TITLE: Transforming Documentation from the XML Doctypes Used on the Apache Website to DITA: A Case Study

AUTHORS(s) & AFFILIATION(s): Donald Leslie, Cambridge Advanced Technology Group, IBM Research, Cambridge, MA

KEYWORD(s): XML, XSL, XSLT, stylesheets, document transformation

PRESENTER / CONTACT PERSON: Donald Leslie

CONTACT EMAIL: donald_leslie@lotus.com

ABSTRACT:

This paper reviews the structural and procedural issues in transforming the Xalan doc set from the XML doctypes used by the Apache XML website to a preliminary version of the DITA doctypes developed by an IBM working group. Xalan is an XSL transformer that Lotus/IBM donated to the Apache open-source XML project in November 1999. The documentation is written in the XML doctypes used on the Apache XML website and transformed to HTML for display on xml.apache.org. As described in detail in other presentations, DITA (the Darwin Information Typing Architecture) provides a modular structure of nested topics that individual documentation groups can extend (with XML inheritance structures) to meet their particular needs. To present an ongoing series of Xalan releases to a number of IBM product groups, we decided to set up an automated build procedure to transform the Xalan XML documentation to DITA, and from DITA to HTML and PDF. This build procedure uses a simple XSL stylesheet to transform a "book" document from the Apache XML into a DITA table of contents. For each "document" in the book, the stylesheet calls an XSLT extension element, which sets up a series of transformations with additional stylesheets to transform the Xalan documentation to DITA. The build supports three methodologies for publishing documentation:

1. Transform the DITA documents to HTML, and serve the HTML
2. Use a servlet to dynamically transform DITA documents to HTML in response to client requests for DITA documents
3. Use Xalan and Apache FOP to transform from DITA to PDF.

Last modified October 28, 2001 by Scott Tilley.