Mojibake
termGarbled text produced when bytes written in one charset are read in another: café shown as café, don’t as don’t.
It is almost never random. Each pair of charsets garbles in its own recognisable way: UTF-8 read as windows-1252 fills a text with à and â€, a DOS code page shows box-drawing characters, EBCDIC read as Latin-1 turns every space into @. Because the wrong reading usually keeps every byte, the damage can often be undone by encoding the text back with the charset that misread it and decoding it with the right one; a byte that was replaced with ? or dropped is gone for good.
Also known as: garbled text, charset garbage