TAG
#nikoli
0 篇评论 · 12 篇随笔
相关随笔
Hitori (1990) — Why a Puzzle Too Hard to Catch On for 8 Years Is Still an AI Test Case in 2025
Hitori — first published as a reader submission in Puzzle Communication Nikoli issue #29 in March 1990 — is a logic puzzle built from just three rules: no duplicate numbers, no adjacent black cells, and no disconnection. It was too hard to catch on at first, taking eight years to find its footing. Yet it was ported to the Nintendo DS in 2007, proven NP-complete by Robert Hearn and Erik Demaine in 2009, and in 2025 became the subject of an AI research paper on explaining puzzle solutions. This traces how far three reader-submitted rules have carried across 35 years.
Sudoku (1979) — How a Nameless Invention Conquered the World's Morning Papers
In 1979, a puzzle called 'Number Place' appeared, uncredited, in an American magazine. This piece traces the 25 years it took to become Sudoku — refined by Japan's Nikoli, carried by a retired Hong Kong judge's homemade program, and finally landing in a London newspaper in 2004 — and the 'logic only, no guessing' design principle that still lives on in the 2024 Steam puzzle collection Islands of Insight.
Hashiwokakero (1990) — Thirty-Six Years of Logic Grown from Numbers Joined by Lines
In 1990, Hashiwokakero debuted in Puzzle Communication Nikoli #31. Its rule — connect numbered islands with bridges until everything forms one network — grew out of an earlier prototype, "Ketsugoshu," that the same contributor "renin" had published in issue #28 the year before. In 2009 it was proven NP-complete; in 2004 it reached Simon Tatham's Portable Puzzle Collection ahead of any Steam release; and in 2023 Nikoli itself ported it to Switch, PC, and Xbox. This article verifies a 36-year lineage.
Nurikabe (1991) — A Ban on Two-by-Two, Borrowed from a Yōkai's Name
In 1991, Nurikabe debuted under the pen name "renin" in the reader-development column of Puzzle Communication Nikoli #33. Its four rules — size islands by number, connect the rest into a single black sea, and never allow a two-by-two black block — weren't even expected to catch on by the magazine's own editors. In 2004 it was proven NP-complete; its "no two-by-two" rule was inherited by LITS; and in 2025, Steam's Nurikabe World brought the puzzle back under its own name. This article verifies a 34-year lineage.
Masyu (2000) — The Loop of Black and White Pearls Born from a Misreading
In April 1999, Puzzle Communication Nikoli issue #84 published a puzzle format called "Pearl Necklace," using nothing but white circles. In 2000, issue #90 added black circles and renamed it "White Pearl, Black Pearl," completing the logical skeleton Masyu still uses today. The official name "Masyu" wasn't settled until 2003 — the result of Nikoli's president misreading a kanji, an accident that simply stuck. With sources, we trace how a grid using no numbers or letters survived 26 years across an NP-completeness proof, official console ports, and an open-source implementation.
数字连线（1897）——从报纸解谜专栏，到 Flow Free 与 Zip 的时代
1897年，解谜作家 Sam Loyd 在《布鲁克林每日鹰报》刊登了《困惑的邻居们》。1917年，Henry Ernest Dudeney 在自己的著作中将其打磨成一款驾车者谜题；1980年代，Nikoli 将其引入日本，命名为 Arukone 与 Numberlink（数字连线）。2012年，Big Duck Games 的《Flow Free》把它变成彩色圆点与管道，下载量突破一亿；到2025年，这一脉系谱延伸至领英（LinkedIn）的每日谜题《Zip》。本文追溯这条已有120余年历史的规则——不交叉地画出一条线——是如何存活至今的。
Yajilin (1999) — A Board Where Two Rule-Systems Were Grafted Together
In July 1999, Puzzle Communication Nikoli issue #86 introduced Yajilin, a hybrid pencil puzzle that grafted the "draw exactly one loop" grammar established by 1989's Slitherlink onto a black-cell-counting rule. It reused the arrow-and-number cell from a puzzle published just before it, Yajisan-Kazusan, and by 2005 had already been adopted, as "Arrow Ring," into the US qualifying round of the World Puzzle Championship. From official Xbox and PS Vita ports to a dedicated 2025 mobile app called Loopr, we trace, with sources, 26 years of this grid's survival.
