A way to know if a Kanji-like Character is Chinese or Japanese.
Google Translate can automatically detect the input language and provide translation of the target language. More over, the Google search engine itself could be using the similar technologie to detect input language and then decide search results in which language they should display.
But often, this faills! Try to search something in complete Chinese with multiple characters on a IP address located in Japan, you will see the problem.
After the Han Unification, it is ture that they can not be differentiated by the unicode. But still, there should be a way to smell the differences, espatially in Chinese we use a lot of Characters and in Japanese people doesn't, even they do, the combination should be different. It is kind of trivial to make this judgement by IP adresse, which could be probably Google is doing as a part of its algorithms. There should be a better way to solve this problem.