CSC151 2007S, Class 49: Object Basics

Admin:
* EC for the Alumna Scholar talk tonight.
* Are there questions on the exam?
* We'll start class by talking about homework 16.
* For tomorrow, you may want to review the reading for today.
  (Particularly if you didn't do it.)

Overview:
* The N paradigms
* Representing Compound Values.
* Introduction to Objects.
* Procedures as Objects.
* Adding State.

/Exam Questions/

//Sorting//

(define grades
  (list
    (list "Sam" 60 60 90 100 60 80)
    (list "Sammy" 95 95 95 95 90 80)
    (list "Samwise" 100 0 0 0 0 0)
    (list "Sammer" 60 70 80 90 100 110)))

(define points
  (list (cons 1 20) (cons 2 10) (cons 2 5) (cons 1 3) (cons 5 10)))

Can we really write make-filter without a lambda?
* I'll try after class
* If SamR can't do it in fifteen minutes or less, everyone gets credit
  (I think)

How concise is concise enough
* I'll post the number of words in my solutions

/Homework 16 - Selection Sort/

index-of-smallest
* Two strategies:
  * Direct recursion
    * Assume we get the index of smallest element in "the rest of the vector"
    * If the thing in pos is smaller, return ops
    * Otherwise return the index of smallest element in rest
  * Helper - keeps track of something

swap
* Straightforward

selection-sort!
* Use a kernel that recurses over the positions

---

/Object-Oriented Programming/

* Computer science is the study of algorithms and data
* Two key questions:
  * How do I think about and represent algorithms?
  * How do I think about and represent data?
* History: Multiple "paradigms"
  * First: Imperative (aka Procedural)
    * Algorithm is a sequence of steps, each of which does computation
      or reads/writes values
    * Traditional recipe
  * Alternate: Functional
    * Algorithms as function definitions and applications
    * Order of operations is more implicit
      (* (+ 3 4) (- 1 1) (sqrt 23))
      * Smart implementations can do less work
    * Tends to mimic mathematical approach to algorithm design
  * Alternate (popular): Object-oriented
    * Model everything in the program as a "thing"/"object"
    * Each object has some properties
    * Each object has some capabilities
* Grinnell believes you need to learn all three paradigms early in your
  career
  * Full blown object-oriented discussed in 152

Motivating example: Representing a library book
* How do you represent a book?
* Issue one: What information do you store?
  * Library of Congress number - String
  * Title - String
  * Author - (list of strings)
  * Number of pages - Integer
  * Subject matters - (list of strings)
* Issue two: How do you group that information together?
  * List
    '("Q76.2.23" "Scheme Scarified" ("Emily Jacobson" "Jacob Emilyson") 2386 (...))
    * How do you find the list of authors?
      (list-ref book 2)
      (caddr book)
  * Vector
  * File
  * Association list
    (('loc "QA76.2.23")
     ('title "Scheme Scarified for Scheming Students")
     ('author "Emily Jacobson")
     ('number-of-pages 2386)
     ('subject-matters "Sentience of computers" "Maliciousness of computers" "Lack of sentience of CS professors" "misspelling"))

Goal of OOP: If you change the representation, the program doesn't break