diff --git a/Cargo.toml b/Cargo.toml
index 55130f4..9ecf5e2 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -4,9 +4,6 @@ version = "0.1.0"
 edition = "2024"
 
 [dependencies]
-noise = "0.9.0"
-image = "0.25.5"
-rayon = "1.10.0"
 [lib]
 name = "worldgen"
 path = "src/lib.rs"
diff --git a/src/.vscode/launch.json b/src/.vscode/launch.json
new file mode 100644
index 0000000..10efcb2
--- /dev/null
+++ b/src/.vscode/launch.json
@@ -0,0 +1,16 @@
+{
+    // Use IntelliSense to learn about possible attributes.
+    // Hover to view descriptions of existing attributes.
+    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "type": "lldb",
+            "request": "launch",
+            "name": "Debug",
+            "program": "${workspaceFolder}/<executable file>",
+            "args": [],
+            "cwd": "${workspaceFolder}"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/src/lib.rs b/src/lib.rs
index 9a2da68..078ba93 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -1,59 +1,70 @@
-use noise::{Fbm, MultiFractal, NoiseFn, Perlin};
-
-/// Generates a 16x16 noise map for a given chunk using fractal noise with multiple octaves
-/// to avoid repetitive patterns.
-///
-/// # Arguments
-/// - `seed`: A number influencing the noise generation (e.g., 42).
-/// - `chunk_x`: The X coordinate of the chunk.
-/// - `chunk_z`: The Z coordinate of the chunk.
-/// - `scale`: The noise scale (smaller = more detailed noise).
-///
-/// # Returns
-/// - A **16x16 noise map** as `[[f64; 16]; 16]`, normalized between -64 and 324.
-pub fn generate_normalized_noise_map(
-    seed: u32,
-    chunk_x: i32,
-    chunk_z: i32,
+pub struct Noise {
     scale: f64,
-) -> [[f64; 16]; 16] {
-    // Spécifier explicitement que Fbm utilise Perlin comme bruit de base
-    let fbm = Fbm::<Perlin>::new(seed).set_octaves(15);
-    let mut noise_map = [[0.0; 16]; 16];
-
-    // Define normalization range
-    let min_range = -64.0;
-    let max_range = 320.0;
-
-    for x in 0..16 {
-        for z in 0..16 {
-            // Convert chunk-local coordinates to global world coordinates
-            let world_x = (chunk_x * 16 + x as i32) as f64 * scale;
-            let world_z = (chunk_z * 16 + z as i32) as f64 * scale;
-
-            // Generate fractal noise value (with multiple octaves)
-            let noise_value = fbm.get([world_x, world_z]);
-
-            // Normalize noise from [-1,1] to [-64,324]
-            let normalized_noise = (noise_value + 1.0) / 2.0 * (max_range - min_range) + min_range;
-
-            // Store the normalized noise value
-            noise_map[x][z] = normalized_noise;
-        }
-    }
-    noise_map
+    amplitude: f64,
+}
+pub struct Vector {
+    x: f32,
+    y: f32,
+}
+impl Noise {
+    pub fn new(scale: f64, amplitude: f64) -> Self {
+        Self { scale, amplitude }
+    }
+    pub fn get(&self, x: f64, z: f64) -> f64 {
+        let xs = x / self.scale;
+        let zs = z / self.scale;
+        self.perlin(xs, zs) * self.amplitude
+    }
+    fn perlin(&self, x: f64, z: f64) -> f64 {
+        // implement Perlin noise here (then simplex because it's harder)
+        todo!()
+    }
+}
+fn dot_product(v1: Vector, v2: Vector) -> f32 {
+    // Calculate the dot product between v1 and v2 using their coordinates.  the result->f32.
+    v1.x * v2.x + v1.y * v2.y
+}
+fn calculate_norm(v1: &Vector) -> f32 {
+    // Calculate the norm of a vector using it's coordinates.  the result -> f32.
+    (v1.x.powi(2) + v1.y.powi(2)).sqrt()
+}
+fn normalize(v1: &Vector) -> Vector {
+    // This function aim that every vector created randomly has the same norm (1).
