TimesFrank wrote... > Again, I don't care so much whether > these particular characters are encoded, but they illustrate a point worth > making, namely that unifications that work in the GUI world don't > necessarily work in an environment where we must use a fixed-pitch font. I think you're missing the point a little. Over the past day, I believe we've established that these are the same things, so they shouldn't be encoded again; and we've established that you don't need two sizes of ordinal indicators. Right? The size of the glyphs in particular fonts, GUI or plain old dot-matrix, is just irrelevant to these characters' identity, since they've so obviously been established. It escapes me why you then say: > Yes, but they would be too small. The SNI glyphs are full-size base > characters, but the ordinal indicator glyphs are superscripts. They're not necessarily superscripts -- and they can be seen with or without little underlines, too. See the attached example of the same string in several fonts... In your font these take up a whole character cell, top to bottom, and have underlines. I have a Courier fixed-pitch font here, and it's got both of these things in it, and I use it in a fixed-pitch VT100 terminal emulator. You have another font there, on some terminal, and the glyphs are a different size/shape, or have a different number of dots. Well, your "A" character is probably a different size, too. But we don't have to encode two "A" characters. This is the same situation. > E0B3 Latin small letter a with underbar SNI Math 04/04 (2) > E0B4 Latin capital letter O with underbar SNI Math 04/09 (2) There's nothing preventing the emulator from mapping the existing feminine and masculine ordinal indicators to whatever byte-code you need for these terminals, then printing them in whatever font you're using for that terminal. Rick -----