I find that the biggest misunderstanding (and note that I am now one of the I18N czars at big river books who runs around evangelizing it) is that people don't get that "unicode" is completely abstract.

Unicode is a set of integer numbers mapped to a big list of characters. It says nothing about how you represent the numbers or how you draw the characters. That's it.

UTF8 UTF16, UCS2, and UCS4 are known as "transfer encodings". Of these, UTF8 is the most useful persistent format and UCS2 is the most common format for manipulation. Converting something into UTF8 is generally described as "encoding your unicode into UTF8". So probably that's where it came from. Converting from UTF8 is decoding unicode from the UTF8 transfer encoding.

Ditto for the other transfer encodings I suppose.