Software Design Strategies

CSS CS Optional — Paper-I, Section-B, Topic V. How to decompose the system — different question from architectural style.

🔑 Underlying Design Concepts

Abstraction — expose essential detail, hide the rest.
Modularity — decompose into independent, well-defined units.
Coupling — interdependence between modules → lower is better.
Cohesion — relatedness within one module → higher is better.
Information hiding (Parnas) — reveal only the interface, hide internals. (design-level ancestor of OOP encapsulation)
Refactoring — improve internal design without changing external behavior. Tested as standalone MCQ in 2025.

Top-down — start from whole system, decompose into smaller sub-modules ("stepwise refinement").
Bottom-up — build smallest reusable components first, combine upward.
Function-Oriented (structured) — decompose around functions/processes; driven by DFDs; data & functions kept separate.
Object-Oriented — decompose around objects/entities; data + operations bundled together. (opposite of function-oriented)
Modularization — guided by coupling/cohesion above.

"Keeps data separate from functions" → function-oriented, not object-oriented.
Refactoring improves code quality/structure, not functionality — doesn't add features or find bugs.

Q. Explain the importance of design patterns under the umbrella of Agile software design and programming.

Ans.

A design pattern is a proven, reusable solution to a common design problem (examples: Singleton, Factory, Observer, Strategy, MVC).
Importance in Agile:
- gives flexible, well-structured code that is easy to change when requirements change
- supports refactoring towards cleaner design in every iteration
- provides a common vocabulary that speeds up communication ("use a Factory here")
- avoids over-design: patterns are applied only when a need appears (YAGNI)
- improves reuse and testability
Testing methodologies are on Software Testing (2016-17).