Kill times...

The terrain looks better now.

The terrain 3D model data can be generated by Truespace.

Install MAC OS on my desktop.

Finally, I found MAC OS for my Intel CPU based computer which called Ledpard. The install file can be found through torrent. The install process is quite simple and easy but the problem is hard to look for the appropriate one.

What I am doing now?

It is a good chance to candidates who may learn about 3D graphic tools without any payment.

There are many free 3D graphic tools, such as, Google Sketch up, Blender, K3D, MeshLab, etc.

There is one more tool called Truespace which I recently have installed on my desktop. I could draw simple model mush easier.

And of course, I could get the model values using collada dae file format which is xml based text. The above shown picture is a simple plane map editer shows plane lands and fast speed of FPS, etc.

Drawing ground with texture.

The result looks more realistic of the ground. The texture is 16 bits color 565(RGB) bitmap.

If anyone want to try the texture with 16 bits 565 bitmap may use the OPENGL function follow.

Ex) glTexImage2D(GL_TEXTURE_2D, 0, GL_RGB5, width, height, 0, GL_RGB, GL_UNSIGNED_SHORT_5_6_5, data);

Here is no activity for a long time!

This post has not a special content. I wish continue my activity which is related with mobile application development.
I am still using old laptop note book and the new desktop which i was bought recently is next with my laptop note book. The desktop run so fast and has more screen capacity. Run to make my new application....

Creating 3d data values for drawing ground

To create 3d data values for drawing ground have to calculate from center (0f, 0f, 0f). The left picture shows the four parts which are divided by center.
Every x and y values are increased and decreased by 1.0 or -1.0 except z values. The increase and decrease values have a same pattern. According many OpenGL examples from Internet, we can draw rectangle with 4 vertex points. Each vertex point has x, y and z value. Therefore, the program has to generate rectangle points’ value. Of course, the generated file has also color values.

int main(int argc, char **argv)
{
int nwidth = atoi(argv[1]);
int nheight = atoi(argv[2]);

fnCVertextFile(nheight, nwidth);
}

Example) program 8 8
When the program was executed like an example, then the vector values will be generated 16 * 16 sizes.

typedef struct {
float x, y, z;
} M3DPoint;

typedef struct {
M3DPoint left, top, right, bottom;
} M3DRect;

// the function writes structure values and string values to file.
void fprintfex(FILE *fvfile, M3DRect row1, char *value)
{
fprintf(fvfile, "%f %f %f %s\n", row1.left.x, row1.left.y, row1.left.z, value);
fprintf(fvfile, "%f %f %f %s\n", row1.top.x, row1.top.y, row1.top.z, value);
fprintf(fvfile, "%f %f %f %s\n", row1.right.x, row1.right.y, row1.right.z, value);
fprintf(fvfile, "%f %f %f %s\n", row1.bottom.x, row1.bottom.y, row1.bottom.z, value);
}

void fnCVertextFile(int aRow, int aColumn)
{
M3DRect row1;

float nleft = 0.0f;
float ntop = 0.0f;

char value1[] = {"0.3 0.3 0.3"};
char value2[] = {"0.5 0.5 0.5"};
char *value;
bool toggle = false;

FILE *fvfile = fopen("vector.dat", "w");
if(fvfile == NULL) {
printf("file cannot open!\n");
exit(-1);
}

// count of vertex
fprintf(fvfile, "%d\n", aRow * aColumn * 4 * 4);

for(int j = 0; j < aRow; j++)
{
for(int i = 0; i < aColumn; i++)
{
row1.left.x = nleft; row1.left.y = ntop; row1.left.z = 1.5f;
row1.top.x = nleft+1.0f; row1.top.y = ntop; row1.top.z = 1.5f;
row1.right.x = nleft+1.0f; row1.right.y = ntop+1.0f; row1.right.z = 1.5f;
row1.bottom.x = nleft; row1.bottom.y = ntop+1.0f; row1.bottom.z = 1.5f;
value = (toggle) ? value1 : value2;
fprintfex(fvfile, row1, value);
toggle = !toggle;
nleft += 1.0f;
}
ntop += 1.0f;
nleft = 0.0f;
toggle = !toggle;
}

// was ommited below section of source
…

Now, my second app was included some MIDI sounds.

