professional-work
Rhyme Generator
An algorithms assignment that finds English rhymes through phonemes. I built a search based on the final vowel and explored the results its rules and dictionary could not cover.
Finding rhymes through sound
Rhyme Generator is a Node.js console program built for an algorithms assignment. It searches by phonemes rather than spelling, using the CMU pronunciation dictionary and a Princeton word list.
I built an array-based stack and phoneme, vowel and rhyme dictionaries, then separated the search into stages. It worked backward from a word’s final phoneme to its last vowel, searched for matching pronunciations and filtered the results through the word list. I applied some functional programming principles along the way.
Is matching a rule enough?
Using the last vowel made the search rule clear, but it did not cover every rhyme someone might want. Rhymes within a word and words missing from the dictionary remained difficult. Dictionary quality, memory use, runtime, input validation and exception handling also needed work.
Reading this alongside the foodservice e-commerce project’s search evaluation work now helps me separate a search procedure from the criteria used to judge its results. This is a question from comparing the two experiences, rather than an evaluation system implemented in the original program.