For some, perhaps the real question is Why XML? (or What is XML?), but assuming you have answered those questions (and are using XML), then the next step is to locate an appropriate data model for your content. This is an important step because you will spend a lot of time and money developing processes and selecting tools to support your chosen data model. XML, by definition, is extensible and allows you to create any valid structure that suits your needs, but before you decide to develop your own, consider the pre-existing options (see Don't Invent XML Languages for a discussion on why not to develop your own). If you can leverage and build on top of someone else's work, why not?

DITA is a data model for authoring and publishing topic-based content. It was developed by IBM for internal use and has since been released to the open-source community (now under the guidance of OASIS). This architecture and data model were designed by a cross-company workgroup representing user assistance teams working throughout IBM. After an initial investigation in late 1999, the workgroup developed the architecture collaboratively during 2000 through postings to a database and weekly teleconferences. Since that time IBM has migrated thousands of pages of content to DITA.

But, why DITA?
Well, assuming your content fits into the topic-based data model, DITA's increasing popularity means that more and more authoring and publishing tools will be developed to support that model. The DITA Open Toolkit allows you to generate many popular output formats (HTML, HTML Help, PDF, Java Help, etc.) from DITA-based content. If you develop your own data model, you'll have to pay to develop those transformations. DITA's modular architecture, supports efficient reuse of content at the word, phrase or topic level. DITA also has the concept of "specialization," which allows you to develop elements of your own that are based on core DITA elements. This helps you to customize DITA to support your particular types of content while continuing to take advantage of the base DITA tools and transformations.

Learn more
The following articles provide additional information:

• Going DITA - by Constantine Hondros Oct. 12, 2005 03:21 PM
  This blog provides insights from a software developer currently involved in a project to migrate a mission-critical document corpus to the DITA architecture.
   
• The Holy Grail of Content Reuse: IBM's DITA XML - Cover Pages, April 25, 2003
  Robin Cover offers a clear, concise overview of DITA basics and benefits in this report.
   
• Is DITA Going to Tip? - Center for Information-Development Management, JoAnn Hackos, PhD, Dec. 2005
  This article explores the reasoning behind the adoption rate of DITA and other publishing technologies.
   
• Why use DITA to produce HTML deliverables? - IBM developerWorks, 07 Oct 2003, Updated 28 Sep 2005
  This article described advantages DITA provides for producing HTML content -- including easy global changes, portability through standards, superior linking and Web management, conditional processing, content and design reuse, and better writing through focused content.
   
• Introduction to the Darwin Information Typing Architecture - IBM developerWorks, 01 Mar 2001, Updated 28 Sep 2005
  This document is a roadmap for DITA: what it is and how it applies to technical documentation.
   
• DITA webinar, "What You Should Know About DITA" - Arbortext, April 28, 2004
  DITA Q&A .. What is DITA?; What’s so great about DITA?; Does DITA allow for use of a custom DTD?; How much of the flexibility and reuse demonstrated during the presentation was due to DITA and how much simply to having XML source?; and more.