Until I get Android test device, The second app will be like this.

Qt4.5.2-tower

Recently, Nokia published the new version of Qt4.5.2-tower which can be operated in many platforms. Previously, I installed the Qt4.5.0-garden for s60 and windows platforms. A couple of days ago, I tried to install S60 5th edition SDK and s60 3rd Feature pack II. I already had N97 0.5 version SDK. The S60 5th edition emulator provided a couple of resolution of mobile screen. So, I uninstalled 3rd edition and N97 edition SDK. Yesterday, I downloaded the Qt4.5.2-tower and I did install process into my old laptop notebook. The install required many things to install the Qt4.5.2-tower. I completed the install process yesterday night due to what I got. I like 2d graphics API. It easily use and so powerful. I am not sure now how an application install into real device. Since I like the emulator and I can launch my testing application into the emulator.

Qt4.5.0-Garden

For Nokia Application developers, Now They have a little strong development toolkit. ‘Qt4.5.0-garden’ which supports library for s60 mobile makes low a barrier then before. It means S60 mobile application developers can develop application more easy. For me, I was trying to learn about Carbide C++ and S60 library but it gave to me very complex. But Qt garden takes away some complex. To launch an application under S60 emulator, it is very simple. The picture shows my first trying of tutorial which was presented from qt garden prereleased web site. At the moment, I only use text editor called Notepad++, Dos command for building tutorial C++ code and emulator. Of course, there are some things should installed before use Qt garden.

Translation bitmap image

Bitmap image contains BitmapHeader, pixels data and palette. The palette only need less than 16 bit color bitmap image, so This article will not include. Microsoft Visual studio has BitmapHeader structure in some header files. It also provides some functions which often give developer difficult to understand. Anyone who may has basic knowledge of c++ and the structure of bitmap file will has no problem with this article. The aritcle will touch only Device Independent Bitmap and pixels data are not compressed.

Code begining here // ->>
#include <stdio.h>
#include <stdlib.h>

#pragma pack(1) // get the structure size 54 not 56

typedef struct {
unsigned char bfType1; // fixed 'B'
unsigned char bfType2; // fixed 'M'
unsigned int bfSize; // Size of File. Zero is acceptable
unsigned short bfReserved1; // Zero
unsigned short bfReserved2; // Zero
unsigned int bfOffBits; // Offset to beginning of bitmap
unsigned int biSize; // Number of Bytes in Structure
unsigned int biWidth; // width of bitmap in pixels
unsigned int biHeight; // height of bitmap in pixels
unsigned short biPlanes; // Planes in target device = 1
unsigned short biBitCount; // Bits per Pixel 1, 4, 8, 16, 24, 32
unsigned int biCompression; // BI_RGB = 0, BI_RLE8, BI_RLE4
unsigned int biSizeImage; // Size of Image Part (often ignored)
unsigned int biXPelsPerMeter; // Always Zero
unsigned int biYPelsPerMeter; // Always Zero
unsigned int biClrUsed; // Number of Colors used in Palette
unsigned int biClrImportant; // Number of Colors that are Important
} TBmpHeader;
<<- // Code end here

The pixels data fills the end of bitmap file after bitmap header. The 1 point pixel data consists of red, green and blue color value each. The each color value has the size of 8 bit.

This is a structure for 24 bit color.

Code begining here // ->>
typedef struct {
unsigned char rgbBlue; //intensity of blue in the color
unsigned char rgbGreen; //intensity of green in the color
unsigned char rgbRed; //intensity of red in the color
} TRGB;
<<- // Code end here

32 bit color adds alpha color value which sometimes use calcuration for mixing two images.

