Character representation

Computer character sets are not adequate for encoding texts any language, even in modern English. In all likelihood this will remain the case for some time to come, though a major improvement in the situation is approaching with the implementation of the new ISO 10646 universal coded character set. Because of the present limitations, the standards committee which drew up SGML included one type of general entity for this purpose and included entity sets for a number of fields — for European languages, mathematics, and publishing — in an appendix of the SGML standard. This work is being expanded by the ISO committee and TEI.1 This article addresses the following topics: 7-bit coded character sets, 8-bit coded character sets, the Universal coded Character Set (UCS), SGML and TEI entity sets, the SGML declaration, the TEI writing system declaration, coded characters and glyphs, standards work, and a conclusion.2 The focus in this article will be upon international and national standards on this subject. Therefore, we are discussing primarily the work of the International The Chair of our technical committee submitted a draft of this article. Comments were received from David Birnbaum, Bert Bos, Steve DeRose, Berend Dijk, and Michael Sperberg-McQueen for which we thank them. The committee made revisions before submitting the final version to the editors of this issue of CHUM. In recognition of longstanding contributions to our work, the committee has elected the above mentioned contributors as honourable members, which increases our numbers five-fold. We also wish to thank the secretary general of ISO, L.D. Eicher, for permission to publish portions of the ISO standards and Jan van den Beld, secretary general of ECMA, for furnishing a complete copy of its official register of all known character sets and current versions of the ECMA standards. We also thank Edwin Smura, registrar of AFII, for supplying a copy of much of their font registry. yVakgroep Alfa Informatica, Rijksuniversiteit Groningen, POB 716, Groningen 9700 AS, The Netherlands. e-mail: [email protected] 1Some languages, such as Chinese and Japanese, require a coded character set of more than 128 or 256 characters. There are national character standards for these languages and unfortunately a number of encoding methods are in use. These languages are excluded from the discussion here. It is a separate subject and should have an article devoted to it. 2Extracts from ISO 646:1991 and ISO 8859 parts 1 to 10 have been reproduced with the permission of the International Organization for Standardization, ISO. The complete standards can be obtained from your national standards organisation or from the ISO Central Secretariat, Case Postal 56, CH-1211 Geneva 20, Switzerland. Copyright remains with ISO. European Computer Manufacturers Association (ECMA) standards are available free from ECMA, 114 Rue du Rhône, CH-1204 Geneva.