Midterm 2 Rubric
CS61A-Summer 09

Problem 1:

2 points for each box and pointer diagram.
1 point for each answer.

All or nothing.


Problem 2:

Part A: 5 points

3 points for knowing x
2 points for knowing y

Part B: 7 points

1 point for each correct circle
-1 point for each wrong circle


Problem 3:

15 points total.

14 pts - trivial mistake
12-10 pts - has the idea
8-4 pts - has an idea
2-0 pts - has an attempt

A single off by 1 indexing error: -1 pt
Multiple off by 1 indexing errors: -3 pts

Did not scale old vector index in vector-ref: -3 pts

No points off for not subtracting by 1 in
(+ 1 (div (- (vector-length vec) 1)) n))

Some people had crazy loops with 5 or 6 counter variables. We tried to follow them the best we could. Let us know if we made a mistake with yours.

Problem 4:

12 points total:

11 pts - trivial mistake
10-8 pts - has the idea
6-2 pts - has an idea

Tried to stream-add a ints to int-matrix instead of using stream-map: - 4 pts

Some tried to use stream-car, thinking that every time in the recursion, the first element is updated. This was a serious error. These solutions tend to get 4 pts total.


Problem 5:

12 points total:

4 points for saving a temporary variable to point to (cdr ls1) or (cdr ls2)
4 points for set-cdr! on ls1
4 points for set-cdr! on ls2

-1 point for trivial mistakes. 

Except for trivial mistakes, your answer had to be completely correct to receive credit.


Problem 6:

12 pts total:

1 point for each number.


Problem 7:

18 points total:

17 pts - trivial mistake
16 - 12 pts - has the idea, knew to modify mc-apply or extend-env and knew that we had to do case analysis and bind arguments differently
10 - 6 pts - has an idea
4 - 0 pts - has an attempt

Some people implemented the new lambda notation but broke the old one. These solutions get 8 pts.

Some people had (list args) but missed (list (params fn)). If changing just (list (params fn)) would fix the problem, we took off 2 points. If just changing (list (params fn)) would not fix the problem, i.e. in mc-apply you changed "args" to (if (atom? (params fn)) (list args) args), then we took off 4 points.

The difference between a 4 and 0 was whether you demonstrated some understanding of how metacircular evaluator works. i.e. putting mc-apply inside the (lambda-exp? exp) clause would be a 0 since lambda expression just creates a procedure; thinking that a lambda expression calls the procedure is a grave mistake.

The difference between a 0 and a 4 was whether your changes seriously broke the existing features of mc-eval.