diff --git a/render_dos.c b/render_dos.c
index cddcb31..c1bc1e9 100644
--- a/render_dos.c
+++ b/render_dos.c
@@ -15,6 +15,8 @@ struct Context
     uint8_t *backbuffer;
     uint8_t *colorbuffer;
     float *depthbuffer;
+    uint8_t *texture;
+    float lightdir[3];
 };
 
 static uint8_t *vga = (uint8_t *)0xA0000;
@@ -154,8 +156,8 @@ static void MapToRes(struct Context *ctx, float a[])
     a[1] = a[1] / a[3];
     a[2] = a[2] / a[3];
 
-    a[0] = (a[0] + 1.0) * 0.5 * (ctx->width - 1);
-    a[1] = (1.0 - a[1]) * 0.5 * (ctx->height - 1);
+    a[0] = (a[0] + 1.0f) * 0.5f * (ctx->width - 1);
+    a[1] = (1.0f - a[1]) * 0.5f * (ctx->height - 1);
 }
 
 static void Cross(float result[], float a[], float b[])
@@ -279,137 +281,248 @@ static void Perspective(float mat[], float fov, float aspectratio, float nearp,
     mat[15] = 0.0f;
 }
 
-static void DrawTriangle(struct Context *ctx, float wp0[], float wp1[], float wp2[], uint8_t texture[], float c0[], float c1[], float c2[], float mvp[], float lightdir[], float n0[], float n1[], float n2[])
+struct ScrVertex
 {
-    float p0[4], p1[4], p2[4];
+    float x, y;
+    float t[2];
+    float n[3];
+    float depth;
+};
 
-    MatMultVec3(p0, mvp, wp0);
-    MatMultVec3(p1, mvp, wp1);
-    MatMultVec3(p2, mvp, wp2);
+static void FillScanLine(struct Context *ctx, const struct ScrVertex *v0, const struct ScrVertex *v1)
+{
+    int x, y, start, end, t0x, t0y, t1x, t1y, tx, ty;
+    float c[3], n[3], depth, *depthbuf, xdiff, depthdiff, tdiff[2], ndiff[3], diffuse;
+    uint8_t *texel;
+
+
+    start = v0->x;
+    if (start < 0)
+        start = 0;
+
+    end = v1->x;
+    if (end > ctx->width)
+        end = ctx->width;
+
+    y = v0->y;
     
-    if (p0[3] != 0.0f && p1[3] != 0.0f && p2[3] != 0.0f)
+    xdiff = v1->x - v0->x;
+    depthdiff = v1->depth - v0->depth;
+    tdiff[0] = v1->t[0] - v0->t[0];
+    tdiff[1] = v1->t[1] - v0->t[1];
+    
+    ndiff[0] = v1->n[0] - v0->n[0];
+    ndiff[1] = v1->n[1] - v0->n[1];
+    ndiff[2] = v1->n[2] - v0->n[2];
+
+
+    for (x = start; x < end; ++x)
     {
-        int x, y, idx, tx, ty, ti;
-        float p[2];
-        float depth;
-        float w0, w1, w2, dot00, dot01, dot11, dot20, dot21, denom, d0, d1, d2;
-        float v0[2], v1[2], v2[2];
-        float texcoord[2];
-        float normal[3];
-        float diff;
-        int aabbmin[2], aabbmax[2];
-        float c[3];
+        float t = ((float)x - v0->x) / xdiff;
+        depth = v0->depth + t * depthdiff;
 
-        MapToRes(ctx, p0);
-        MapToRes(ctx, p1);
-        MapToRes(ctx, p2);
-    
-        aabbmin[0] = Max(Min3(p0[0], p1[0], p2[0]), 0.0f);
-        aabbmin[1] = Max(Min3(p0[1], p1[1], p2[1]), 0.0f);
-        aabbmax[0] = Min(Max3(p0[0], p1[0], p2[0]), ctx->width - 1);
-        aabbmax[1] = Min(Max3(p0[1], p1[1], p2[1]), ctx->height - 1);
+        depthbuf = &ctx->depthbuffer[x + y * ctx->width];
+        if (*depthbuf < depth)
+            continue;
 
-        for (y = aabbmin[1]; y <= aabbmax[1]; ++y)
-        {
-            for (x = aabbmin[0]; x <= aabbmax[0]; ++x)
-            {
-                p[0] = x;
-                p[1] = y;
+        *depthbuf = depth;
+        
+        tx = (v0->t[0] + t * tdiff[0]) * ctx->tex_w;
+        ty = (1.0f - (v0->t[1] + t * tdiff[1])) * ctx->tex_h;
 
