Self-Referencing Source Code
Picture of me
I'm Jay Dan Howard!
I believe compassion makes tech worthwhile

Evolving list of loosely held, hopefully not pretentious sounding, and strong beliefs about software engineering, mayhaps only applying to the web world:

  • Code that feels good to write is productive code
  • Teams should use auto-formatters
  • Unit test files that have more lines of mocks than tests is a really bad sign since they cement contracts that will be null and void when business needs and implementations change
  • Dogma is as rampant in tech as it is anywhere else people are involved. We should always stay open-minded
  • Everything in software is about getting things to communicate "well"
    • To people: other developers, business stakeholders, users, API consumers
    • To machines: clients to services, services to services, processes to processes
    • To present-tense code: interfaces between functions, classes, network calls
    • To past-tense code: gracefully handling old versions of code after deploying to prod
    • To future-tense code: today's code will happily swap in and out with tomorrow's code
  • We need to remember performance is a measurement of the latency and throughput of the communication between or within machines
  • Performance claims that aren't measurable tend to be a result of engineers being shy about being called artists. We should not be shy
  • Large multi-hundred-line PRs indicate problems in the process
    • Poor developer utilization
    • Slower and more painful feedback cycles (requested changes have to happen in the service code and in tests)
    • Behind-the-scene iteration from single developer battling dragons