160 likes | 329 Views
Abstraction: Procedures as Parameters. CMSC 11500 Introduction to Computer Programming October 14, 2002. Roadmap. Motivation: Two too similar procedures? Procedural Abstraction Laziness is a virtue Procedures as parameters Contracts Unnamed procedures: Lambda Scope Binding variables
E N D
Abstraction:Procedures as Parameters CMSC 11500 Introduction to Computer Programming October 14, 2002
Roadmap • Motivation: Two too similar procedures? • Procedural Abstraction • Laziness is a virtue • Procedures as parameters • Contracts • Unnamed procedures: Lambda • Scope • Binding variables • Creating local variables: Let
Two Too Similar Procedures (define (square x) (* x x)) (define (squarelist alon) (cond ((null? alon) ‘()) (else (cons (square (car alon)) (squarelist (cdr alon)))))) (define (double x) (+ x x)) (define (doublelist alon) (cond ((null? alon) ‘()) (else (cons (double (car alon)) (doublelist (cdr alon))))))
Procedure Comparison • Both procedures: • Consume a list • Perform some operation on each element • Return new list • Differences: • Different name • Different operation • Different recursive call
Abstract Procedure • Problem: • Want different functions for “abstractop” (define (abstractlist alon) (cond ((null? alon) ‘()) (else (cons (abstractop (car alon)) (abstractlist (cdr alon))))
Solution: Procedures as Parameters • Can pass a procedure as a parameter • Bind to formal parameter (define (map proc alist) (cond ((null? alist) ‘()) (else (cons (proc (car alist)) (map proc (cdr alist)))))) (define (squarelist alist) (map square alist)) (define (doublelist alist) (map double alist))
Why Abstract? • Laziness can be a virtue • Abstract procedures localize key structure • Creates single point of control • Changes can be made in single location • Generally simplifies programs • Avoid copying and modifying code segments
Contracts • Describe input and output requirements of function • fn: <in1>{<in2>…<inn>} <out> • where <in>, <out> are types of input & output • Original: squarelist: alon alon • Now: map: (X -> Y) (listof X) -> (listof Y) • where X,Y are parameters - any type
More Examples • Triplelist • Scale list • Add2 list
Unnamed Functions: Lambda • Issue: Defining lots of trivial procedures • E.g. add2, double, triple… • Solution: unnamed function • Can’t be recursive • (lambda (<par1>{<par2>..<parn>}) exp) • E.g. “Add2” (lambda (x) (+ x 2)) • Apply like other procedures • ((lambda (x) (+ x 2)) 10) -> 12
Redefining with Lambda (define (doublelist alon) (map (lambda (x) (+ x x)) alon)) (define (triplelist alon) (map ???? alon)) (define (scalelist scale alon) (map ???? alon))
Let: Defining Local Variables • Issue: Store partial results • Solution: Embed in unnamed procedure • Parameters bind names to values • Let easier to read • e.g. f(x,y)=x(1+xy)^2+y(1-y)+(1+xy)(1-y) (define (f x y) (let ((a (+ 1 (* x y))) (b (- 1 y))) (+ (* x (square a)) (* y b) (* a b))))) (define (f x y) (lambda (a b) (+ (* x (square a)) (* y b) (* a b))) (+ 1 (* x y))(- 1 y)))
Scope: What’s in a Name? • Where does a variable get its value?
Summary • Procedural abstraction • Procedures as parameters • Contracts • Single point of control • Laziness is a virtue • Unnamed functions with lambda
Next Time • Defining local variables with let • Scope • Bound and free variables • Procedures as return values