alt Hacker NewsMarkdown support isn't a bad idea, actually, as long as they don't break the most important (IMO) property of Notepad: binary WYSIWYG. I.e. if I type in some plain text and then open the file with anything else (including after moving to another machine/platform, or even viewing raw data stream in transit or on drive), I can trust to see that text, as is, and nothing else. In particular, if I restrict myself to lower 127 bytes, I expect byte-to-byte correspondence.
(Modulo CR/LF, of course.)
FWIW, Notepad has had support for BoM detection and wide-characters (UTF-16/UCS-16) for some while. That said, IMO, most simple editors at this point should default to UTF-8 encoding and only LF for line endings.