Indiana Jones and the Temple of Legacy Code - William Brander - NDC Oslo 2026

Indiana Jones and the Temple of Legacy Code - William Brander - NDC Oslo 2026

Title history
  • Indiana Jones and the Temple of Legacy Code - William Brander - NDC Oslo 2026 — observed on 20261103-1336 UTC (YouTube)

by William Brander · View the channel on PSI

Applied Sciences & Engineering Computing & Cybersecurity UMComputer programmingUMZSoftware Engineering
🎙 William Brander 👥 230K 📅 November 3, 2026 ⏱ 53 min 👁 4 📄 expert opinion 🧭 2026-11-03
Available in: English (current) Français

Keywords

legacy systemsupportabilityunderstandingmutation testingobservability

Summary

William Brander opens by recounting his childhood experience with a poorly designed Indiana Jones video game, setting up an analogy between exploring dangerous temples and working on legacy code. He defines legacy systems as naturally occurring states rather than intentional failures. He introduces two key attributes for assessing a legacy system: understanding (how well changes can be predicted) and supportability (the infrastructure enabling confident changes). Plotting these on a quadrant, he identifies four situations, with low understanding and low supportability being the worst. To improve supportability, he recommends reproducible builds, pinning dependency versions, removing machine-specific build requirements, setting up CI, and writing at least one test—any test—to validate the pipeline. He highlights mutation testing (e.g., Stryker.NET) as a way to assess test effectiveness and suggests using AI agents to generate tests and run mutation testing. For enhancing understanding, he advocates adding observability to the simplest flow first, then to the specific flow being modified, and using causal spans to map the system’s architecture. He also mentions mining version control history (‘source control archaeology’) to reveal why code evolved as it did. The talk concludes by emphasizing that legacy code is inevitable and can be managed through incremental, safe improvements.

199 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk delivers valuable, actionable advice for developers dealing with legacy systems. The quadrant framework (understanding vs. supportability) is a clear mental model that helps prioritize efforts. The step-by-step guidance—reproducible builds, pinned dependencies, CI, baseline tests, mutation testing, and observability—forms a coherent strategy. The argumentation is sound and grounded in personal experience, though it lacks quantitative evidence or case studies. The use of humor and relatable analogies (Indiana Jones, folding a fitted sheet) makes the content engaging and memorable. The speaker effectively argues that legacy code is inevitable and that gradual, low-risk improvements are more effective than rewrites.

Scientific Rigor, Source Quality, Title Accuracy

The talk does not cite external academic sources; it relies on the speaker’s professional experience. The only sources mentioned are the conference links (ndcconferences.com, ndcoslo.com) and a comic from Abstruse Goose, which is not formally referenced. The technical concepts (mutation testing, OpenTelemetry, reproducible builds) are well-known and correctly described, but no official documentation is referenced. The title accurately reflects the content, using a playful analogy that aligns with the talk’s tone and subject. The speaker’s credibility is enhanced by his evident familiarity with the topic, but the lack of citations reduces scientific rigor. Overall, the information is practical and plausible, but not evidence-based.

216 words

Title / Content Match

The title cleverly uses the Indiana Jones analogy to frame the talk's content about exploring and surviving legacy code, accurately reflecting the speaker's approach.

Quality & Reliability

7/10

The speaker provides practical, experience-based advice without empirical validation. The framework and techniques are plausible and actionable but lack formal references or metrics.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

  • NDC Conferences — Conference organizer and host of this talk.
  • NDC Oslo — Specific conference site where this talk was recorded.

Contribution & Novelties

The talk’s original contribution is the framing of legacy code as a quadrant based on understanding and supportability, which provides a practical diagnostic tool for engineers. It also offers a concrete sequence of actions to improve a legacy system incrementally. While the individual techniques are known, the structured approach and the analogy to Indiana Jones make it memorable.

Pour aller plus loin :

  • Mutation testing — A key technique discussed for evaluating test quality.
  • OpenTelemetry — The open-source observability framework recommended for mapping system flows.
  • Reproducible builds — The practice of ensuring builds are deterministic, as advocated in the talk.
  • Working Effectively with Legacy Code — A foundational book by Michael Feathers that aligns with the talk’s themes.

118 words

Radar Profile

The radar profile shows high marks in technical level and quantity of information, with moderate scores in quality and reliability, reflecting a practical, experience-driven talk with actionable content but limited external validation.

Reliability 7/10