URLs and HTML
"Chopping up" URLs in is completely appropriate. See the appendix of the URL RFC (http://www.faqs.org/rfc/rfc1738.txt). I can't understand why any mailer that doesn't implement this bothers to pick URLs out of text/plain content at all, since you can't do it unambiguously.
As far as HTML, hardly any complex messages are really worth the effort of describing appropriately, while it's incredibly easy for an imitation word processor to make you think your message is k3w7 when in fact its structure is garbage.