[jasomill]

All character setes do not encode subsets of Unicode.

Two well-known counterexamples that come immediately to mind:

1. Mac OS Roman includes a non-Unicode Apple logo.

2. The Atari ST character set includes two non-Unicode characters that combine to create an Atari logo, and 4 non-Unicode characters that combine to create a picture of J.R. "Bob" Dobbs [1].

[1] https://en.wikipedia.org/wiki/J._R._%22Bob%22_Dobbs

[zzo38computer]

Yes, I know that (I was aware of both of these cases, as well as others). In those cases, there are characters that do not correspond to any Unicode characters.

Nevertheless, what I was saying is that many programs seem to be designed as though other character sets do encode subsets of Unicode, but actually they are different character sets and are not Unicode.

However, what I meant was, in addition to things like the examples that you have, other less obvious cases. Even if characters do correspond to Unicode (and sometimes there is more than one way to do it, which is the case for several PC characters), they are not necessarily supposed to work in the same way.

At best, Unicode could be used as an approximation if the other character sets cannot be used, although sometimes there are other ways to do it.