November 4, 2025
Pencils down, drama up
Learning from Sudoku Solvers (2007)
Sudoku showdown: tests vs 12 lines — and the comments are savage
TLDR: A 2007 blog pitted Ron Jeffries’ test-first wandering against Peter Norvig’s quick, compact Sudoku solver; later authors added sharp critiques of test-driven development. Today’s commenters revived the fight: some champion Knuth’s algorithm or math models, others dismiss the redux and want deeper threads—illustrating evergreen method-vs-result drama.
An old-but-spicy 2007 post is back, pitting a meandering, test-first Sudoku project by Agile icon Ron Jeffries against Peter Norvig’s blink-and-it’s-done solver — and the comments are doing the heavy lifting. Author Peter Seibel drops a zinger, dubbing Jeffries’ detour-filled approach a pattern of “going in circles,” while praising Norvig’s lean, readable setup. Add in critiques of TDD (test-driven development) from other veterans, and you’ve got a retro flamewar with fresh kindling.
The crowd quickly picks sides. One camp cheers the “12 lines and chill” energy. Another defends TDD as great when you don’t know the path, just maybe not for pencil-puzzle math. Then the nostalgia bomb drops: a commenter asks if dancing links — a clever method from computing legend Donald Knuth — is “out of fashion,” igniting a throwback debate. A different voice goes full math-nerd and suggests treating the puzzle as a 3D grid (“9x9x9”) of yes/no switches — a tidy way to make a solver via math constraints, no vibe checks required. And the meta-snarkers? They’re here too, pointing to a newer, meatier HN thread and side-eyeing the short 2007 write-up.
It’s part code philosophy brawl, part algorithm nostalgia tour, and all comment-section theater — exactly how the internet likes its Sudoku.
Key Points
- •The article contrasts Ron Jeffries’s TDD-based Sudoku solver attempt with Peter Norvig’s concise solver.
- •It links Jeffries’s multi-post effort that emphasized board representation and tests over solving progress.
- •Peter Seibel’s commentary (cited in updates) critiques Jeffries’s focus and praises Norvig’s compact representation.
- •Peter Norvig later mentioned this post in Peter Seibel’s book Coders at Work.
- •An update links to Andrew Dalke’s “Problems with TDD,” broadening the critique of TDD’s limitations.