Joel Spolsky's latest article talks about text encoding. I agree with him in that what the article covers is the minimum acceptable amount of information regarding character sets and encoding methods for any programmer whose software ever touches text.

His claim about UTF-8's universalness left me wondering one detail, though.. I have been led to believe that to send email in Japanese, one should encode the message in ISO-2022-JP instead of UTF-8 to ensure its readability for all receivers. It would certainly be much easier to encode everything in UTF-8 instead. Knowing that 5 years ago most email programs would have had trouble displaying anything but ISO-8859-1 (Latin-1), I still tend to think UTF-8 probably is not email-safe. If you have any experience on this matter, please comment!