I'm hoping to implement a fuzzy match algorithm in TSQL (without MDS) that compares full names. The names are coming from separate manual inputs with no controls over what's entered. One of the systems also tends to cut off the end of names as it has a low character limit. As an extreme example of the sort of data I'm looking at:
System1Name System2Name
----------- -----------
Katherine G Livingstone Kat George Livingst
I'm definitely not looking for a perfect match here, but hopefully someone can recommend an algorithm for something like this. I've looked into Jaro-Winkler distance and Soundex, but I'm not sure if I'm better off using a distance based algorithm or a phonetic algorithm with this kind of data. Thoughts?