From fdc@watsun.cc.columbia.edu Fri Sep 4 17:32:50 1992 Received: from watsun.cc.columbia.edu by dkuug.dk via EUnet with SMTP (5.64+/8+bit/IDA-1.2.8) id AA13598; Fri, 4 Sep 92 17:32:50 +0200 Received: by watsun.cc.columbia.edu (5.59/FCB) id AA27205; Fri, 4 Sep 92 11:30:53 EDT Date: Fri, 4 Sep 92 11:30:51 EDT From: Frank da Cruz To: "Jurgen Bettels" Cc: Multi-byte Code Issues , i18n@dkuug.dk Subject: Re: request for feedback on character set identification proposal In-Reply-To: Your message of Fri, 4 Sep 1992 10:46:53 -0500 Message-Id: X-Charset: ASCII X-Char-Esc: 29 In adapting the Kermit file transfer protocol to handle conversion of character sets, we have addressed many of the issues raised by Jurgen Bettels in his paper, "Identification of Character Collections and Encodings". Our basic conclusion was that only registered ISO standard character sets should be used within the transfer syntax, and local (often proprietary) character sets, and conversion between them and a standard transfer character set, were the business of the application software on the local computer. A simple notation was developed to allow the sender to identify the transfer character set to the receiver. A major issue concerns the invertible translation between a proprietary character set, such as IBM code pages or the character sets used on the Macintosh, and the corresponding standard character set, such as ISO 8859-1. IBM is (to our knowledge) the only major originator of proprietary character sets that has published official translations. In the absence of official translations, the maker of each application is left to devise her own, resulting in multiple incompatible translations. The paper "A Kermit Protocol Extension for International Character Sets" describes in detail the problems of character-set identification and translation and our approach to solving them. It can be obtained via anonymous FTP from watsun.cc.columbia.edu [128.59.39.2] as kermit/e/isok7.txt, or, on BITNET/EARN, by sending e-mail to KERMSRV@CUVMA containing the text "MAIL ISOK7.TXT". The techniques described in the paper have been successfully implemented in Kermit software for the IBM PC, UNIX, VAX/VMS, IBM mainframes, and many other computers. Christine M. Gianone Frank da Cruz Columbia University, New York City