Functional Problem Solving (CSC 151 2014F) : Assignments

Assignment 6: Exercises in Recursion


Due: 10:30 p.m., Tuesday, 4 November 2014

Summary: You will apply the basic and helper recursion patterns to a variety of problems.

Purposes: To practice writing a variety of procedures that perform recursion over lists and numbers. To give you more experience verifying preconditions and writing local procedure bindings.

Collaboration: If you have been assigned to a group, you must work with your assigned partners on this assignment. The partner assignments are available at http://www.cs.grinnell.edu/~rebelsky/Courses/CSC151/2014F/partners/assignment.06.html. You may discuss this assignment with anyone, provided you credit such discussions when you submit the assignment.

Wrapper (Prologue): Individually read through this assignment and make sure that you understand what is required. Then use the form available at http://bit.ly/151hw6pro to indicate (a) how long you think this assignment will take and (b) what you think will be the most challenging aspect of this assignment.

Wrapper (Epilogue): When you are done with the assignment, fill out the form available at http://bit.ly/151hw6epi to indicate (a) how long the assignment took, (b) what the most challenging part of the assignment was, and (c) something important you learned from doing the assignment. If you find that the assignment took much less or much more time than you expected, also include (d) a note as to what might have led to that difference.

Submitting: Email your answer to . The subject of your email should have the form CSC 151.01 Assignment 6: Exercises in Recursion and should contain your answers to all parts of the assignment. Scheme code should be in the body of the message.

Warning: So that this assignment is a learning experience for everyone, we may spend class time publicly critiquing your work.

Preliminaries

Your procedures need not verify their preconditions, except as directed in the instructions to particular problems.

Assignment

Problem 1: A Safer Sum

Write and document a procedure (safe-sum values) that, given a list of values as a parameter, computes the sum of the numeric values in the list. That is, safe-sum should ignore all non-numeric values.

> (safe-sum (list 1 2 3))
6
> (safe-sum (list 3 'a 'b 5))
8
> (safe-sum (list 'a 'b))
0

Problem 2: Closest to Zero

a. Write a procedure, (closest-to-zero values), that, given a list of real numbers (including both positive and negative numbers), returns the value closest to zero in the list. Your solution should use basic recursion.

Hint: Think about how, given two numbers, you determine which is closer to zero.

Note: If there are multiple values equally close to zero, and all are closer than any other value, you may return any of them.

b. Write a second version of closest-to-zero that uses helper recursion. That is, you should have an additional helper procedure that takes closest-so-far and remaining as parameters.

c. Explain which version of closest-to-zero you prefer and why.

Problem 3: Averaging Colors

Averaging two colors is a fairly simple task: simply average each of the respective red, green, and blue components, as in the following procedure.

;;; Procedure:
;;;   irgb-average
;;; Parameters:
;;;   irgb1, an integer-encoded RGB color
;;;   irgb2, an integer-encoded RGB color
;;; Purpose:
;;;   Compute the "average" of two integer-encoded RGB colors.  
;;; Produces:
;;;   irgb-avg, an integer-encoeded RGB color
;;; Preconditions:
;;;   [No additional]
;;; Postconditions:
;;;   (irgb-red irgb-avg) is the average of (irgb-red irgb1) and 
;;;     (irgb-red irgb2)
;;;   (irgb-green irgb-avg) is the average of (irgb-green irgb1) and 
;;;     (irgb-green irgb2)
;;;   (irgb-blue irgb-avg) is the average of (irgb-blue irgb1) and 
;;;     (irgb-blue irgb2)
(define irgb-average
  (lambda (irgb1 irgb2)
    (irgb-new (quotient (+ (irgb-red irgb1) (irgb-red irgb2)) 2)
              (quotient (+ (irgb-green irgb1) (irgb-green irgb2)) 2)
              (quotient (+ (irgb-blue irgb1) (irgb-blue irgb2)) 2))))

We also saw how to average a list of colors in the lab on Recursion Basics. But what if we want to do something slightly different? Given a list of colors, we want averages, but only of neighboring elements in the list.

