01 Introduction

  • Vasa Syndrome:

    • A famous historical failure in shipbuilding is used as an analogy for software projects that suffer from unclear requirements, shifting specifications, and inadequate testing.
    • Requirements
      • Changing shipbuilding orders
      • No specifications for modified keel
      • Shifting armaments requirements
    • Teams
      • Shipwright’s death
    • Metrics
      • No way to calculate stability, stiffness, or sailing characteristics
    • Quality Assurance (QA): Testing, reviewing
      • Failed pre-launch stability tests

  • Complex software engineering issues

    • Requirements (Architecture & Design)
      • Software running with kernel-level privilege can crash the system (BSOD) and prevent booting: High-privileged software can cause catastrophic system failures, like a Blue Screen of Death (BSOD), making the system unusable.
      • Content updates more continuous than software updates: Continuous content updates outpaced the software's ability to accommodate them, introducing instability.
    • QA
      • Insufficient testing of content and parser: Lack of thorough testing of both the content and the parsers (which interpret the content) allowed defects to go unnoticed.
    • DevOps
      • No staged roll-outs: The absence of gradual deployment of updates (staging) increased the risk of introducing wide-scale issues.
      • No way to fix issue remotely after impact: Once problems were introduced, there was no mechanism to resolve them remotely, leading to greater operational downtime.
    • Metrics
      • Skewed incentives: Misaligned goals or motivations across teams or stakeholders led to decisions that negatively impacted the product's quality or performance.
    • Licenses
      • Limited liability: A lack of accountability or liability limits may have discouraged sufficient quality assurance, leading to negligence in testing or development.

  • CrowdStrike Incident:

    • A modern example of software engineering failure due to insufficient testing, lack of staged rollouts, and difficulty in addressing post-release issues remotely.

  • Key Themes:

    • Software engineering as a human process
    • Requirements and Specifications
    • Metrics and Measurement
    • Software Quality: Testing + CI + Security
    • Continuous Deployment and DevOps
    • Software Project Teams
    • Managing Time, Teams, and Risks
    • Software Architecture and Design Docs
    • Scaling and Performance, Trade-offs
    • AI/ML in Software Engineering
    • Open-Source Software
    • Strategic Thinking about Software