I noticed this immediately too. But since there are a fixed set of possible languages, seems like a dropdown box and/or multi-select would do the trick and remove the possibility of user error.
We use the language reported by the GitHub api, so the list of languages may change without notice. We should definitely normalize capitalization and I think there are node libraries edit distance we could leverage too. Thanks for the advice!
I'd like to be able to search "lisp" and get both Common Lisp and Emacs Lisp. I tried that and was surprised when nothing showed up.
I think that would be best handled by manually coding a set of language synonyms, rather than substring searches. Shorter language names could cause problems in the latter, such as C and Go.