Before you can design and implement your own configurations and specializations you must have a basic understanding of configuration and specialization.

DITA is designed to let you design new markup and new document types in a way that allows any general-purpose DITA processor to usefully process documents that use the new markup. This in turn enables blind interchange of DITA documents, because DITA users knows they can use and process, at least minimally, any conforming DITA document they get from any source. In particular, in the context of a map, you can combine topics of any type and get usable, if not optimal, results from any general-purpose DITA processor.¹

The specialization feature of DITA, the ability to have your own markup design while still ensuring blind interchange of your content with other DITA users, is unique among all currently-existing standard XML applications (and most, if not all, private XML applications).

Specialization is the one truly unique and distinguishing aspect of DITA. No other aspect of DITA is exclusive to DITA. All of DITA's modularity features—maps, topics, key-based addressing, etc.—can either be found to one degree or another in other XML applications or could be added to those applications without too much trouble. This is not to discount the value of these features of DITA—they represent deep thought and years of practical experience and are quite valuable in themselves—but they are not distinguishing in the way that specialization is.

Even if you have no use for any aspect of DITA having to do with modularity or reuse, you still have a use for specialization simply because it enables reliable interchange in a way that no other XML application does. Even if your only interchange partner is your future self, DITA still offers dramatic and compelling advantages.

In short, one can see DITA as an architecture for the management of XML vocabularies.

These tutorials show you how to apply the architecture to specific types of requirements.