My guess is that the reality is "98% of websites are valid UTF-8 documents". A large portions contain only ASCIIs so they happen to be, just indistinguishable w

My guess is that the reality is "98% of websites are valid UTF-8 documents". A large portions contain only ASCIIs so they happen to be, just indistinguishable with truly UTF8 encoded ones until they break.