Code begining here // ->>
typedef struct {
unsigned char rgbBlue; //intensity of blue in the color
unsigned char rgbGreen; //intensity of green in the color
unsigned char rgbRed; //intensity of red in the color
unsigned char rgbAlpha; //intensity of Alpha in the color
} TRGBA;
<<- // Code end here

This routine fills bitmapheader structure by using width, height and color bit. The third argument is color type. It can use with 16, 24 and 32 value.

Code begin here // ->>
TBmpHeader FillBitmapHeader(unsigned int width, unsigned int height,
unsigned short bitmapType)
{
TBmpHeader MyBitmapHeader;

MyBitmapHeader.bfType1 = 0x42;
MyBitmapHeader.bfType2 = 0x4D;
MyBitmapHeader.bfSize = 0x00000000;
MyBitmapHeader.bfReserved1 = 0x0000;
MyBitmapHeader.bfReserved2 = 0x0000;
MyBitmapHeader.bfOffBits = 0x36; // the location of bitmap data
MyBitmapHeader.biSize = 0x28; // Size of the BItmapHeader

MyBitmapHeader.biWidth = width;
MyBitmapHeader.biHeight = height;
MyBitmapHeader.biPlanes = 0x0001; // Plane target device 1

MyBitmapHeader.biBitCount = bitmapType; // Bits per Pixel 1, 4, 8, 16, 24, 32
MyBitmapHeader.biCompression = 0x00000000; // BI_RGB = 0, BI_RLE8, BI_RLE4

MyBitmapHeader.biSizeImage = 0x00000000;
MyBitmapHeader.biXPelsPerMeter = 0x00000000;
MyBitmapHeader.biYPelsPerMeter = 0x00000000;
MyBitmapHeader.biClrUsed = 0x00000000;
MyBitmapHeader.biClrImportant = 0x00000000;

return MyBitmapHeader;
}
<<- // Code end here

16 bit color can consist 32 bit color but each RGBA color has to decrease there own size half. Therefore, the result image can not avoid color loss.

Code begin here // ->>
void ConvertBmp32to16(const char *srcfname, const char *dstfname)
{
TBmpHeader MyBitmapHeader;
FILE *finput = fopen(srcfname, “rb”);
if(finput == NULL) return;
fread(&MyBitmapHeader, sizeof(TBmpHeader), 1, finput);
TRGBA * rgbdata;
unsigned short *rgbdata16;

int SSize = MyBitmapHeader.biWidth * MyBitmapHeader.biHeight;
int Size = SSize * 4;

rgbdata = (TRGBA *)malloc(Size);

fread(rgbdata, Size, 1, finput);
fclose(finput);

rgbdata16 = (unsigned short *)malloc(SSize*2);

unsigned short r, g, b, a;
for(int i=0; i < SSize; i++) {
b = rgbdata[i].rgbBlue >> 4;
g = rgbdata[i].rgbGreen >> 4;
r = rgbdata[i].rgbRed >> 4;
a = rgbdata[i].rgbAlpha >> 4;
rgbdata16[i] = (r << 12) + (g << 8) + (b << 4) + a;
}

MyBitmapHeader = FillBitmapHeader(MyBitmapHeader.biWidth, MyBitmapHeader.biHeight, 16);

FILE *fpout = fopen(dstfname, “wb”);
if(fpout == NULL) return;
fwrite(&MyBitmapHeader, sizeof(TBmpHeader), 1, fpout);
fwrite(rgbdata16, SSize*2, 1, fpout);
fclose(fpout);
free(rgbdata);
free(rgbdata16);
}
<<- // Code End here

After execute the ConvertBmp32to16() routine, a bitmap image will be generated. The aim of this routine is for keeping alpha value after translation. So, you don't have to confirm the generated image.

