XML was originally intended primarily to support the authoring and production of technical documents and this is still an important application of XML. While there are many compelling reasons to use XML for technical documents, such as the ability to produce many different output and deliverable forms from a single set of source files, the ability to re-use the same information object among several units of publication is a particularly important one because of its implications for increased information accuracy and reduced authoring and translation costs. Except for the weak and dangerous external parsed entity mechanism, the XML specification provides no direct support for re-using information objects. Thus, authoring support and production systems must provide that support themselves. Often this is done using proprietary or purpose-built mechanisms, such as those provided by typical XML-aware content management systems. However, the W3C's XML Inclusions (XInclude) specification provides a basic mechanism that standardizes the way that XML information components are linked in order to establish use-by-reference (re-use) relationships. This simple mechanism is relatively easy to implement using most XML processing technologies (XSLT, Java, etc.). The use of XInclude provides a standard, generic, non-proprietary facility for creating and managing compound documents.

This paper first discusses the basic technical and practical aspects of XML information re-use. It then describes basic approaches for using XInclude in typical authoring support and document production systems to create and manage technical documents organized as compound documents that include re-used components. Discusses some ways that the base XInclude specification can be extended to better support the needs of technical document authoring, in particular, the use of specialized element types in place of the standard xi:include element type. In addition to the practicalities of implementing XInclude in XML editors and processing tools, the paper also discusses information management issues raised by the use of shared information objects and outlines information management policies and work practices that can be used to address these issues. Finally, the paper discusses, in general terms, real authoring support and production systems that use these techniques for managing use-by-reference.