← notes

Solving the right problems

• Source

Notes

• There is a huge demand for engine programmers, even though most studios use off-the-shelf ones (unity, ue)
• Comments on advice he has seen:
  • There is more value in solving the most statistically likely problem. He started this point by referring to Single responsibility principle from SOLID.
  • There is value in reinventing the wheel. That's how we learn how the wheel works. (*)also, from several of Casey's talks. The wheel has been around for thousands of years, marginally changed. None of these tools have outlived this test of time
  • Design first is nonsense. The requirements and constraints are not apparent until the problem is tackled.
  • Abstraction as a core concept is not engineering. You code for reality.
  • Memory is one of the most important things in your program. Any shared resource is.
  • "Premature optimization ..." is misleading and often an excuse for not putting in the work.
• Things he expects:
  • Understand how your hardware works
  • Understand how the OS works
  • Have an area of expertise
• Practice:
  • Explore gaps in knowledge
  • Practice is ephemeral. Throw it away, you are not building a product
• Problem solving:
  • Solving the right-problem is an iterative process.
  • Understand the CONTEXT of the problem.
    • Things that can be assumed to be known.
    • Different problems require different solutions.
    • Identify user's needs.
    • Identify and balance constraints.
    • What-if game is a waste of time.
    • Future-proofing is a fool's errand. Everything has a lifespan.
    • Simplification has an inherit limit, defined by the complexity of the problem.
  • Understand the VALUE of solving the problem.
    • You need to be able to articulate why solving the problem is valuable
    • You need to understand how much time is it worth putting into the problem.
  • Understand the COST of solving the problem.
    • Predicted dev cost.
    • Opportunity cost.
    • Maintenance cost.
  • Understand the PLATFORM you are solving the problem for.
    • Reality is the actual problem. Not an abstract theoretical one.
    • There is no such thing as platform independant.
    • Know your tools.
  • Understand the DATA of the problem.
    • Everything is a data problem. If you don't understand the data, you don't understand the problem.
    • The purpose of all programs is to transform data.