Yves said: > > > Are there any reason we can't remove case mapping from > > > our spec? (Other than something like "because it is a > > > natural English convention..." :-) > > > > Yes, there is a good reason: removing the table would introduce confusion. > > But we already know that there may be some confusion at some point, as you > acknowledge a couple days ago, if some later revision of Unicode introduces > new characters with case distinctions (guaranteed for 3.2 IIRC).
This is true, but at this point the characters with case distinctions introduced are truly rare (an archaic Greek koppa not used in modern Greek, four Cyrillic letters used only in Kildin Sami, eight Cyrillic letters used only in Komi in an obsolete orthography no longer in use). And the candidates are getting scarcer and rarer. > So this > really needs to be pointed out, so that people realize that it may break at > any time depending on the applications they use. (The optimistic in me > thinks that applications will strive to offer timely updates in order to > offer a better user experience w/o the confusion.) I agree. By the way, a propos of the note that originally started this thread, has anyone else actually realized that there really is a http://www.cobet.ru/ And on the title of the home page, this business consultancy group playfully posts "COBET.py", illustrating the cross-script dual pun on "Soviet" and the Paraguayan ccTLD. (They apparently haven't bothered to also create a www.cobet.py in reality.) And going back to points made by John, Mark, and others, one should note that Unicode is not the issue here. *Every* major character encoding used in Russia includes both ASCII and Cyrillic, and the same well-known overlaps of Latin letters and Cyrillic letters occur in all of them, from ISO 8859-5 to Code Page 855 to Windows 1251 to KOI-8 to Mac Cyrillic. Unless Cyrillic were eliminated from domain names altogether (which I doubt the Russians, Ukrainians, Bulgarians, and Serbs would be very happy about), these kinds of visual overlaps will occur. And as Mark said: > I don't think a solution is possible in DNS. --Ken > > YA > > >
