CISC 280 review sheet for second midterm exam, April 17, 2001

Exam covers all of chapter 2 except sections 2.3.4 (Huffman codes) and 2.5.3 (symbolic algebra).

sample questions with some answers

  1. What is a recognizer? __________
    Ans: a predicate that is true just for a specific type. For instance null? pair? list? procedure? number? sales-employee?
  2. What is a constructor? __________
    Ans: a procedure for creating a new instance of some object type. For instance make-rat cons new-cx.
  3. What is a proper list? __________
    Ans: null or a pair whose cdr is a proper list.
  4. What is a generic operator? __________
    For ans material first review paragraph one of 2.5, then check out which procedures/operations are being called generic and why.

    The following apply to the lists L = ((1 2) (3 4)) and M = (5 6 7).
    ...if you think any of these ans may be wrong, just check in scheme. Also it is good practice to write box and arrow diagrams for these.
  5. (append L M) returns __________.
    Ans: ((1 2) (3 4) 5 6 7)
  6. (list L M) returns __________.
    Ans: (((1 2) (3 4)) (5 6 7))
  7. (cons L M) returns __________.
    Ans: (((1 2) (3 4)) 5 6 7)
  8. (cons (list L (+ 1 3)) (append (list M) M)) returns __________.
    Ans: ( (((1 2)(3 4)) 4) (5 6 7) 5 6 7) -- whew!
  9. (cdr (car L)) returns __________.
    Ans: (2) -- caution, "2" is not correct!
  10. (length (list L M L L M)) returns __________.
    Ans: 5 -- trick question: because easier than you might think!

  11. Write (reverse L), the procedure that returns a list which has the items of L in it, but in reverse order.

    Ans: Reverse is harder than one might think because of the asymmetric basic access to a list: car is easy, last element is work to get. It is rather easy to write a wrong solution. ; (reverse L1) ->L2. Assumes L1 is a proper list, returns a list of the same elements, but in the opposite order.
    (define (reverse L) 
      (define (reverse-iter L1 L2)
    	  (if (null? L1) 
    	      L2 
    	      (reverse-iter (cdr L1) (cons (car L1) L2)) ))
      (reverse-iter L '()) )
    

  12. Write a procedure split which takes a list as argument and returns a list of length 2. The two items in this output list are lists. The first one is the list of all the elements of the input list that are in odd numbered positions. The second list is all the elements in even numbered positions.
    > (split (list 1 2 3 4 5 6 7 8 9 0))  ; input may have even length
    ---> ((1 3 5 7 9) (2 4 6 8 0))
    
    > (split (list 1 2 3))  ; input may have odd length
    ---> ((1 3) (2))
    
    > (split (list 1))  ; input may be short
    ---> ((1) ())
    
    > (split '() )  ; input may be very short
    ---> (() ())
    
    > (split (list "a" 1 "b" 77 (* 2 4) "c" '(1 2))) ; it doesn't matter what's in the list 
    ---> (("a" "b" 8 (1 2)) (1 77 "c"))
    

    Ans: Here are three solutions representing different approaches to the problem (my favorite is the second one):
    ;; Solution 1: do and just use one half of the job.
    (define (split L)
      ; procedure for getting the odd position elements
      (define (odds L)
    	  (cond ((null? L) '() )
    		((null? (cdr L)) (list (car L))) ; ask self: why not just (car L)?
    		(else (cons (car L) (odds (cddr L))) ) ))
      (if (or (null? L) (null? (cdr L)))
          (list (odds L) '())
          (list (odds L) (odds (cdr L))) ))
    
    ;;;;;;;;;;;;;;;;
    ;; Solution 2: do both halves with procedures that call each other recursively.
    (define (split L)
      (define (odds L) (if (null? L) '() (cons (car L) (evens (cdr L)))))
      (define (evens L) (if (null? L) '() (odds (cdr L))))
      (list (odds L) (evens L)) ) ;; short and sweet!
    
    ;;;;;;;;;;;;;;;;
    ;; Solution 3: iterative
    (define (split L)
      ;; procedure to add odd-position elements of L to odd-L, 
      ;; and add even-posn elts of L to even-L.
      (define (split-iter L odd-L even-L)
    	  (if (null? L)
    	      (list odd-L) even-L)
    	      (split-iter (cdr L) even-L (cons (car L) odd-L)) ))
      (split-iter L '() '()) )
    ;;;;;;;;;;;;;;;;;;
    ;; Solution 3a: actually that split-iter is going to return the lists in 
    ;; reverse order.  So the last line should be:
      (map reverse (split-iter L '() '())) )
    

    Stuff on painters

  13. Write a painter (named "en") to draw the letter "N" with its vertical bars ON the side edges of the frame.
    ;; Ans:
    (define en 
      (define p1 (make-vect 0 0))
      (define p2 (make-vect 0 1))
      (define p3 (make-vect 1 0))
      (define p4 (make-vect 1 1))
      (segments->painter (list (make-segment p1 p2)
    			   (make-segment p2 p3)
    			   (make-segment p3 p4) )))
    
