Abstraction and Program Design: Revision Notes (Computer Science A 3.1)

These notes recap Topic 3.1. For the full explanations and worked examples, read the study guide.

Recap

• Abstraction reduces complexity by focusing on the main idea and hiding irrelevant detail.
• Data abstraction names data without exposing how it is represented. In a class the data are attributes: variables declared in the class outside any method or constructor.
• Procedural abstraction names a process. A method can be used knowing only what it does, not how.
• Method decomposition breaks a large behaviour into smaller methods, each with one job.
• Parameters generalise a method so it works for a range of arguments.
• You may change a method’s internals without telling its callers if the signature and what it does stay the same.
• Design a class before coding it: list the attributes and the behaviours, in natural language or as a diagram.

Key relationships

Idea	What it hides	Java form	Example
Data abstraction	How the data are stored	Attribute (instance or class variable)	private int seatsLeft;
Procedural abstraction	How the process works	Method	sellTickets(int count)
Instance variable	Nothing: it is per-object state	No static	each Booth has its own soldHere
Class variable	Nothing: it is shared state	static	one totalSold for all booths
Parameter	The specific value used	Variable in the method header	addTax(double amount)

Assumptions to keep in mind

• A local variable is not an attribute: it lives inside one method or constructor.
• An object’s state is its attributes and their current values.
• A design should list all attributes and behaviours the specification needs, not a sample.

Mistakes to avoid

• Calling a method “abstract” because it has no code. Abstraction is about hiding detail, not missing code.
• Listing only attributes (or only behaviours) in a design.
• Repeating the same block twice instead of writing one method with a parameter.
• Changing a method’s signature while believing it is an “internal” change.
• Treating a class variable as if each object had its own copy.

Mini example to remember

A Screening object has the attributes filmTitle, seatsLeft and price, and the behaviours sellTickets(int), getSeatsLeft() and getPrice(). sellTickets(2) on a screening with 3 seats left returns true and leaves 1; a second sellTickets(2) returns false and leaves 1. The data (attributes) and the process (method) are each given a name, and the callers use them without seeing the insides. In the design diagram, the attributes go in the middle box and the behaviours in the bottom box.

Quick self-check

1. A class has private int lives; and a method loseLife(). Which is data abstraction and which is procedural abstraction? Answer: lives is data abstraction (an attribute), loseLife() is procedural abstraction (a method).
2. A method contains the same three lines twice. What should you do and why? Answer: extract them into one method (with parameters if the values differ), to remove duplication and leave one place to fix.
3. You speed up a method but keep its header and behaviour. Must callers change? Answer: No.
4. static int count; appears in Pet. If three Pet objects exist, how many copies of count exist? Answer: one.