httpwebwitch - 3:22 am on Jul 17, 2010 (gmt 0)
Yes, it is the case that my data was not properly encoded. Now I'm certain that everything is UTF8 - going in, coming out, top to bottom. I've tested on 6 different browsers on 3 operating systems, entering Chinese, Cree, Hebrew, Arabic, and French. Everything is A-OK.
I do still have some old data with bad encoding, but it's only a few thousand rows; I can find & replace the naughty characters and it shouldn't take too long to get everything smoothed out. In fact it's half done already; most of the naughty bits were dashes, apostrophes, copyright symbols, and the occasional accented vowel. Oh, and there is some Spanish in there with that funny little squiggle thing over the "n". Got those.