Write a procedure, (rgb-averages colors), that, given a list of colors, computes a new list of colors, by averaging subsequent pairs of colors. For example, if the input list is the standard seven rainbow colors (red, orange, yellow, green, blue, indigo, and violet), the output list will consist of a red-orange average, an orange-yellow average, a yellow-green average, a green-blue average, a blue-indigo average, and an indigo-violet average.

The length of the resulting list will be one less than the length of the input list.

Problem 4: Riffling Lists

Write and document a function (riffle first second) that produces a new list containing alternating elements from the lists first and second. If one list runs out before the other, then the remaining elements should appear at the end of the new list.

> (riffle (list 'a 'b 'c) (list 'x 'y 'z))
'(a x b y c z)
> (riffle (list 'a 'b 'c) (iota 10))
'(a 0 b 1 c 2 3 4 5 6 7 8 9)

Problem 5: The Nth Element

Write and document a procedure, (my-list-ref lst n), that extracts element n of a list. Your procedure should verify its preconditions and give helpful error messages. For example,

> (my-list-ref (list "red" "orange" "yellow" "green" "blue" "indigo" "violet") 5)
"indigo"
> (my-list-ref (list "red" "orange" "yellow" "green" "blue" "indigo" "violet") 0)
"red"
> (my-list-ref 3 (list "red" "orange" "yellow" "green" "blue" "indigo" "violet"))
my-list-ref: expected a list as first argument, given 3
> (my-list-ref (list "red" "orange" "yellow" "green" "blue" "indigo" "violet") 15)
my-list-ref: index too large for list

Even though this procedure does the same thing as list-ref, you should not use list-ref to implement it. Instead, your goal is to figure out how list-ref works, which means that you will need to implement this procedure using recursion.

Hint: When recurring, you will need to simplify both the numeric parameter (probably by subtracting 1) and the list parameter (probably by taking its cdr).

Problem 6: Generalized Numeric Sequences

You may recall that in our early work with lists, we built lists of numbers by using iota and then using map to scale the elements or increment the elements. For example, if wanted a list of six multiples of five, starting with ten, we might write the following expression.

> (map (o (l-s + 10) (l-s * 5)) (iota 6))
'(10 15 20 25 30 35)

Now that we know numeric recursion, we can write a recursive procedure that generates such simple arithmetic sequences.

Document and write a procedure, (sequence start increment bound), that produces a list in which the initial element is start, each other element is increment greater than the previous element, and all of the elements are strictly less than bound.

> (sequence 10 5 37)
'(10 15 20 25 30 35)

You must implement sequence using recursion. (You may use helper recursion, if you deem it useful to do so.) You may not use map or anything similar.

As you've noted, when we're careless with numeric recursion, we sometimes write procedures that recur forever. Hence, your sequence procedure must not only check its obvious preconditions, it must also ensure that the parameters are such that it is possible to create a finite sequence with the given parameters.

Extra Credit: Splitting Lists

Define and test a Scheme procedure, (unriffle lst), that takes a list as argument and returns a list of two lists, one comprising the elements in even-numbered positions in the given list, the other comprising the elements in odd-numbered positions.

> (unriffle (list 'a 'b 'c 'd 'e 'f 'g 'h 'i))
'((a c e g i) (b d f h))
> (unriffle (list))
'(() ())
> (unriffle (list 'a))
'((a) ())
> (unriffle (list 'b))
'((b) ())
> (unriffle (list 'a 'b))
'((a) (b))

Hint: There are many ways to solve this problem. Before writing code, try solving it by hand to develop your algorithm.

Important Evaluation Criteria

We will apply our standard evaluation criteria to this assignment. That is, solutions should be correct, clear, well documented, and tested (either with unit tests or with manual experiments).