an attempt to accelerate processing of vertices

I got a wonderful idea on OpenGL with Java.

To render real objects with OpenGL, it's necessary to process a great amount of vertices or normals. So I would like skip processing, like reading from external files and generate arrays with so much elements dynamically, on running.

I found that it is possible. The concept is following using Java static variables.
a class defining vertices data:

public class ComplexObject {
 public static FloatBuffer vertexBuffer;
 static{
  ByteBuffer vbb = ByteBuffer.allocateDirect(74 * 3 * 4);
  vbb.order(ByteOrder.nativeOrder());
  vertexBuffer = vbb.asFloatBuffer();

  vertexBuffer.put(new float[]{
    0.53f, 0.3f, 3.55f,
    /* definitions of vertices go on */
  });
  vertexBuffer.position(0);
 }
}

To call the defined data:

gl.glVertexPointer(3, GL10.GL_FLOAT, 0, ComplexObject.vertexBuffer);

What I did next is scratched a Java program to convert Wavefront .obj files into Java code. This worked fine, and generated codes so easily.

It is when I compile one of generated so long Java code. I encountered a problem that Java compiler alerts following:
The code for the static initializer is exceeding the 65535 bytes limit

In Java, many part of source code elements is limited to 65536 bytes:
http://java.sun.com/docs/books/jvms/second_edition/html/ClassFile.doc.html

It's been hard. To do next is implementing split method into converter or look for another solution.