CS 3413 Data Communications, Fall 2004 Midterm Exam 2 Comments

I added 5 points to everyone's grade, mainly because of problem 5.

The grade distribution (after adding the 5 points) was as follows:

Comments about individual problems:
  1. Just before the exam I was asked about this and gave an answer to the class. Some people did not listen.
  2. You needed to assume the number of bits needed for each sample.
    A reasonable number would be 32 for stereo or 16 for mono.
    Eight bits per channel would make the music sound as if it were coming over a telephone line.
  3. Some people still don't get this. Expect a problem like this one on the final exam.
  4. Almost everyone got this one right.
  5. Only 4 people in the class did this problem correctly.
    Most did not know what the Hamming distance of a code was.
    It is important because it tells you how many bit errors can be detected and corrected.
    The book discusses this distance but does not give it a name.
  6. Most people were able to do part a) but not part b).
  7. Some people tried to do this by incorrectly applying a formula.
    It is better to do this problem from first principles.
    Remember that you were given the 1-way propagation delay and you need to use the round trip delay.
  8. Remember that an ack for a frame in the sender's window acknowledges all previous frames as well.
  9. In general, people did poorly on this problem.
    If you got the first column wrong and got the rest wrong because of it, come and see me and tell me what you did. If your reasoning was correct you can get additional points.