In the original question, you have utf-8 text being reinterpreted as latin-1. (I'd recognize that leading A-circumflex anywhere!)
Your html files must have a charset declaration, and it must match the character set you use to do the editing.
Some folks will tell you to use entities-- named or numerical-- for all non-ASCII characters. This is serious overkill. It makes your html unreadable, and makes it take up twice as much room (in bytes, disk space and bandwidth) as you go beyond the Latin-1 range. And it's obviously impossible if your base language uses a non-Roman script, so why do English web pages have to stay in the entity ghetto?