This is my alternative to the Levenstein string distance algorithm. By looking for similarity locally, I made it really fast! Complexity O(n*constant) rather than O(n^2)
The basic idea behind it using a fast and dirty LCS (lowest common substring) length to determine the similarity between two strings.
A helper class that implements Levenstein and length distances is also provided.
Project is written in .Net 3.5 C#.