URL Slug Generator

The naive slug is lowercase text with every non-alphanumeric run replaced by a hyphen, and it quietly destroys words. "Crème Brûlée" becomes "crme-brle", because stripping non-ASCII removes the letter along with its accent. Normalising to Unicode NFD first splits an accented letter into a base letter plus a combining mark, so removing the marks leaves the letter behind and the result is "creme-brulee".

NFD does not cover everything, and that is the second half of the problem. Letters like ß, æ, ø, ł, đ and þ are not decorated ASCII — they are distinct letters — so they decompose to nothing and vanish entirely, turning "Straße" into "strae". Those are transliterated from an explicit table instead. The table is deliberately limited to the Latin script: one that claimed to transliterate Greek, Cyrillic or CJK would be wrong in languages its author does not read, so Unicode mode keeps those characters instead, which is legal in a URL and has been since RFC 3987.

Ampersands and at signs are spelled out rather than dropped, because "Rock & Roll" reducing to "rock-roll" loses a word. Separators are collapsed and trimmed, truncation prefers a word boundary unless that would leave almost nothing, and optional stop-word removal never returns an empty slug — a title made entirely of stop words is still a title.