Last comments for similar_text()

On 23. Jun 2015 06:29 PoLyGLoT wrote:

similar_text is weird. I always heard it was (the num of similar characters)*200/ (length of string 1 + length of string 2). "Aluminum" "Spatula" Num letters in common = 4, which works out to be 53.3% similarity. similar_text ==> 2 letters in common, for 26.6% similarity. Why the discrepancy? I'd appreciate your input!