Anyone who has seriously wrangled web content and especially XML will have encountered headaches with encoding. You see junk characters or little rectangles (signifying unknown characters) in your browser instead of the correct ones.
In general, they happen when the rendering component is expecting one encoding but the stream it receives is encoded with another. One common case is when a web page purports to be encoded in UTF-8 (via an HTTP header and/or a meta tag) but actually contains Windows-1252 encoding because its contents were pasted directly in from a Word document (for example).
These kinds of mistakes are to be expected from web amateurs but even pure web companies can make a hash of it. This recent example is from Bloglines:

My exploration yielded the following information.
- I saw the same problem in Firefox, Camino and Safari.
- There is no encoding problem with the article itself on oreillynet.com.
- The Atom feed appears to be encoded properly.
- The Bloglines frame source shows that this HTML is written using JavaScript and that the JS strings have junk in them.
An extract from the source is here:

Is it hard wrangling encoding in JavaScript? Do you need to be careful with string constants in JavaScript? I haven’t used JavaScript with non-ASCII characters so I am not sure.
(Oh yes, the French is wrong, too. The word is égalité, not equalité.)