All algorithms are ranging over the bytes of strings, which leads to different results for multibyte and singlebyte UTF-8 characters.
Example using WagnerFischer(a, b, 1, 1, 1):
"szellemhaj" - "szellemhajo" distance is 1
"szellemhaj" - "szellemhajó" distance is 2
Ranging over runes of strings would result 1,1 distance.
It's is slower but more accurate for non-ascii characters.
All algorithms are ranging over the bytes of strings, which leads to different results for multibyte and singlebyte UTF-8 characters.
Example using WagnerFischer(a, b, 1, 1, 1):
"szellemhaj" - "szellemhajo" distance is 1
"szellemhaj" - "szellemhajó" distance is 2
Ranging over runes of strings would result 1,1 distance.
It's is slower but more accurate for non-ascii characters.