The levenshtein function in PHP. Because the string function names have to kept as inconsistent as possible.

Does anyone know a good source for learning the Damerau-Levenshtein distance(i'm trying to implement a simple spell corrector) and i haven't really found anything detailed on google yet. And also how does Damerau-Levenshtein distance fair to Levenshtein distance in terms of memory or time cost?

Since Google is failing me...

Given a user input (string query) and a list of larger strings (like email bodies or something), what's the best way to search and rank the list of strings against the user input.

So far I have implemented levenshtein distance but it doesn't really seem to do extremely well. (Short strings rank very well against each other, whereas long strings **containing** exact matches will go lower in the list)

Should I be splitting the input and the list by word and then averaging the distances?

The only thing I have tried is removing complete non-matches from the list by not including them if the distance is equal to the length of the largest string