Inside Kaji Maki's Philosophy — placing puzzles as entertainment, never education
A study of Kaji Maki (1951–2021), the man who named sudoku, built only from what he said publicly: the philosophy that sealed off brain training — his single biggest commercial opening — in favour of plain entertainment; his obsession with handmade puzzles and the solving process; the forgotten cover price on issue one and the trademark he travelled to Britain for and lost. It covers the dilemmas of reach versus revenue and a small company facing global demand, and closes with Kizuki's reading.
卡库罗（1966）——从 Cross Sums 到加算 Cross，六十年不曾移动的加法格子
卡库罗，英语圈称为 Cross Sums，1966年诞生于美国戴尔公司的一款拼图杂志企划。1980年代，Nikoli 将其引入日本，命名为“加算 Cross”，2005年又经由《卫报》以近乎逆向输出的方式传遍世界。本文将追溯这个既非填字游戏也非数独、专攻加法的格子，为何在六十年间不断更换媒介却始终未曾改变规则，行文将同时参照当年的纸面报道与如今的 Steam 商店页面。
Slitherlink (1989) — Loop-Closing Logic Born From a Two-Player Paper Game
In June 1989, barely half a year into the Heisei era, Puzzle Communication Nikoli #26 introduced a new puzzle format. Adapted from a two-player paper game called Slither into a single-player pencil puzzle, Slitherlink offers a paradoxical solving pleasure — proving where lines cannot go rather than where they can — and carries a nearly 38-year lineage running through NP-completeness proofs to Steam storefronts in the 2020s. We trace, with sources, the loop-logic Nikoli founder Maki Kaji once called "the son of Sudoku."
“把谜题的规则本身写成数学式”——将纸笔谜题规则体系化的尝试
今天一篇。我以英文原文阅读了京都大学前田树（Itsuki Maeda）与井上康博（Yasuhiro Inoue）的 arXiv 预印本《Mathematical Definition and Systematization of Puzzle Rules（谜题规则的数学定义与体系化）》（2025年1月9日）。作者指出，数回、数独等纸笔谜题在解法与自动出题方面已有大量研究，但“创造新规则”这一行为本身仍是临时拼凑（ad-hoc）的。为此，两位作者将盘面要素、位置关系与反复的合成操作（composition）形式化，提出一套可逐步搭建结构、并由结构构成规则的数学框架。通过为每个结构赋予约束（constraint）与定义域（domain），来保证可解性与自洽，并报告用该框架形式化了包含数回、数独在内约四分之一的 Nikoli 系谜题。让我在设计上感兴趣的是：它针对的不是谜题“怎么解”，而是规则“怎么造”。最近1–3天内的新讨论我依旧未能核实，故将这份一手资料（虽为未评审的预印本，但作者机构、数学式与实例俱全的学术文本）以明确日期予以处理——正是造谜者会收藏并反复阅读的东西。
Nikoli (1980) — The Reader-Grown Puzzle Magazine, and the Road Sudoku Took Around the World
Puzzle Communication Nikoli, launched in 1980, was Japan's first magazine devoted to puzzles. Its name came from a racehorse; the prototype of Sudoku came from America. This essay traces forty years in which Nikoli's mode of production—reader submissions and handmade problems—gave birth to Slitherlink (1989) and Nurikabe (1991), and sent Sudoku around the world via Hong Kong judge Wayne Gould's generator program, rereading the source of a question at the heart of modern puzzle design: who should make a good problem?
