90 likes | 112 Views
Mutability. SWE 332 Fall 2015 Paul Ammann. Data Abstraction Operation Categories. Creators Create objects of a data abstraction Producers Create objects of their type based on existing objects Typically used in immutable data types Mutators Modify objects of their type
E N D
Mutability SWE 332 Fall 2015 Paul Ammann
Data Abstraction Operation Categories • Creators • Create objects of a data abstraction • Producers • Create objects of their type based on existing objects • Typically used in immutable data types • Mutators • Modify objects of their type • Typically used in mutable data types • Observers • Take objects of their type as inputs and return results of other types
Adequacy of a API • Should provide enough operations so that users can manipulate objects conveniently and efficiently • Should have at least three of the four category operations • Should be fully populated • Possible to obtain every possible abstract object state
Some Examples • Polynomials – We’ve already seen this Poly.java • Complex numbers (pp 74-75 in Bloch) Complex.java • Note use of producers instead of mutators
Mutable/Immutable Transform • Mutable Stack example in Bloch (page 56) • Stack.java • Goal: Transform to an immutable version ImmutableStack s = new ImmutableStack(); s = s.push(“cat”); s = s.push(“dog”); • What do these objects look like?
Mutator Producer Consider a void mutator method in class C: public void mutator1(…) Corresponding producer method: public C producer1(…) Consider a non-void mutator method in class C: public S mutator2(…) Corresponding observer/producer methods are: public S observerPart(…) public C producerPart(…) Note that non-void mutator needs to be split into two methods. Example: pop() in Stack vs. pop(), top() in ImmutableStack
Typical Transformation • Typical method in mutable class Foo: public void foo(T1 t1, T2, t2, …) {modify “this”} • Immutable version of Foo: public Foo foo(T1 t1, T2, t2, …) { Foo f = … … return f; } • Functional programming vs. procedural programming.
Disadvantage: Performance • Typical approach: • Provide immutable class • Provide mutable companion class for situations where performance is an issue • Clients choose on performance needs • Example in Java Library: • String (Immutable) • StringBuilder (Companion Mutable Class) • Static factories can cache frequently used items
Caching Frequently Used Items • Goals: • Control creation of immutable objects • Limit occurrence of given object to 0 or 1 • Return existing object if possible, • Else construct new object • Flyweight Pattern • Requires a table of existing items • Table design is a key decision • Requires hiding constructors • Enables “fast” equals(), hashCode() • Memory leakage may be an issue