From keld@dkuug.dk Sat Apr 18 00:44:41 1992 Received: by dkuug.dk (5.64+/8+bit/IDA-1.2.8) id AA00258; Sat, 18 Apr 92 00:44:41 +0200 Date: Sat, 18 Apr 92 00:44:41 +0200 From: Keld J|rn Simonsen Message-Id: <9204172244.AA00258@dkuug.dk> To: wg15rin@dkuug.dk Subject: SC2 resolution to SC22 X-Charset: ASCII X-Char-Esc: 29 I had an action Item to find the SC22 and SC2 resolutions on 10646, here is the SC2 answer to the SC22 liaison statement: (adopted at the SC2 plenary in Rennes, France, october 1991). Liaison statement from SC2 to SC22 in answer to their comments on DIS 10646. JTC1/SC2 wishes to thank SC22 for its thoughtful comments on the the working document for the 2nd DIS 10646. SC2 appreciates SC22's concerns, including those affecting the C language. In response SC2 notes: a) ISO/IEC 10646, UCS, is designed as a 4-octet code (UCS-4) with a 2-octet Basic Multilingual Plane (UCS-2). Thus, the UCS is not a single-octet based code and is not intended to be implemented as a 1-octet C "char" data type (where the NUL octet terminator is used). SC2/WG2, in response to comments from National Bodies, decided to allocate the characters of ISO 8859-1, in both UCS-2 and UCS-4 by zero extension of their 1-octet representation. This maintains the numerical equality and enables future implementations of UCS-2 and UCS-4 as C "wchar_t" data type (or a possible similar new data type). As an alternative ISO/IEC 10646 defines the "UCS Transformation Format" (UTF) for a transformed variable-length representation of UCS to avoid data with C0 and C1 control octets. Data transformed from the UCS by the UTF can be processed as C multi-byte character strings. b) SC2 agrees with SC22 that the UCS encoding of character data will coexist with other coded character sets such as ISO/IEC 646, ISO/IEC 8859 or others. The UCS standard defines escape sequences according to ISO 2022 to announce the UCS data and to return from the UCS back to the ISO 2022 environment. (Courtesy of Johan van Wingen).