This reminded me to visit Apache's OpenOffice web site, where I was immediately greeted by an egregious encoding bug.
Their front page styles itself Apache OpenOffice �. That symbol probably won't survive poor LWN's 1990s web coding, but it's the Unicode replacement character U+FFFD, it appears when something went wrong either in transcoding or in some other text operation that's not permitted to report an error.
Generally when you see that it means somebody can't get their encodings quite straight, they've maybe tried to emit a registered trademark symbol, or a fancy quote mark, or really anything outside ASCII and their software has rather let them down. Obviously you don't want to trust software written by people like that with anything very important, especially if it's non-ASCII.