—
or —
Is there a difference between these? Is one better-supported than the other?
SGML parsers (or XML parsers in the case of XHTML) can handle —
without having to process the DTD (which doesn't matter to browsers as they just slurp tag soup), while —
is easier for humans to read and write in the source code.
Personally, I would stick to a literal em-dash and ensure that my character encoding settings were consistent.