Hi, Every 7 bits encodings will always be valid UTF8 as the range U+0000..U+007F is explicitly (documented as) valid in UTF8.
So, there will always be the possibility to find legitimate UTF-8 encoded documents matching sequences found in ISO-2022-JP, like: '\x1B\x28\x42' which is "control, left parenthesis, LATIN CAPITAL LETTER B". If you want to start speaking about probabilities that a given document is UTF8 or ISO-2022-JP according to a frequency analysis of common patterns, you're not talking about `isutf8` but `chardet`. The role of isutf8 is *only* to tell if a given bit sequence is valid in utf8, it means that false positives are acceptable (ASCII is UTF-8 compatible, so if a document is encoded in ASCII, isutf8 will falsely tell it's valid, which is acceptable) but false negatives are not acceptable: every encoding error have to be reported. Bests, -- Julien Palard https://mdk.fr
