Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!zaphod.mps.ohio-state.edu!usc!snorkelwacker.mit.edu!bloom-beacon!eru!hagbard!sunic!dkuug!dkuugin!keld From: keld@login.dkuug.dk (Keld J|rn Simonsen) Newsgroups: comp.std.internat Subject: Re: ISO 10646 and Unicode Message-ID: Date: 12 Jan 91 18:34:37 GMT References: <00941D67.9F327C20@cunyvms1.gc.cuny.edu> <2346@enea.se> Sender: news@slyrf.dkuug.dk Lines: 21 It was stated in this group some time ago that UNICODE could be viewed as one of the most useful implementations of ISO 10646. ISO 10646 would allow many kinds of subsetting and UNICODE could be one of them, was the message in this article (which is now expired on my system). Reading the specifications of ISO 10646 and from the knowledge I have of UNICODE, it is not possible to have UNICODE as a version or subset of ISO 10646. It is true that ISO 10646 can have subsets, but they are all with encodings defined in ISO 10646, for instance a 16 bit subset has to use the encoding of a ISO 10646 plane - all the characters in the subset must use the same codes for the two bytes as the 10646 specifies. As far as I know there is no direct relationship between 10646 and UNICODE encodings. For example UNICODE uses byte codes 0-31 and 127-159, which is otherwise reserved for control characters. A feature of UNICODE is thus that it can have more (about 33 % more) characters in 16 bits than 10646, at the price of losing compatibility with ISO standards like ISO 646 and 8859. Keld Simonsen