Code begin here // ->>
void ConvertBmp16to32(const char *srcfname, const char *dstfname)
{
TBmpHeader MyBitmapHeader;
FILE *finput = fopen(srcfname, “rb”);
if(finput == NULL) return;
fread(&MyBitmapHeader, sizeof(TBmpHeader), 1, finput);
TRGBA * rgbdata;
unsigned short *rgbdata16;

int SSize = MyBitmapHeader.biWidth * MyBitmapHeader.biHeight;
int Size = SSize * 2;

rgbdata16 = (unsigned short *)malloc(Size);

fread(rgbdata16, Size, 1, finput);
fclose(finput);

rgbdata = (TRGBA *)malloc(SSize*4);

unsigned short r, g, b, a;
for(int i=0; i < SSize; i++) {
b = (rgbdata16[i] >> 4) & 0x0f;
g = (rgbdata16[i] >> 8) & 0x0f;
r = (rgbdata16[i] >> 12) & 0x0f;
a = rgbdata16[i] & 0x0f;
rgbdata[i].rgbBlue = b << 4;
rgbdata[i].rgbGreen = g << 4;
rgbdata[i].rgbRed = r << 4;
rgbdata[i].rgbAlpha = a << 4;
}

MyBitmapHeader = FillBitmapHeader(MyBitmapHeader.biWidth, MyBitmapHeader.biHeight, 32);

FILE *fpout = fopen(dstfname, “wb”);
if(fpout == NULL) return;
fwrite(&MyBitmapHeader, sizeof(TBmpHeader), 1, fpout);
fwrite(rgbdata, SSize*4, 1, fpout);
fclose(fpout);
free(rgbdata);
free(rgbdata16);
}
<<- // Code end here

After execute the ConvertBmp16to32() routine, a bitmap image will also be generated. The image color shows
some color loss.

The shown below routine is translation routine which converts 24 bit color into 16 bit color bitmap.

Code begin here // ->>
void ConvertBmp24to16_555(const char *srcfname, const char *dstfname)
{
TBmpHeader MyBitmapHeader;
FILE *finput = fopen(srcfname, “rb”);
if(finput == NULL) return;
fread(&MyBitmapHeader, sizeof(TBmpHeader), 1, finput);
TRGB * rgbdata;
unsigned short *rgbdata16;

int SSize = MyBitmapHeader.biWidth * MyBitmapHeader.biHeight;
int Size = SSize * 3;

rgbdata = (TRGB *)malloc(Size);

fread(rgbdata, Size, 1, finput);
fclose(finput);

rgbdata16 = (unsigned short *)malloc(SSize*2);

unsigned short r, g, b, a;
for(int i=0; i < SSize; i++) {
b = rgbdata[i].rgbBlue >> 3;
g = rgbdata[i].rgbGreen >> 3;
r = rgbdata[i].rgbRed >> 3;
rgbdata16[i] = (r << 10) + (g << 5) + b;
}

MyBitmapHeader = FillBitmapHeader(MyBitmapHeader.biWidth, MyBitmapHeader.biHeight, 16);

FILE *fpout = fopen(dstfname, “wb&rdquo);
if(fpout == NULL) return;
fwrite(&MyBitmapHeader, sizeof(TBmpHeader), 1, fpout);
fwrite(rgbdata16, SSize*2, 1, fpout);
fclose(fpout);
free(rgbdata);
free(rgbdata16);
}
<<- // Code end here

Code begin here // ->>
int main(int argc, char **argv)
{
ConvertBmp32to16(“bitmap32.bmp”, “bitmap16r1.bmp”);
ConvertBmp16to32(“bitmap16r1.bmp”, “bitmap32r1.bmp”);
ConvertBmp24to16_555(“bitmap24.bmp”, “bitmap16r2.bmp”);
return 0;
}
<<- // Code end here

This will be my second Android game application...

Since, I am trying to complete this application. The application still miss some routines but the routines will be made soon.

Sphere on my Android emulator...

