Mechanized reasoning has proved effective in avoiding serious mistakes in software and hardware, and yet remains unpopular in the practice of mathematics. My thesis is aimed at making mechanization easier so that more mathematicians can benefit from this technology. Particularly, I experimented with higher-dimensional types, an extension of ordinary types with a hierarchy of stacked relations, and managed to mechanize many important results from classical homotopy theory in the proof assistant Agda. My work thus suggests higher-dimensional types may help mechanize mathematical concepts.
Ki'hâihuà thuilí í'king ē'tàng pībián nńgthé kap ngē'thé tiong giâmtiōng ê tshò'ngōo, tsóng--sī tī sòoha̍k giánkiù si̍tbū iáu hántit iōng. Pún lūnbûn ê tsongtsí sī beh sú ki'hâihuà khah khuài hōo sòoha̍kka sú'iōng. Siôngsè lâi kóng, guá sú'iōng ko'uî luīhîng (higher-dimensional types), iā tsiūsī tī itpuann ê luīhîng tíngkuân ka'thiam to'kai'tsân ê kuanhē, tsiong kóotián tônglûn lílūn (classical homotopy theory) ê kuí'nā hāng tiōng'iàu ê sîngkó, tī tsìngbîng hú'tsōo kangkhū Agda lāi'té sūnlī ki'hâihuà. Guá ê giánkiù sîngkó hiánsī ko'uî luīhîng kiámtshái ē'tàng pangtsōo sòoha̍k khàiliām ê ki'hâihuà.
My thesis failed to explicitly attribute many fundamental concepts, such as h-levels, to Vladimir Voevodsky. One should be aware that Voevodsky's contribution is much greater than the impression given by the text.
These errata were fixed in the updated PDF but not in the submitted (official) version.
Minimalism with GitHub-like CSS (licensed under MIT and modified by me).