C++ – Design Patterns and Idioms (CPP3)

Course content:

Hide details

• Clear, maintainable code written for people
1. Use familiar practices and design patterns
2. Describe software as components and behavior
3. Separate behavior from implementation details

• Object-oriented software design
1. Program composed of collaborating components
2. Components have assigned responsibilities
3. Reason about individual components and roles

• Classic object-oriented programming
1. Represent components as class instances
2. Instances define component state and relations
3. Classes define behavior and component interfaces

• Inheritance and replacing ancestor methods
1. Adjust behavior by overriding methods
2. Template Method design pattern
3. Static implementation in modern C++ (CRTP)

• Reuse of code written once
1. Share general code via inheritance
2. Compose objects from ready-made building blocks
3. Generic data types and algorithms

• Separating interface and concrete implementation
1. Abstract classes as interface definitions
2. Swap implementations at runtime (dynamic polymorphism)
3. Strategy design pattern

• Dependencies between components in C++
1. Ownership (RAII), borrowing and references
2. Pointers, references and polymorphism
3. Object interactions and method calls

• Tools of generic programming
1. Composition using generic templates
2. Policy-based design pattern
3. Move variability to compile time

• Principles of good software design
1. Component independence and localized changes (SRP)
2. Depend on interfaces, not implementations (DIP)
3. Closed for modification, open for extension (OCP)

• Changing component behavior at runtime
1. State machine and State design pattern
2. Modern approach using std::variant
3. Differences from classic state machines

• Connecting interfaces with connectors
1. Separating different domains (Bridge)
2. Bridging incompatible interfaces (Adapter)
3. Exposing a simpler interface (Facade)

• Chaining through a single interface
1. Add behavior to an existing system (Decorator)
2. Chain possible handlers (Chain of Responsibility)
3. Protecting and controlling access (Proxy)

• Linear data structures and sequences
1. Traversing data structures (Iterator)
2. Differences between linked list and chain
3. Ending without null pointers (Null Object)

• Composing more complex objects and components
1. Branching by concrete type (Visitor)
2. Tree-like object structures (Composite)
3. Solutions using std::variant and std::visit

• Creating diverse object types
1. Factory for creating many kinds of objects
2. Abstract Factory for families of objects
3. Use lambdas and std::function for registration

• Single-instance objects
1. Create a single instance in modern C++ (Singleton)
2. Managing global resources
3. Communicating state with the OS

• Multi-step object construction
1. Builder object as an intermediate step
2. Methods to modify builder and final build
3. Split builder into multiple phases

• Signaling and event handling
1. Central coordinating component (Mediator)
2. Registering event subscribers (Observer)
3. Signals with templates and lambdas (signals & slots)

• Generation, streaming and lazy evaluation
1. Process sequences and pipelines (iterators, views)
2. Use message queues for decoupling
3. Generation using std::generator

• Representing data changes
1. Describe a specific change (transaction, command)
2. Sequence of successive changes (journal)
3. Use mini-languages or bytecode (interpreter)

• Storing and restoring state
1. Create snapshots of state (snapshot, memento)
2. Manage history of snapshots (history, recovery)
3. Combine snapshots and transactions (incremental backup)

• Optimizing data handling
1. Data structures with instant snapshots (copy-on-write)
2. Share mostly-immutable parameters (Flyweight)
3. Cloning objects and the "rule of zero" (prototype)