Both 'setModel()' and 'setModelEX()' methods work with making points of sphere surface. Now I get my drawing sphere routine. I recommend a URL which provides some glut functions that can be used under Android. akjava-android-project

class GLBall {
private int VertexCount, mallocCount;
private FloatBuffer mVertexBuffer;
private FloatBuffer mNormalBuffer;
public float z_angle;
private int slices, stacks;

public GLBall()
{
mVertexBuffer = null;
mNormalBuffer = null;
}

public void setModel(double radius, int lats, int longs)
{
double lat0, z0, zr0, lat1, z1, zr1, lng, x, y;
float normal[];
float vertices[];

int mindex = 0, nindex=0;

VertexCount = (lats+1) * (longs+1) * 2;
mallocCount = VertexCount * 3;

normal = new float[mallocCount];
vertices = new float[mallocCount];

stacks = lats+1;
slices = longs+1;

for(int i = 0; i <= lats; i++) {
lat0 = (double)Math.PI * (-0.5 + (i - 1) / (double)lats);
z0 = (double)Math.sin(lat0) * radius;
zr0 = (double)Math.cos(lat0) * radius;

lat1 = (double)Math.PI * (-0.5 + i / (double)lats);
z1 = (double)Math.sin(lat1) * radius;
zr1 = (double)Math.cos(lat1) * radius;
for(int j = 0; j <= longs; j++) {
lng = 2 * Math.PI * (j - 1) /(double)longs;
x = Math.cos(lng);
y = Math.sin(lng);
normal[nindex++] = vertices[mindex++] = (float)(x * zr0);
normal[nindex++] = vertices[mindex++] = (float)(y * zr0);
normal[nindex++] = vertices[mindex++] = (float)z0;
normal[nindex++] = vertices[mindex++] = (float)(x * zr1);
normal[nindex++] = vertices[mindex++] = (float)(y * zr1);
normal[nindex++] = vertices[mindex++] = (float)z1;
}
}
mVertexBuffer = FloatBuffer.wrap(vertices);
mNormalBuffer = FloatBuffer.wrap(normal);
z_angle = 0.0f;
}

public void setModelEX(float radius, int astacks, int aslices)
{
int i, j, vindex = 0, nindex = 0;
float slicestep, stackstep;

stackstep = ((float)Math.PI) / astacks;
slicestep = 2.0f * ((float)Math.PI) / aslices;

VertexCount = astacks * (aslices+1) * 2;
mallocCount = VertexCount * 3;

float normal[] = new float[mallocCount];
float vertices[] = new float[mallocCount];

for (i = 0; i < astacks; i++)
{
float a = i * stackstep;
float b = a + stackstep;

float s0 = (float)Math.sin(a);
float s1 = (float)Math.sin(b);

float c0 = (float)Math.cos(a);
float c1 = (float)Math.cos(b);

float nv;
for (j = 0; j <= aslices; j++)
{
float c = j * slicestep;
float x = (float)Math.cos(c);
float y = (float)Math.sin(c);

nv=x * s0;
normal[nindex++] = nv;
vertices[vindex++] = nv * radius;

nv=y * s0;
normal[nindex++] = nv;
vertices[vindex++] = nv * radius;

nv=c0;
normal[nindex++] = nv;
vertices[vindex++] = nv * radius;

nv=x * s1;
normal[nindex++] = nv;
vertices[vindex++] = nv * radius;

nv=y * s1;
normal[nindex++] = nv;
vertices[vindex++] = nv * radius;

nv=c1;
normal[nindex++] = nv;
vertices[vindex++] = nv * radius;
}
}
mVertexBuffer = FloatBuffer.wrap(vertices);
mNormalBuffer = FloatBuffer.wrap(normal);
z_angle = 0.0f;

this.stacks = astacks;
this.slices = aslices+1;
}

public void draw(GL10 gl) {
int triangles;
gl.glDisable(GL10.GL_TEXTURE_2D);
gl.glVertexPointer(3, GL10.GL_FLOAT, 0, mVertexBuffer);
gl.glNormalPointer(GL10.GL_FLOAT, 0, mNormalBuffer);

