Technical information publishing is all about putting the right information in the hands of users at the right time and without a lot of clutter. But the problem is that each user is different, and one oyster’s pearl is another oyster’s kidney stone. The metaphor applies to computer-based training (one man’s important lesson is another man’s boring repeat), maintenance and performance support systems (one woman’s critical step is another woman’s insulted intelligence), and many other cases where technical information may need to be not only customized but individualized to be useful.

Recent models for technical information delivery are dynamic, providing data at an appropriate user level for every topic in the content, and even blurring the line between performance support and training by enabling software to interject training content “just in time.” To support this model, core ideas of reuse come into play because screen composition on-the-fly requires reusable content to be packaged in different ways for different users and different situations. Dynamic content selection, then, can be viewed as an extension of the common definition of reuse.

In this paper, we discuss the idea of a content pool (a generalization, for example, of the DocBook “Information Pool”) where collections of useful information float around sunning themselves until they are needed in a particular context or presentation. To make the dynamic content selection technique work, we use a combination of hierarchically structured content and automated links. Using an XML markup scheme that we will discuss, our applications ask members of the content pool to “speak up” if they are relevant. And of course, being extensible by nature, they are happy to oblige.

The key to dynamic reuse is defining how information applies in different contexts. Our technique gives authors a method for defining what an element of content is all about, but doesn’t require them to define explicitly how to package it. This gets authors out of the assembly and formatting business, and back to creating content.

Our markup strategy has two parts: identifying Object Types and defining Roles. Authors are not required to build explicit links, only categorize the content they create. And yet the occurrences of an Object Type reveal the relationships inherent in the document. The application software looks at related Object Types and identifies the Roles where they appear, then automatically builds a network of links to help users traverse from one context to another.

The benefit of dynamic reuse is that applications can quickly take on multiple personalities by changing how relationships are revealed and how content is packaged based on those relationships. It’s simply a matter of replacing XSLT scripts on demand at runtime, which we enable through the use of reusable software components. Thus, an application can change its look, its organization, its navigation, and it’s content to accommodate different situations, preferences, or users.