David Peterson <[log in to unmask]> wrote: > The IPA character [a] is different from the Roman character "a". > The Unicode character for both, though, is identical. This could > lead to big problems--especially if this is changed (which it ought > to be) in the future. Mark J. Reed, On 28/01/2010 12:37: > Huh??? The IPA symbol <a> is in no way different from <a>. Unicode > encodes symbols, not semantics. The various places that identical > symbols show up twice is solely due to backward-compatibility issues > with pre-Unocode text. All of the IPA symbols that are identical with > Latin letters are actually identical with them. What do you gain by > treating them distinctly? Garth Wallace, On 28/01/2010 19:06: > How are they not the same character? They look identical. Only the > context is different, but if you distinguish characters by just > context then you'd have to have separate characters for English <a>, > French <a>, Spanish <a>, Polish <a>... David Peterson, On 28/01/2010 20:32: > I thought this was the whole point of Unicode. Notice that there's > little difference between the Roman "a" and the Cyrillic "a", yet > the Cyrillic "a" gets its own codepoint. Just how different does > "different" have to be? > > Another problem (aside from the caps) is italicization. Not all > characters italicize the same way. Most IPA characters don't. > Also notice that if you italicize "a" in many fonts you get the > closed variant, which is an entirely different character in IPA. Mark and Garth are right about Unicode characters, of course, but I agree with David that IPA <a> and roman <a> are not the same -- e.g. IPA <a> must be double-decker and roman <a> needn't be. More generally, I would maintain that characters are defined relative to a particular graphological system (by which I mean both scripts and systems like IPA), and that the (system-specific) definition includes both form (e.g. alloglyphic range) and meaning. I wish Unicode had adopted a system in which character codes consisted of a first part that identifies the graphological system and a second part that identifies the character within that system. I think such a scheme would have been far freer from controversies, disputes, analytical solecisms, etc., much more straightforwardly extensible, much less cumbersome, and so on. --And.