As part of the Texcraft project I've been re-implementing Donald Knuth's TeX in a modular way. One of the goals of Texcraft is to be able to use these modular pieces in different contexts. hyphenate.dev is a website that uses my re-implementation of TeX's hyphenation algorithm to visually illustrate how the algorithm chooses hyphens.
This is cool, good luck with your project! For this page, you may want to pick as the default examples some longer words (like say "nevertheless") that show both positive (odd number) and negative (even number) patterns.
Also for anyone else reading this, Frank Liang's thesis "Word Hy-phen-a-tion by Com-put-er" (https://tug.org/docs/liang/) is a great read, and the data structure it uses (packed tries) is clever too. The section of the TeX program that describes hyphenation is also an interesting read, and Knuth added a further twist (what we may call a hash-packed trie) in his frequent-words literate program for Bentley's column.
Thank for the suggestion, implemented it! Now all suggestions have at least one hyphen and one position where the patterns disagree (one pattern says hyphenate, another to not hyphenate). I discovered you don't necessarily need long words to get "interesting" hyphenation results; e.g. https://hyphenate.dev/zero.
Also for anyone else reading this, Frank Liang's thesis "Word Hy-phen-a-tion by Com-put-er" (https://tug.org/docs/liang/) is a great read, and the data structure it uses (packed tries) is clever too. The section of the TeX program that describes hyphenation is also an interesting read, and Knuth added a further twist (what we may call a hash-packed trie) in his frequent-words literate program for Bentley's column.