Computing and Design

Session - 1

• System - A system consists of interconnected components with expected behavior.

Complexity

• what are 5 signs of complexity?
  • Large number of components
  • Large number of interconnections
  • Many irregularites
  • A long description
  • A teams of designers, implements, or maintainers

sources of complexity

• what are the sources of complexity ?

  • Comes from multiple requirements imposed by designers.

  • High utilization increases complexity.

  • More requirements and interactions add cascading complexity.

    

Coping with complexity

• What are the methods to cope with complexity?
  • Divide and concur methord using subsystem
  • Reduces debugging time and allows incremental improvement.

example
no. of lines ( N) ~ no. of bugs
N - 100 lines code - 100 Bug
to find 1 Bug, need to read N lines ( for finding 1 bug, - read 100 lines, )
100 bugs - 100 * 100 = 10000 times - $N^2$

Dividing code into K Modules - auth, user, product, demo, config, K = 5 Modules
eg: bug in auth, no. of lines to read is - N/K = 100/5 = 20 lines

debugging time

• time to fix 1 bug - 20 lines, N/K
• totol number of bugs - 100, which is N
• time to fix all bugs = time to fix 1 bugs * total number of bugs
  $=\frac{N}{K}\ast N=N^2/K$

date: 31/01/2025

Abstraction

• Separation of interface from internals.
• Often used with modularity.
• Ensures components interact in a well-defined manner.

Layering

• Organizes modules into layers where each builds on the one below.
• Reduces interconnections and simplifies design.
• Common in computing (e.g., OS, programming languages).

Hierarchy

• Components are organized into subsystems in a tree-like structure.
• Reduces the number of interactions among modules.
• Simplifies system design and debugging.

Why Modularity, Abstraction, Layering, and Hierarchy Aren't Enough

• These techniques assume designers fully understand the system.
• Choosing the right modularity, abstraction, layering, or hierarchy is difficult.
• Experience from previous systems helps guide decisions.

Iteration in System Design

• Start with a simple working system and gradually evolve it.
• Helps manage complexity and reduces risk.
• Document assumptions to manage future changes.
• Risks:
  • Bad-news diode: Bad news may not spread within the organization.
  • Second-system effect: Overconfidence from initial success can lead to overcomplicated redesigns.