Software Project Mgmt & Quality: Metrics, Models & Standards

Key Concepts & Frameworks

1. Software Project Management Models

Software projects follow structured life cycle models to ensure systematic development. The three primary models are:

classDiagram
    class Waterfall {
        +Sequential phases
        +Documentation-heavy
        +Rigid structure
    }
    class Agile {
        +Iterative/incremental
        +Customer collaboration
        +Flexible scope
    }
    class Spiral {
        +Risk-driven
        +Combines Waterfall + Prototyping
        +Iterative refinement
    }
    Waterfall --> "Evolved into" Agile
    Agile --> "Inspired" Spiral

Comparison Table:

Worked Example: eSewa’s Payment System

• Model Used: Agile (Scrum framework)
• Why? eSewa’s team needed to adapt to Nepali government policy changes (e.g., tax regulations) and user feedback (e.g., mobile app usability).
• Process:
  • Sprint 1: Basic payment gateway (UPI integration).
  • Sprint 2: Added QR code payments after user demand.
  • Sprint 3: Fixed bugs in high-traffic periods (e.g., Dashain sales).

2. Quality Management Models

Quality is measured using metrics and standards. Two key models:

ISO/IEC 9126 (Software Quality Model)

Defines six quality characteristics evaluated at different levels:

mindmap
  root((ISO/IEC 9126))
    Functionality
      Suitability
      Accuracy
      Interoperability
      Security
    Reliability
      Maturity
      Fault Tolerance
      Recoverability
    Usability
      Understandability
      Learnability
      Operability
    Efficiency
      Time Behavior
      Resource Utilization
    Maintainability
      Analyzability
      Changeability
      Stability
    Portability
      Adaptability
      Installability
      Conformance

Example: Ncell’s MyNcell App

• Quality Focus: Usability (simple UI for prepaid top-ups) and Reliability (99.9% uptime during load shedding).
• Metric Used: Mean Time Between Failures (MTBF) to track server stability.

Capability Maturity Model Integration (CMMI)

A five-level maturity model for process improvement:

stateDiagram-v2
    [*] --> Level1: Initial
    Level1 --> Level2: Managed
    Level2 --> Level3: Defined
    Level3 --> Level4: Quantitatively Managed
    Level4 --> Level5: Optimizing
    Level5 --> [*]

Worked Example: NEPSE’s Trading System

• Current Level: Level 3 (Defined) – Standardized processes for share trading.
• Goal: Move to Level 4 by implementing automated risk detection (e.g., fraud alerts).

3. Software Quality Metrics

Quantifiable measures to assess project health:

4. Risk Management in Projects

Risks are classified into technical, project, and business risks. The risk management process follows:

flowchart TD
    A["Identify Risks"] --> B["Analyze Risks"]
    B --> C["Prioritize Risks"]
    C --> D["Mitigate Risks"]
    D --> E["Monitor Risks"]
    E -->|"New Risks"| B

Worked Example: Pathao’s Driver App Launch

• Risk: Low driver adoption in rural areas.
• Mitigation:
  • Partnered with local NGOs for driver training.
  • Offered incentives (e.g., free rides for first 100 bookings).
• Result: 30% higher sign-ups in Chitwan vs. Kathmandu.

5. Cost Estimation Techniques

Three widely used methods:

Worked Example: Khalti’s API Development

• Estimated LOC: 20,000
• COCOMO Estimate:
  • Organic Mode (Small team): 20 person-months (~₹1.2 crore)
  • Semi-Detached (Experienced team): 15 person-months (~₹90 lakh)
• Actual Cost: ₹85 lakh (saved 5% via Agile).

6. Quality Assurance (QA) Techniques

• Static Testing: Code reviews, walkthroughs.
• Dynamic Testing: Unit, integration, system, acceptance testing.
• Automated Testing: Selenium (UI), JUnit (unit tests).

In the Real World

1. eSewa’s Fraud Detection System

   • Idea: Defect Density Metrics
   • How? eSewa tracks false transaction alerts (defects) per 10,000 transactions. Goal: <0.1% false positives.
   • Impact: Reduced fraud losses by 40% in 2023.

2. Ncell’s Network Reliability

   • Idea: MTBF (Mean Time Between Failures)
   • How? Ncell’s 4G towers have an MTBF of 30 days (industry average: 20 days).
   • Result: Fewer customer complaints during load shedding.

3. Daraz’s Order Fulfillment

   • Idea: Cyclomatic Complexity
   • How? Daraz’s inventory system has a complexity of 5 (3 if statements for stock checks).
   • Why? Low complexity = easier debugging during Diwali sales rush.

Exam Tip

1. Models > Theory: Always compare Waterfall vs. Agile vs. Spiral in exams. Use eSewa/Khalti as case studies.
2. Metrics are Key: Memorize defect density, cyclomatic complexity, and SV (Schedule Variance) formulas. Apply to Ncell/Daraz scenarios.
3. Risk Management: Expect short-answer questions on risk identification vs. mitigation. Use Pathao’s driver adoption example.
4. CMMI Levels: Know Level 3 (Defined) is the minimum for government projects (e.g., NTC’s billing system).
5. Cost Estimation: COCOMO is favored in TU exams. Practice converting LOC to cost (use ₹15,000 per person-month as a rough rate).

Final Visual Summary:

mindmap
  root((Software Project & Quality Mgmt))
    Project Models
      Waterfall
      Agile
      Spiral
    Quality Models
      ISO 9126
      CMMI Levels
    Metrics
      Defect Density
      Cyclomatic Complexity
    Risk Management
      Identify → Analyze → Mitigate
    Cost Estimation
      COCOMO
      FPA