The current implementation of "Estimate reading time" seems to handle "words" as they are split by whitespaces for all languages.
However, in some non-western languages such as Japanese or Chinese, words are not split by whitespaces. Instead, the morphological analysis is required to split sentences into words. Thus, the current "Estimate reading time" is not accurate for those languages.
So, it would be nice if the time calculation logic is improved for those languages.