-                if (!IsOnRight(p0, p1, p) || !IsOnRight(p1, p2, p) || !IsOnRight(p2, p0, p))
-                    continue;
+        if (tx < 0)
+            tx = 0;
+        else if (tx >= ctx->width)
+            tx = ctx->width - 1;
 
-                idx = x + ctx->width * y;
-                
-                v0[0] = p1[0] - p0[0];
-                v0[1] = p1[1] - p0[1];
-                v1[0] = p2[0] - p0[0];
-                v1[1] = p2[1] - p0[1];
-                v2[0] = p[0] - p0[0];
-                v2[1] = p[1] - p0[1];
+        if (ty < 0)
+            ty = 0;
+        else if (ty >= ctx->height)
+            ty = ctx->height - 1;
 
-                dot00 = Dot2(v0, v0);
-                dot01 = Dot2(v0, v1);
-                dot11 = Dot2(v1, v1);
-                dot20 = Dot2(v2, v0);
-                dot21 = Dot2(v2, v1);
+        n[0] = v0->n[0] + t * ndiff[0];
+        n[1] = v0->n[1] + t * ndiff[1];
+        n[2] = v0->n[2] + t * ndiff[2];
 
-                denom = dot00 * dot11 - dot01 * dot01;
-                w1 = (dot11 * dot20 - dot01 * dot21) / denom;
-                w2 = (dot00 * dot21 - dot01 * dot20) / denom;
-                w0 = 1.0f - w1 - w2;
+        Normalize(n);
+        diffuse = Max(0.5, Dot(n, ctx->lightdir));
 
-                depth = p0[2] * w0 + p1[2] * w1 + p2[2] * w2;
+        texel = &ctx->texture[(tx + ty * ctx->tex_w) * 3];
+        c[0] = texel[0] / 255.0f * diffuse;
+        c[1] = texel[1] / 255.0f * diffuse;
+        c[2] = texel[2] / 255.0f * diffuse;
 
-                if (depth > 0.0 && ctx->depthbuffer[idx] > depth)
-                {
-                    ctx->depthbuffer[idx] = depth;
-                    idx = idx * 3;
-
-                    texcoord[0] = w0 * c0[0] + w1 * c1[0] + w2 * c2[0];
-                    texcoord[1] = w0 * c0[1] + w1 * c1[1] + w2 * c2[1];
-
-                    normal[0] = w0 * n0[0] + w1 * n1[0] + w2 * n2[0];
-                    normal[1] = w0 * n0[1] + w1 * n1[1] + w2 * n2[1];
-                    normal[2] = w0 * n0[2] + w1 * n1[2] + w2 * n2[2];
-                    
-                    Normalize(normal);
-                    diff = Max(0.5, Dot(normal, lightdir));
-
-                    tx = texcoord[0] * ctx->tex_w;
-                    ty = (1 - texcoord[1]) * ctx->tex_h;
-                    ti = (tx + ty * ctx->tex_w) * 3;
-
-                    c[0] = texture[ti] / 255.0f * diff;
-                    c[1] = texture[ti + 1] / 255.0f * diff;
-                    c[2] = texture[ti + 2] / 255.0f * diff;
-                
-                    SetPixel(ctx, x, y, c);
-                }                
-            }
-        }
+        SetPixel(ctx, x, y, c);
     }
 }
 
-// static void OutputPPM(FILE *file, struct Context *ctx)
-// {
-//     int i, numpixels;
+static void SortByY(struct ScrVertex** s)
+{
+    struct ScrVertex* temp;
+
+    if (s[0]->y > s[1]->y)
+    {
+        temp = s[0];
+        s[0] = s[1];
+        s[1] = temp;
+    }
+
+    if (s[1]->y > s[2]->y)
+    {
+        temp = s[1];
+        s[1] = s[2];
+        s[2] = temp;
+    }
+
+    if (s[0]->y > s[1]->y)
+    {
+        temp = s[0];
+        s[0] = s[1];
+        s[1] = temp;
+    }
+}
+
+static void DebugDrawPoint(struct Context *ctx, int x, int y)
+{
+    float c[] = { 1.0f, 0.0f, 0.0f };
+
+    if (x < 0 || x >= ctx->width || y < 0 || y >= ctx->height)
+        return;
+
+    SetPixel(ctx, x, y, c);
+}
+
+static void LerpScrVertex(const struct ScrVertex *a, const struct ScrVertex *b, struct ScrVertex *res, float t)
+{
+    res->x = a->x + (b->x - a->x) * t;
+    res->y = a->y + (b->y - a->y) * t;
+    res->t[0] = a->t[0] + (b->t[0] - a->t[0]) * t;
+    res->t[1] = a->t[1] + (b->t[1] - a->t[1]) * t;
+    res->n[0] = a->n[0] + (b->n[0] - a->n[0]) * t;
+    res->n[1] = a->n[1] + (b->n[1] - a->n[1]) * t;
+    res->n[2] = a->n[2] + (b->n[2] - a->n[2]) * t;
+    res->depth = a->depth + (b->depth - a->depth) * t;
+}
+
+static void DrawTriangle(struct Context *ctx, float wp0[], float wp1[], float wp2[], float c0[], float c1[], float c2[], float mvp[], float n0[], float n1[], float n2[])
+{
+    float p[3][4];
+    struct ScrVertex v[3], nv, *s[4], *tempv;
+    int i;
+    float nf;
+
+    MatMultVec3(p[0], mvp, wp0);
+    MatMultVec3(p[1], mvp, wp1);
+    MatMultVec3(p[2], mvp, wp2);
+
+    for (i = 0; i < 3; ++i)
+    {
+        if (p[i][3] == 0.0f)
+            return;
+
+        MapToRes(ctx, p[i]);
+
+        v[i].x = p[i][0];
+        v[i].y = p[i][1];
+        v[i].depth = p[i][2];
+    }
+
+    v[0].t[0] = c0[0];
+    v[0].t[1] = c0[1];
     