+    Vector {
+        x: v1.x / calculate_norm(v1),
+        y: v1.y / calculate_norm(v1),
+    }
 }
-
 #[cfg(test)]
 mod tests {
-    use super::*;
-
+    use crate::{Vector, calculate_norm, dot_product, normalize};
     #[test]
-    fn test_generate_normalized_noise_map() {
-        let noise_map = generate_normalized_noise_map(456, 0, 0, 0.1);
-        assert_eq!(noise_map.len(), 16);
-        for row in noise_map.iter() {
-            assert_eq!(row.len(), 16);
-        }
+    fn test_dot_product() {
+        assert_eq!(
+            dot_product(Vector { x: 1.0, y: 0.0 }, Vector { x: 0.0, y: 1.0 }),
+            0.0
+        );
+        assert_eq!(
+            dot_product(Vector { x: 1.0, y: 0.5 }, Vector { x: 0.2, y: 1.0 }),
+            0.7,
+        );
+        assert_eq!(
+            dot_product(Vector { x: 1.0, y: 0.5 }, Vector { x: -0.2, y: -1.0 }),
+            -0.7,
+        );
+    }
+    #[test]
+    fn test_calculate_norm() {
+        assert_eq!(calculate_norm(&Vector { x: 0.5, y: 0.5 }), 0.5_f32.sqrt());
+        assert_eq!(calculate_norm(&Vector { x: 0.7, y: 0.3 }), 0.58_f32.sqrt());
+        assert_eq!(
+            calculate_norm(&Vector { x: -0.7, y: -0.3 }),
+            calculate_norm(&Vector { x: 0.7, y: 0.3 })
+        );
+    }
+    #[test]
+    fn test_normalize() {
+        let v1 = Vector { x: 0.5, y: 0.5 };
+        assert_eq!(calculate_norm(&normalize(&v1)).round(), 1_f32);
     }
 }
diff --git a/src/main.rs b/src/main.rs
deleted file mode 100644
index 00c2ec6..0000000
--- a/src/main.rs
+++ /dev/null
@@ -1,69 +0,0 @@
-use image::{GrayImage, Luma};
-use rayon::prelude::*;
-use worldgen::generate_normalized_noise_map; // Import rayon for parallel processing
-
-fn main() {
-    // Define parameters
-    let radius = 300;
-    let chunk_size = 16;
-    let num_chunks = 2 * radius + 1; // number of chunks per dimension (65 here)
-    let img_width = num_chunks * chunk_size;
-    let img_height = num_chunks * chunk_size;
-
-    // Generate all chunk coordinates from -radius to +radius
-    let chunk_coords: Vec<(i32, i32)> = (-radius..=radius)
-        .flat_map(|cx| (-radius..=radius).map(move |cz| (cx, cz)))
-        .collect();
-
-    // Compute noise maps for each chunk in parallel using Rayon
-    let chunk_results: Vec<(i32, i32, [[f64; 16]; 16])> = chunk_coords
-        .into_par_iter()
-        .map(|(cx, cz)| {
-            let noise_map = generate_normalized_noise_map(42, cx, cz, 0.001);
-            (cx, cz, noise_map)
-        })
-        .collect();
-
-    // Create the big image with the appropriate dimensions
-    let mut big_img: GrayImage = GrayImage::new(img_width as u32, img_height as u32);
-
-    // Write each chunk's noise map into the global image at the corresponding position
-    for (cx, cz, noise_map) in chunk_results {
-        // Compute pixel offset: we shift de coordonnées de chunk pour obtenir des indices positifs
-        let offset_x = ((cx + radius) * chunk_size) as u32;
-        let offset_y = ((cz + radius) * chunk_size) as u32;
-        write_chunk_to_image(&mut big_img, offset_x, offset_y, noise_map);
-    }
-
-    // Save the generated image to a file
-    big_img
-        .save("big_noise_map.png")
-        .expect("Failed to save image");
-    println!("Image saved as big_noise_map.png");
-}
-
-/// Writes a 16x16 chunk noise map into the provided image at the specified offset.
-/// Noise values are normalized from the range [-64, 324] to [0, 255].
-fn write_chunk_to_image(
-    img: &mut GrayImage,
-    offset_x: u32,
-    offset_y: u32,
-    noise_map: [[f64; 16]; 16],
-) {
-    let min_val = -64.0;
-    let max_val = 324.0;
-    let scale = 255.0 / (max_val - min_val);
-
-    // Iterate over each pixel in the 16x16 noise map
-    for (row_index, row) in noise_map.iter().enumerate() {
-        for (col_index, &value) in row.iter().enumerate() {
-            // Map noise value from [-64, 324] to [0, 255]
-            let pixel_value = (((value - min_val) * scale).round() as u8).min(255);
-            img.put_pixel(
-                offset_x + col_index as u32,
-                offset_y + row_index as u32,
-                Luma([pixel_value]),
-            );
-        }
-    }
-}