Session Nine, Monday February 4th

I didn't use any slides today.  Instead, we had a very interactive session where we looked at a variety of issues that arise when we switch from processing pictures with a single loop (normally done when you modify a picture as a whole) to processing pictures with a pair of nested loops (normally done when you are focusing only on part of the picture).

These issues included:

• What actually happens when you run a pair of nested loops.  We walked through our makeGrayscale() below (written in class on Friday) that used nested loops to show what order the data is processed.

def mysteryMethod(picture):
   for x in range(1,getWidth(picture)):
      for y in range(1,getHeight(picture)):
         pixel = getPixel(picture,x,y)
         redPart = getRed(pixel)
         greenPart = getGreen(pixel)
         bluePart = getBlue(pixel)
         val = (redPart + greenPart + bluePart)/3
         setColor(pixel,makeColor(val,val,val))

• Data access with 0-based vs. 1-based starting points.  Recall that JES treats pictures with the concept of (1,1) being the upper left hand corner (despite some typos in JES to the contrary).  While this isn't necessarily a big deal, there are several issues that accompany this. 

  The largest of these issues is that range(a,b) returns the list of numbers starting at a and counting by 1s until it reaches b-1.  For example:  range(1,5) returns the list of numbers [1,2,3,4] but NOT 5.  Thus, while we are tempted to say something like range(1,getWidth(picture)) to count the pixels across a picture this actually leaves the last column untouched.

  We demonstrated this by looking carefully at our makeGrayscale() method above to see that the last row and last columns remained untouched (OOPS!)

Finally, we spent a few minutes talking about what is involved with mirroring part of a picture onto the other part of a picture and how we would do this (independent of code).