-//     fprintf(file, "P3\n%d\n%d\n255\n", ctx->width, ctx->height);
-
-//     numpixels = ctx->width * ctx->height;
+    v[1].t[0] = c1[0];
+    v[1].t[1] = c1[1];
     
-//     for (i = 0; i < numpixels * 3; ++i)
-//     {
-//         fprintf(file, "%d ", (int)(ctx->colorbuffer[i] * 255.0f));
-//     }
+    v[2].t[0] = c2[0];
+    v[2].t[1] = c2[1];
 
-//     fprintf(file, "\n");
-// }
+    v[0].n[0] = n0[0];
+    v[0].n[1] = n0[1];
+    v[0].n[2] = n0[2];
 
-// static void OutputVT100(FILE* out, const struct Context *ctx)
-// {
-//     int r1, g1, b1, r2, b2, g2, x, y, idx, rowoffset;
-//     fprintf(out, "\x1b[1;1H\n");
+    v[1].n[0] = n1[0];
+    v[1].n[1] = n1[1];
+    v[1].n[2] = n1[2];
+    
+    v[2].n[0] = n2[0];
+    v[2].n[1] = n2[1];
+    v[2].n[2] = n2[2];
 
-//     for (y = 0; y <= ctx->height - 2; y += 2)
-//     {
-//         fprintf(out, "\n");
-//         for (x = 0; x < ctx->width; ++x)
-//         {
-//             idx = (x + y * ctx->width) * 3;
-//             rowoffset = ctx->width * 3;
-//             r1 = ctx->colorbuffer[idx] * 255;
-//             g1 = ctx->colorbuffer[idx + 1] * 255;
-//             b1 = ctx->colorbuffer[idx + 2] * 255;
-//             r2 = ctx->colorbuffer[idx       + rowoffset] * 255;
-//             g2 = ctx->colorbuffer[idx + 1   + rowoffset] * 255;
-//             b2 = ctx->colorbuffer[idx + 2   + rowoffset] * 255;
+    if (IsOnRight(p[0], p[2], p[1]))
+        return; /* backface cull */
 
-//             fprintf(out, "\x1b[38;2;%d;%d;%dm\x1b[48;2;%d;%d;%dm▀", r1, g1, b1, r2, g2, b2);
-//         }
-//     }
-// }
+    s[0] = &v[0];
+    s[1] = &v[1];
+    s[2] = &v[2];
+    s[3] = &nv;
+    SortByY(s);
+    
+    nf = (s[1]->y - s[0]->y) / (s[2]->y - s[0]->y);
+    LerpScrVertex(s[0], s[2], &nv, nf);
+    nv.y = s[1]->y;
+    
+    if (s[3]->x < s[1]->x)
+    {
+        tempv = s[1];
+        s[1] = s[3];
+        s[3] = tempv;
+    }
+    
+    // for (i = 0; i < 4; ++i)
+    // {
+    //     DebugDrawPoint(ctx, s[i]->x, s[i]->y);
+    // }
+
+    if (s[0]->y != s[1]->y)
+    {
+        int start, end, sc_y;
+
+        start = s[0]->y;
+        if (start < 0)
+            start = 0;
+
+        end = s[1]->y + 1;
+        if (end > ctx->height)
+            end = ctx->height;
+
+        for (sc_y = start; sc_y < end; sc_y++)
+        {
+            struct ScrVertex scv[2];
+            float t = ((float)sc_y - s[0]->y) / (s[1]->y - s[0]->y);
+
+            LerpScrVertex(s[0], s[1], &scv[0], t);
+            LerpScrVertex(s[0], s[3], &scv[1], t);
+
+            scv[0].y = sc_y;
+            FillScanLine(ctx, &scv[0], &scv[1]);
+        }
+    }
+
+    if (s[1]->y != s[2]->y)
+    {
+        int start, end, sc_y;
+
+        start = s[3]->y + 1;
+        if (start < 0)
+            start = 0;
+
+        end = s[2]->y + 1;
+        if (end > ctx->height)
+            end = ctx->height;
+
+        for (sc_y = start; sc_y < end; sc_y++)
+        {
+            struct ScrVertex scv[2];
+            float t = ((float)sc_y - s[3]->y) / (s[2]->y - s[3]->y);
+
+            LerpScrVertex(s[1], s[2], &scv[0], t);
+            LerpScrVertex(s[3], s[2], &scv[1], t);
+            
+            scv[0].y = sc_y;
+            FillScanLine(ctx, &scv[0], &scv[1]);
+        }
+    }
+}
 
