It took me way too long to work out why e-mails from Outlook-using colleagues often had a loan 'J' at the end of them.
It seems that there's some easy way for users to indicate they want a smiley face in a message, possibly simply typing ':-)'. And Outlook helpfully[*1] decides to display that as a smiling face symbol, '☺'. But, rather than using the Unicode character for that symbol, it hatefully uses it from one of those pre-Unicode symbol fonts, where symbol glyphs were used in place of ordinary characters, such that what is a 'J' in the text is displayed as a face in that particular font. Which appears to display correctly in the HTML part. And then Outlook automatically generates a plain text part of message, hatefully by simply stripping all formatting information, replacing it with nothing. This of course means the 'J' is now just a plain 'J', with nothing indicating it should be anything else. And the plain and HTML parts are no longer true alternatives of the same content. All this for no reason: Outlook could simply use the Unicode '☺', which does the right thing regardless of font. Or, if the user simply typed the three Ascii characters ':-)' in the first place, Outlook could perhaps DO NOTHING AT ALL WITH THEM, BECAUSE IT ISN'T A COINCIDENCE THAT THEY CAN BE TRANSMITTED LIKE THAT IN PLAIN TEXT. Managing to break the transmission of three printable Ascii characters is actually quite a feat. Smylers [*1] Um, maybe I mean, hatefully there too. -- http://twitter.com/Smylers2