REPL-driven development
summary
Argues for heavy REPL use with concrete demos: live coding in plain Python, then using bpython's undo and F7 editor round‑trip, showing tab completion, disassembly and inspect, and using IPython %run and autoreload versus deep reload. Explains the module reload caveats (from foo import bar doesn’t update) and shows embedding shells, pdb/werkzeug drops, doctest workflow and F5/F6 auto‑reload vs hard restart so you know when a REPL helps and when to restart.
AI-written summary · about
chapters (9)
- 0:00Start
- 9:24Use the REPL more
- 20:52What is REPL-driven development
- 24:09Any questions about anything besides reloading
- 33:13What about "real software"
- 33:30reload (and execfile)
- 35:16IPython's deep reload
- 35:42IPython's autoreload
- 37:14don't trust reloading
related talks
32mFeedback and the REPL
32mREPL, do you speak it
55mREPLs All The Way Up: A Rubric For Virtuous Feedback Loops
28mInstant-feedback, instant-debugging Python coding
50mREPLs All The Way Up: A Rubric For Better Feedback Loops
35mREPL driven development
36mRewind: Python Time-Travel Debugger
36mReactive Notebooks for Python