gl.glDisableClientState(GL10.GL_TEXTURE_COORD_ARRAY);
gl.glEnableClientState(GL10.GL_VERTEX_ARRAY);
gl.glEnableClientState(GL10.GL_NORMAL_ARRAY);

triangles = slices * 2;
for(int i=0; i < stacks; i++)
gl.glDrawArrays(GL10.GL_TRIANGLE_STRIP, i*triangles, triangles);

gl.glDisableClientState(GL10.GL_NORMAL_ARRAY);
gl.glEnableClientState(GL10.GL_TEXTURE_COORD_ARRAY);
gl.glEnable(GL10.GL_TEXTURE_2D);
}
}

Drawing sphere again

With Mingw I can use OpenGL function within main.cpp. I tried to use OPenGL function in my sphere.cpp but link error occured When i compiled and linked. So, I moved draw() function into my main.cpp. Once sphere
object is made, then the object will can share and will draw many with different color and angle.

main.cpp>

void drawSphere(TSphere sphere)
{
glEnableClientState(GL_VERTEX_ARRAY);
glEnableClientState(GL_NORMAL_ARRAY);
glVertexPointer(3, GL_FLOAT, 0, sphere.modelvalue);
glNormalPointer(GL_FLOAT, 0, sphere.normalvalue);
glDrawArrays(GL_QUAD_STRIP, 0, sphere.VertexCount);
glDisableClientState(GL_VERTEX_ARRAY);
glDisableClientState(GL_NORMAL_ARRAY);
}

sphere.h>

#ifndef __SPHERE_H__
#define __SPHERE_H__

#include
#include
#include

typedef struct {
float x, y, z;
} Vertex3D;

class TSphere {
public:
TSphere();
void SetModelValue(float radius, int slices, int stacks);
public:
int VertexCount;
Vertex3D *modelvalue;
Vertex3D *normalvalue;
};

#endif

sphere.cpp>

#include "sphere.h"

TSphere::TSphere()
{
//
}

void TSphere::SetModelValue(float radius, int slices, int stacks)
{
float lat0, z0, zr0, lat1, z1, zr1, x, y, lng;
int index = 0;

VertexCount = (slices+1) * (stacks+1) * 2;
modelvalue = new Vertex3D[VertexCount];
normalvalue = new Vertex3D[VertexCount];

for(int i = 0; i <= slices; i++) {
lat0 = M_PI * (-0.5 + (i - 1) / (float)slices);
z0 = sin(lat0) * radius;
zr0 = cos(lat0) * radius;

lat1 = M_PI * (-0.5 + i / (float)slices);
z1 = sin(lat1) * radius;
zr1 = cos(lat1) * radius;

for(int j = 0; j <= stacks; j++) {
lng = 2 * M_PI * (j - 1) / stacks;
x = cos(lng);
y = sin(lng);
modelvalue[index].x = x * zr0;
modelvalue[index].y = y * zr0;
modelvalue[index].z = z0;
normalvalue[index].x = x * zr0;
normalvalue[index].y = y * zr0;
normalvalue[index].z = z0;
index++;
modelvalue[index].x = x * zr1;
modelvalue[index].y = y * zr1;
modelvalue[index].z = z1;
normalvalue[index].x = x * zr1;
normalvalue[index].y = y * zr1;
normalvalue[index].z = z1;
index++;
}
}
}

Drawing sphere without glut...

Finally, I got a picture which is about drawing sphere. The routine does not contain glut function and neither glBegin() function. It is very simple to use. I would like to appreciate to http://ozark.hendrix.edu/~burch/cs/490/sched/feb8/. You can get c source code by visiting the url but u can see little bit different.

Probably, I will change this routine to c code and Android Java code ...

