Skip to content

Programming Languages

Learn a language well enough to deliver software, then use runtime knowledge to explain what the program actually does under load and failure.

Choose a track from the problem in front of you. Use Go or Python when you need to build and operate software in that language. Use Language Internals when a type, allocation, compiler, runtime, numeric, security, or interoperability decision is unclear.

flowchart LR Need["Engineering task"] --> Build{"Build in Go?"} Build -->|"Yes"| Go["Go track"] Build -->|"No or uncertain"| Internals["Language Internals"] Go --> Evidence["Tests, traces, profiles"] Internals --> Evidence

The destination is evidence: a test, compiler diagnostic, benchmark, profile, trace, or production measure that confirms the mental model.

Tracks

Track Start here Use it to
Go Set up, write, organize, and operate Go software. Deliver services and tools with explicit concurrency, error, API, data, and debugging decisions.
Python Build maintainable Python applications, APIs, workers, and data-facing services. Make explicit choices about interfaces, concurrency, runtime behavior, and operations.
Language Internals Study concepts shared across languages and runtimes. Predict behavior, compare language choices, and diagnose problems below source syntax.

A practical study loop

  1. Pick one behavior you cannot yet predict.
  2. Write the expected result before running the program.
  3. Build the smallest example that can disprove the expectation.
  4. Inspect the relevant compiler, runtime, memory, or protocol evidence.
  5. Change one variable and repeat.
  6. Record the rule and its limit in your own words.

Level progression

Level Responsibility
Junior Make a small program behave correctly and explain the result.
Middle Choose maintainable boundaries and verify component integration.
Senior Protect system invariants across performance, failure, and evolution.
Professional Govern adoption, ownership, migration, and measurable outcomes.

Move to the next level when you can reproduce the current level's evidence without copying the example.