Typesetting SGML Documents Using TjjX
Andrew E. Dobrowolski ArborText, Inc. internet: aedQarbortext.corn
Since its publication as an international standard in 1986, the Standard Generalized Markup Language (SGML) has become a preferred document-markup standard within many industries. Many users have developed their own document type definitions (DTDs) that define the elements (tag sets) for their documents. However, if SGML is to become a universally accepted standard of document interchange, then a standard way of specifying formatted output and a means of producing that output will be needed.
The U.S. government's Computer-aided Acquisition and Logistic Support (CALS)initiative selected SGML as the standard for text interchange. The output specification section of the CALS standards proposed the Formatted Output Specification Instance (FOSI) as the means of formatted output specification interchange.
TJ$ can be used as the formatting engine to implement FOSI-based formatting. But without extending w,not every FOSI formatting request can be fulfilled. Conversely, certain
capabilities cannot be formulated in terms of FOSI characteristics. However, a FOSI/m-based formatting system would be a major advance towards fulfilling the document interchange needs of a growing community of SGML users.
Document Interchange Standards
In the past ten years, w has become a well known and widespread language for typesetting technical documents. From its original base of universities and colleges, it has spread to such an extent that people in industries with only incidental needs for publishing have heard about it. A large part of w ' s appeal comes from its portability, since the program is in the public domain and has been ported to quite a number of operating systems. There is no standard for the way a document is "marked up"; this is dependent on the macro package used. Given the right macro package and fonts, the formatted output of two different
implementations on two different machines will produce identical results.
By contrast, generic markup systems identify document structures without making assumptions about the end application of the document. This makes the same document useful to various pro- grams and for various applications. Generic markup has been around in several flavors for over ten years.
These dissimilar flavors were a hindrance to its util- ity. To remove this hindrance and to promote the portability and acceptance of generic markup, an international standard (IS) specification for generic markup was established in 1986. Since then, SGML (Standard Generalized Markup Language) has be- come extremely important to industry, especially in areas where huge quantities of data have created a document-management nightmare. Today a large number of programs can read and write SGML on a variety of platforms.
The U.S. government's Computer-aided Acqui- sition and Logistic Support (CALS) initiative gave SGML additional clout by selecting SGML as the standard of text interchange between the Depart- ment of Defense and its subcontractors. However, SGML contains no information pertaining to the printed representation of a document or to the meaning attached to the markup. The compan- ion standard to SGML that addresses standardized formatting specifications, the Document Style Se- mantics and Specification Language (DSSSL), is
TUGboat, Volume 12 (1991), No. 3 -Proceedings
of the 1991 Annual Meeting