Slide 3
Slide 3 text
Be warned that what follows is just an experiment.
There is nothing prescriptive / exemplary / recommended about the experiment and about its parameters, e.g.
• The programming paradigm used (imperative programming - shudder)
• The contents of the copilot-instructions.md file used
• The Github Copilot mode used (Agent)
• The LLM model used (Claude Sonnet 4.6)
• The particular prompts used - e.g. their length, their approach to expressing details, etc.
Just because the above parameters had the intended consequences for the limited purpose of re-enacting a
classic sample code kata, it doesn’t mean that they are of any particular merit.
They were not the results of a planned strategy and/or tactics aimed at securing a desired outcome in terms of
e.g. optimality, effectiveness, efficiency, productivity.