For example)

SPhere1 := TSPhere.Create;
SPhere1.SetModelValue(0.75, 8, 8);

in my render() routine...
begin
...
glLoadIdentity;
glTranslatef(-1.5, 0.0, -3.0);
glRotatef(Angle, 1.0, 1.0, -1.0);
glColor3f(0.9, 0.9, 0.9);
SPhere1.Draw();
glColor3f(1.0, 1.0, 1.0);
...
end;

type
TXCustomVertex = packed Record
fX, fY, fZ : single;
end;

TSphere = class
public
Constructor Create;
Procedure Draw();
procedure SetModelValue(r: single; lats, longs : Integer);
protected
private
clats, clongs, vertexCount : Integer;
ModelValue : Array of TXCustomVertex;
NormalValue : Array of TXCustomVertex;
end;

Constructor TSphere.Create;
Begin
//
end;

procedure TSphere.SetModelValue(r: single; lats, longs : Integer);
var
i, j, index : Integer;
lat0, z0, zr0, lat1,z1, zr1, lng, x, y : Single;
begin
clats := lats;
clongs := longs;
vertexCount := (lats+1) * (longs+1) * 2;

SetLength(ModelValue, (lats+1) * (longs+1) * 2);
SetLength(NormalValue, (lats+1) * (longs+1) * 2);

index := 0;
for i := 0 to lats do begin
lat0 := PI * (-0.5 + (i - 1) / lats);
z0 := sin(lat0) * r;
zr0 := cos(lat0) * r;

lat1 := PI * (-0.5 + i / lats);
z1 := sin(lat1) * r;
zr1 := cos(lat1) * r;

for j := 0 to longs do begin
lng := 2 * PI * (j - 1) / longs;
x := cos(lng);
y := sin(lng);
ModelValue[index].fX := x * zr0;
ModelValue[index].fY := y * zr0;
ModelValue[index].fZ := z0;
NormalValue[index].fX := x * zr0;
NormalValue[index].fY := y * zr0;
NormalValue[index].fZ := z0;
index := index + 1;
ModelValue[index].fX := x * zr1;
ModelValue[index].fY := y * zr1;
ModelValue[index].fZ := z1;
NormalValue[index].fX := x * zr1;
NormalValue[index].fY := y * zr1;
NormalValue[index].fZ := z1;
index := index + 1;
end;
end;
end;

Procedure TSphere.Draw();
Begin
glEnableClientState(GL_VERTEX_ARRAY);
glEnableClientState(GL_NORMAL_ARRAY);

glVertexPointer(3, GL_Float, 0, ModelValue);
glNormalPointer(GL_Float, 0, NormalValue);

glDrawArrays(GL_QUAD_STRIP, 0, vertexCount);

glDisableClientState(GL_VERTEX_ARRAY);
glDisableClientState(GL_NORMAL_ARRAY);
End;

My Mog Roulette game likes this...

Probably, it will be completed in the future or may not.

GNUstep

GNUstep is a way of using with Objective-C under Window or other Platform. Probably many know or not that the language can support same functionality as other languages do. The language is being used with "Cocoa" for building an application for IPhone. The language is quit interesting for learning. There are two installation files should download and install. Before install it, Mingw should install first. Open a new tab this link , get gnustep-system-0.9.1-setup.exe and gnustep-core-0.9.2-setup.exe. After that, install it by order. GNUstep Renaissance will give some special things. go and get it . Probably need its source because should compile and install. Unzip it to MSYS working folder. The working folder will be created by clicking at MSYS shell of GNUstep at first time. The above shown picture shows an execution figure of simple application.

Since, the process can go advance...

Clear screen shot of roulette under OpenGL

One of Nehe's OpenGL tutorials gives me an example how to use. The roulette machine will smoothly spin with OpenGL. but since, this picture means alpha valued image works with OpenGL.

To apply for alpha value

To get a proper alpha valued image, some tasks should be included in the previous list. Thank you for Paul Nicholls, I found FloodFill() routine which can be used with TBitmap32.
The above shown picture is alpha valued image. When the generated image is opened by Gimp, it will be the same as the shown above picture. I include its source.

Click -> available download