DITA2InDesign Project Overview

The DITA2InDesign plugin is a processor that takes DITA content (maps or individual topics) and, through a series of steps, composes that content as sophisticated pages using all the features of the Adobe InDesign product.

Adobe InDesign is a desktop publishing application designed to meet the typography and layout requirements of publishers. It is widely used for highly-designed publications including glossy magazines, trade books, marketing brochures, as well as more mundane publications. It is not as widely used for technical documents because it is not, itself, an XML authoring environment, although it does provide some XML support. In the marketplace, InDesign is largely replacing Quark XPress as the dominant interactive layout system.

InDesign is part of the Adobe Creative Suite family of products, which includes products such as Illustrator and PhotoShop, as well as many others. The current Creative Suite version is CS3. Thus when talking about InDesign versions you must specify the Creative Suite version. InDesign CS3 is the current version.

InDesign provides XML import features. However, these features are not sufficient by themselves to import and layout out XML as sophisticated as the typical DITA-based publication. In particular, the XML import feature does not automatically create new pages to accommodate arbitrary amounts of content nor can it apply different page masters based on element in context rules. It does provide some handy features for making it easy to map elements to specific character and paragraph styles.

This means that getting DITA-based content (or any non-trivial, variable XML) into InDesign requires custom scripting to manage the creation of new pages with the appropriate page masters and the placement of content into linked frames on those pages. That scripting, while non-trivial, is not too hard in CS3, but it is, by its nature, document type and layout specific.

Getting XML into InDesign also requires that the XML be carefully prepared for import as follows:

All whitespace must be reduced to just those characters that should be displayed as content. In addition, all newlines must be removed except those that should be taken by InDesign as paragraph markers.
Elements must be added to act as anchors for generated text. The text itself can be generated using InDesign styles but there must be an XML element to apply the styles to.

In addition, the XML preparation can do any re-ordering or additional grouping, although InDesign's XML mechanism is flexible enough to allow structures to be laid out in an order that is different from its XML source order.

Given these aspects of InDesign, a process that automates the import and layout of XML into InDesign must have the following components:

An XML-to-XML transform that prepares the XML for import into InDesign. Because of limitations in InDesign's built-in XSLT engine, this transform must be run outside of InDesign.
A set of InDesign scripts that manage the XML import process and implement the XML-to-InDesign-Component process, creating new pages and frames as necessary to accommodate all the content in the incoming XML.
An InDesign template document to serve as the import target. Because there is a mapping from XML elements in context to InDesign styles and page masters, there is a necessary dependency between the XML-to-XML transform and the scripts and the templates as the transforms and scripts must know what the style and page master names are.

For the DITA2InDesign plug-in the plan is to define a set of "standard" InDesign component names that can then be used in any InDesign template to enable flowing of DITA content into that template. This allows the XML-to-XML transform and the import script to be completely generic, such that using a different layout for a given publication is simply a matter of changing the typographic details in the template without changing the names of things.

In the case where the existing templates don't provide a layout distinction that you need, you would create a modified template that adds the named styles or page layouts you need and then use the same sort of extension technique used for the Open Toolkit HTML plug-in to extend the XML-to-XML transform to apply your new style and page layout names where needed. Ideally this would require just a few lines of relatively simple XSLT code.

In the case where you needed more sophisticated automatic placement or sizing functionality you would need to extend the InDesign scripts, which would be more involved.

For layouts that require more sophistication in layout automation than the DITA2InDesign plug-in can provide the Typefi product would probably be an appropriate solution. Typefi is an InDesign plug-in that provides very sophisticated automated layout heuristics that go far beyond what could be done for the DITA2InDesign plug-in.

When implemented, you would use the plug-in by first applying the XML-to-XML-for-Import transform using normal Open Toolkit mechanisms. The result of that transform would be one or more XML instances, each instance representing the content to be used to create a single InDesign document. You would then use the plug-in provided InDesign scripts to import the generated XML into InDesign. You would then use InDesign's features to adjust the formatted result as needed. This import process could be started automatically using InDesign's command-line. It could also be integrated with InDesign server to provide a fully-automated batch process.

Note that there is no plan or intention to guarantee the ability to get DITA XML back out of InDesign with this plug-in. It is intended to be a one-way process from authoritative DITA source to InDesign-composed pages. However, to the degree that the original DITA markup is preserved in the generated XML it should be possible to export XML back out of InDesign following text changes made in InDesign and run a transform that essentially reverses the original augmentation. However, that process is not in the scope of this plug-in at this time (of course, if somebody wanted to contribute that functionality it would be very welcome).