You have to always decode.

No, you don't have to with US-ASCII and iso-8859-1.

But that wasn't part of Juerds claim.

I agree. He didn't mention any relation between the first 128 characters of iso-8859-1 and UTF-8. No idea why you bring this up.

iso-8859-1 is an encoding (of a subset of Unicode)

Unicode is a character set, not an encoding, so that sentence is broken.

iso-8859-1 is both a character set and an encoding. The iso-8859-1 character set is a subset of the Unicode character character set, but this property does NOT explain why iso-8859-1 works without being decoded.

No, you don't have to with US-ASCII and iso-8859-1.

Sure you do. It's not a difficult encoding, but it still is an encoding.

Unicode is a character set, not an encoding, so that sentence is broken.

Wait. You are saying that a sentence of the form "X is an encoding of Y" is broken in English is Y isn't an encoding?

I guess that "UTF-8 is an encoding of Unicode" is equally broken. For the reason that Unicode isn't an encoding in that sentence either.

Sure you do

Why do you keep insisting there's a need to apply an identity function? Or are you saying decoding ASCII isn't an identity function? It's hard to tell since you're nay-saying without any explanation.

You are saying that a sentence of the form "X is an encoding of Y" is broken in English [if] Y isn't an encoding?

I seem to have misread what you said. (Perhaps I missed the "of"?) Striken.