 static void OutputVGA(const struct Context *ctx)
 {
@@ -417,16 +530,6 @@ static void OutputVGA(const struct Context *ctx)
         memcpy(vga, ctx->colorbuffer, 320 * 200);
 }
 
-// Function DumpMat(Real Array mat)
-//     Declare Integer i
-//     For i = 0 to 15
-//         Output mat[i] & " "
-//         If i % 4 == 3
-//             Output ""
-//         End
-//     End
-// End
-
 static void *MyMalloc(size_t n)
 {
     void *mem = malloc(n);
@@ -443,9 +546,6 @@ int main(int argc, char **argv)
 {
     struct Context ctx;
 
-    // float bg[] = { 0.25f, 0.25f, 0.25f };
-    float lightdir[] = { 0.3f, 1.0f, 0.4f };
-    
     float view[16];
     float proj[16];
     float mvp[16];
@@ -454,7 +554,7 @@ int main(int argc, char **argv)
     float targetpos[] = { 0.0f, 0.0f, 0.0f };
     float upvector[] = { 0.0f, 1.0f, 0.0f };
 
-    float aspectratio = 1.46;
+    float aspectratio = 1.3f;
 
     float p0[3], p1[3], p2[3];
     float c0[3], c1[3], c2[3];
@@ -462,7 +562,6 @@ int main(int argc, char **argv)
 
     int datasize, i, vertices, tp;
     float *data;
-    uint8_t *texture;
 
     int temp, i0, i1, i2;
     float camangle;
@@ -470,11 +569,6 @@ int main(int argc, char **argv)
     FILE *in = stdin;
     FILE *out = stdout;
 
-    // getch();
-
-    // SetMode(0x3);
-    // return 0;
-
     if (argc > 1)
     {
         in = fopen(argv[1], "r");
@@ -503,7 +597,10 @@ int main(int argc, char **argv)
 
     ctx.colorbuffer = ctx.backbuffer;
 
-    Normalize(lightdir);
+    ctx.lightdir[0] = 0.3f;
+    ctx.lightdir[1] = 1.0f;
+    ctx.lightdir[2] = 0.4f;
+    Normalize(ctx.lightdir);
     
     Perspective(proj, DegToRad(90.0f) / aspectratio, aspectratio, 0.1f, 50.0f);
 
@@ -520,13 +617,13 @@ int main(int argc, char **argv)
     fscanf(in, "%d\n%d\n", &ctx.tex_w, &ctx.tex_h);
     tp = ctx.tex_w * ctx.tex_h * 3;
 
-    texture = MyMalloc(tp);
+    ctx.texture = MyMalloc(tp);
 
     for (i = 0; i < tp; ++i)
     {
         int v;
         fscanf(in, "%d\n", &v);
-        texture[i] = v;
+        ctx.texture[i] = v;
     }
 
     fprintf(stderr, "textura rdy\n");
@@ -592,16 +689,12 @@ int main(int argc, char **argv)
             n2[1] = data[i2 + 6];
             n2[2] = data[i2 + 7];
     
-            DrawTriangle(&ctx, p0, p1, p2, texture, c0, c1, c2, mvp, lightdir, n0, n1, n2);
+            DrawTriangle(&ctx, p0, p1, p2, c0, c1, c2, mvp, n0, n1, n2);
         }
     
-    
-        // OutputPPM(out, &ctx);
-        // OutputVT100(out, &ctx);
         OutputVGA(&ctx);
     }
 
-
     return 0;
 
 }