25th May 2008
Google is actually using a similar algo for domain name splits
Today someone mistyped my domain as thherarestwords.com - note two “h” and going to that link from url I’ve noticed something interesting. Both my algorithm (that I’ve described here) and Google’s one made the same major mistake with only a minor difference.
Notice both algorithms got “rest” and “words”, but not “rarest” as it’s a very rare word. So that makes me think that google is doing it the same way as mine algo does 
We just might have a little difference in coefficients in weighting evalutaion.
|