  14. Write a painter to draw this figure:
    Ans: no figure! go figure?
  15. What does this-painter draw?
    (define this-painter (beside (beside en (flip-vert en)) en) )
    

    Ans: Do it in scheme and see!
  16. Define a procedure (mirror p:Painter) that produces a painter that draws p twice, inverted, flip-vert, in the top half and unchanged in the bottom half of the frame.
    ;; Ans:
    (define (mirror p:Painter)
      (lambda (frame)
    	  (below p:Painter (flip-vert (beside p:Painter p:Painter))) ))
    
  17. Define a procedure (circus-mirror u:UPO p:Painter) that produces a painter that draws p twice, once with u applied to p in the top half, and once, unchanged, in the bottom half of the frame. (UPO means unary painter operator.)
    ;; Ans:
    (define (circus-mirror upo painter)
      (lambda (frame)
    	  (below painter (upo painter))))
    
  18. Define a procedure (new-mirror u:UPO) that produces a UPO so that ((new-mirror u) p) is a painter that draws the same thing as (circus-mirror u p).
    ;; Ans:
    (define (new-mirror u)
      (lambda (p)
    	  (circus-mirror u p) ))
    

    Stuff on sets

  19. Write an efficient procedure to form the union of two sets of numbers represented as sorted lists.
    ;; Ans:
    (define union (lambda (s1 s2)
      (cond ((null? s1) s2)
    	((null? s2) s1)
    	((< (car s1) (car s2)) (cons (car s1) (union (cdr s1) s2)))
    	((= (car s1) (car s2)) (cons (car s1) (union (cdr s1) (cdr s2))))
    	((> (car s1) (car s2)) (cons (car s2) (union s1 (cdr s2))))
    	)))
    
  20. What advantages does the sorted tree set representation have over the sorted list representation? What advantages does the sorted list representation have over the sorted tree representation?
    Ans: when balanced, sorted tree will have faster member-set? and adjoin-set? but sorted list will have faster union and intersection when the two lists are about the same length.
  21. In the following code join-em is a procedure which (like append) takes two lists and forms the list consisting of all the elements in the first list followed by all the elements in the second list. What difference does it make in the number of steps to do tree->list on a tree T with n nodes if join-em is a one step procedure or is a k step procedure, where k is the number of items in the first list?
    (define (tree->list T)
      (if (empty-tree? T)
          '()
          (join-em (tree->list (left-branch T))
    	       (cons (node-value T) 
    		     (tree->list (right-branch T)) )) ))
    

    Ans: join-em is called at every node of the tree so it is called n times when there are n nodes in the tree. If each use is a one step cost the total cost is O(n). If each use is the cost of the size of the first, left, tree, then the cost is the sum of the sizes of all left branch trees, which at worst can be the sum of all the numbers from 1 to n-1, which is O(n^2). Remark: This question is based on material we didn't cover, or at any rate didn't emphasize so far. We didn't discuss this matter so far this term. So don't expect it's brother on the exam.
  22. Mercy hospital and the VA hospital have decided to use a common system to support their patient records. Each hospital keeps a record on new patients consisting of the patient's name, date of admission, and diagnosis. However, Mercy uses a list of the form (name date diag), whereas VA uses (diag name date). In each hospital, many programs deal with these records, so it is not possible to simply change one to conform to the other. To handle the combined record keeping system, it is necessary to provide the constructors make-Mercy-record and make-VA-record and provide the generic selectors name, date, and diag.

    Sketch how you might use data-directed programming to do this, using manifest type and using put and get on a table of procedures.

    ;; Ans:
    (define (install-pkg-Mercy)
      (put 'make-Mercy-record 'Mercy (lambda (n dt dg) (attach-tag 'Mercy (list n dt dg))))
      (put 'name 'Mercy car)
      (put 'date 'Mercy cadr)
      (put 'diag 'Mercy caddr)
      )
    
    (define (install-pkg-VA)
      (put 'make-VA-record 'VA (lambda (n dt dg) (attach-tag 'VA (list n dt dg))))
      (put 'name 'VA cadr)
      (put 'date 'VA caddr)
      (put 'diag 'VA car)
      )
    
    (define (make-Mercy-record . args) (apply (get 'make-Mercy-record 'Mercy) args )
    (define (name record) (apply-generic 'name record))
    ;; etc.
    
    

Here is a quick summary of some procedures studied and other terminology,

Pairs: car cdr cons pair? null?

Uses of pairs abstraction barriers - levels from basic scheme up to application functions

Sequences (lists): nth length list append reverse list?

Symbols and quote '

Symbolic differentition: Expressions are naturally processed using tree recursion. Simplification is an issue that Make-sum, etc can handle.

rational numbers: interval arithmetic:

List processing: (map f L) (filter p? L)

More list processing: (flatten L) (extent L)

Painters: segments->painter,
Painter ops: transform-painter,
unary ops: flip-vert, rotate90, erase, etc.
binary ops: beside, below,
higher order ops: square-of-four,
recursive ops: right-split, corner-split, up-split, square-limit,

Frames: constructors and selectors

Multiple representations, Manifest type, Data directed programming

Some good exercises from chapter 2 are: 2.18, 2.21, 2.22, 2.23, 2.24