Modes to submit points

• He begins with an algorithm for generating a sierpinski gasket

  SIERPINSKI
  
     
   Pick an initial point (x,y,0) at random inside the triangle
     
   Select one of the three verticies at random
     
   Find the point q halfway between p and the randomly selected vertex.
     
   Display q 
     
   replace p with q
     
   go to step 2
  
      

• He proposes three ways to draw a picture
  • Immediate mode

    • At Startup
      Initialize graphics
      When display is needed
      while not done generating points
      	generate a point
      	display a point
      	       

    • Points must be recomputed each time.
      • Probaby too expensive if we redraw the data.
    • Points must be sent from CPU to GPU every time.
    • Memory usage is reduced.
  • Retained mode graphics

    • At Startup
      Initialize Graphics
      
      while not done generating points
      	generate a point
      
      When display is needed
      Send All points TO GPU
      Display All Points on GPU
      	    

    • Points are only generated once.
    • But each time the scene is redrawn, the points are sent to the GPU
    • This was the way we did things for a long time.
    • Especially when we were using the CPU to do the rendering.
  • Another Method

    • At Startup
      Initialize Graphics
      
      while not done generating points
      	generate a point
      Send points to GPU
      
      When display is needed
      When Needed: 
      Display All Points Stored in GPU
      	    

    • This means that data is only sent to the GPU once.
• Two Dimensional Application
  • We can think of a point in two or three space.
  • If we use two space, OpenGL will add a third dimension, with a value of 0
  • A vertex is a point in space
    • This may be a point
    • It may be an endpoint for a line
    • It may be one vertex in a polygon. (Triangle, quadrilateral, ...)
  • By default our drawing occurs in Normalized Device Coordinates (NDC)
    • A cube (-1,-1,-1) to (1,1,1)
    • OpenGL mapps this to Device Coordinates
    • For now, we need to specify our coordinates in this space.
    • Later we will specify our coordinates in object coordinates.
  • He introduces his vec structure next. Look over the source code, you should be fine with this.
  • OpenGL Data Types And Functions
    • OpenGL supports a number of data type names.
    • These are most likely #typedef from C types.
    • But they make programs far more portable.
    • See table 1.1 here
    • OpenGL functions begin with gl and follow our naming convention (glDrawArrays, glClear, glEnableVertexAttribArray).
    • Sometimes they come in faimily of funtions
      • Verticies are 2d, 3d or even 4d.
      • Verticies are ints, floats, doubles, ...
      • In this case the convention is name dimension, type
      • Look at the glVertex family.
  • OpenGL Constants begin with GL_ and follow our constant naming convention.