applied to it. It ended up with the exact kind of behavior I was looking for, so the level design could start, at last!

The Level design

As I knew I couldn’t spend too much time on the graphics, I headed for the simplest method I could think of to build the levels : “lego bricks”. This technique, based on blocks, presented two benefits to me :

  • it is cheap in terms of time investment, because you only need to prepare a few blocks and you’re free to work with the GameStart editor. I didn’t have to constantly exchange back-and-forth between 3dsMax and GameStart.
  • it let’s the door open to a future extension that brings a lot of value to a game : a level editor that the players can work with to assemble new levels from scratch.

It has drawbacks too, of course. It’s a bit heavier for the engine to handle a large list of split blocks, and the overall visual aspect tends to be less realistic.

What’s next ?

Now that the game is almost done, with a series of 24 different levels, all we need is to polish the final builds both for iOS and Android platforms. As soon as the game is released, I’ll be able to focus on the next big implementation : the level editor.

In the meantime, you may enjoy the gameplay video below :

Preparing a trailer for Astlan

Today, as we have planned a walk to the Aerotrain’s test site, I decided to bring my iPhone and a tripod.
The Aerotrain railway makes a perfect landscape to create a fun & intriguing trailer for my (almost finished) game ‘Astlan’.

For the record, the Aerotrain prototype was designed during the 70’s as an high-speed alternative to the regular railways. A test site running on 25km, made of a concrete rail was built in order to test this prototype (max speed recorded was around 400km/h).

Nowadays, only the concrete rail remains, making a perfect place for urban exploration.

Here are a few mockups I made of the kind of framing I want to shoot once on the site. We’ll see if the final result is close enough. Or not πŸ™‚

Astlan, sunday late update.

Both Saturday and Sunday were profitable days. I managed to stick on AstroLander and work a dozen of details, such as :

  • Normalize the audio tracks
  • Improve the touch/click reactivity of the main GUI
  • Write a few common function to create UI buttons and ultimately refactor the existing UI code
  • Implement a “Pause Menu” when in-game. Works perfectly and stops the game completely (display, stopwatch, physics, particles and whatnot)
  • Fix an inconsistent design in the score calculations
  • Renamed a dozen of scene files and their respective classes

Not bad, for a Sunday πŸ™‚

Ideas for my next project

While it is too early to expose here the idea I have regarding my next game project, I can say it will involve robots…animated robots.

The challenge here will be to animate the robots using a procedural technique as I want them to be able to place their hands randomly.

For this purpose I will probably use two things :

1. This robot I modeled years ago (here in a shameless photoshopping)


2. The physic engine of GameStart, and especially joints, to assemble a sort of ragdoll, so that I shall only need to animate the hands if I want to move the arms πŸ™‚

Stay tuned.
I need to finish and release Astlan first, anyway!

Update on CosmicLander

Here I a quick summary of the most recents update I made to my latest game project:

  • The name was finally changed to Astlan, as in Astro Lander. I hope it will avoid any possible copyright issue.
  • 20 levels out of 24 are now completed, with different kinds of traps, such as doors, rotating barriers or lasers.
  • The game UI is slowly taking shape, thanks to the help of an amazing local artist who made the intro screen, displayed somewhere in this page. So far, the main title, settings screen and level selection menu are functional.
  • I recently manage to throttle my physics routines when the game happens to slow down too much. This was a matter of concern to me, especially regarding the Android version.
  • Yup, you read it right. There’s an Android version around πŸ™‚

I will post a playable version including the updates listed here before January 2012 so that people will be able to test, and hopefully make feedbacks πŸ™‚



Cosmic Lander, a new minigame project

I recently Β started to work on a new indie game on my spare time.

The initial goal was to develop a series of really small games within very short development cycles, in order to address a largest spectrum possible with my favorite game engine to date: GameStart

It turns out that it rapidly evolved into something a bit more ambitious, even I’ve been able to prioritize the tasks, and especially to work on a gameplay before starting the real artworks.

From a basic “Lunar Lander” it ought to be, it moved to a gameplay with a series of closed levels (caves ?) in which the player must seek for a list of items, while struggling against the gravity.

The graphics are basic so far, as mentioned above, mainly because I want to focus on what I’m not confident with : the level design. The visual aspect of the caves might eventually get painted or textured, but the real challenge here is to create funny & challenging levels.

So far, the game features 4 levels, no score, the basic win/die gameloop, a minimap and a few bonus & traps. The physics engine plays an important part in the gameplay.

I’m currently wondering to what extent it could be relevant to create a level editor inside the game, in order to have people create their own levels… Want to share your opinion? πŸ™‚