Your average joe doesn't get mad about em-dashes and en-dashes.

Well, he should! Correctly typeset text is a right of the reader, not a privilege.

For the record, all horizontal lines in the middle of bits of text are not created equal. There are at least four kinds.

A hyphen is used to connect two or more words into one, where ambiguity would result were they separate: a "sub-aqua club" rather than a "sub aqua club".

An en rule (&endash; in HTML, often represented as two hyphens where unavailable) is used for ranges: 4--12 megabytes per song. It's also used, with spaces each side, as a parenthetical separator -- rather in this manner.

An em rule (&emdash; or three hyphens) is used without spaces as a parenthetical separator---or occasionally to indicate direct speech---but is slightly old-fashioned these days.

The minus sign of mathematics is different again, typically longer than a hyphen but shorter than an en rule.

As for quote marks, it's more complicated still. The only quote-like characters in ASCII are the ' and " characters. Those are known as "unsexed" quotes (because they don't slant one way or the other) and are incorrect in typeset text except when used for the units of feet (or minutes) and inches (or seconds).

Whether for quotation of direct speech, or for scare quotes like those around "unsexed" above, the correct double quotes to use in typeset English text (other European languages are different) are the "66-like" and "99-like" ones available in HTML as “ and ” -- I don't have a Unicode spec handy to look up the exact character numbers. The correct single quotes to use are the "6-like" and "9-like" ones ‘ and ’. Apostrophe, the mark found in it's and similar words, is technically a different glyph from ’ but appears identical in all sane fonts.

The problem with all these sexed quotes is that they do not appear either in ASCII or in standard ISO8859-1. They appear in the reserved range (0x80-0x9F) of Microsoft code pages, and of Macintosh and Acorn encodings too -- but in three different places in those three encodings. That's why pages authored by Macintosh weenies sometimes look odd in Windows browsers (and vice versa): people have used the literal characters (unportable) rather than the entities (portable). The characters do not appear on standard keyboards (even French keyboards do not have the guillemet characters labelled).

Unix zealots tend to represent ‘ with grave accent and ’ with ', which looks wrong in almost all fonts these days.

This has been quite a long post already, so I'll stop ranting now, but I also hate people who think they can set text in small caps by using normal capitals in a smaller point size. Penguin Books do this on their spines these days. Penguin Books, for whom Jan Schichold once worked. How are the mighty fallen.

Peter