In this article, I want to discuss how we can put auto-classification technology to work for us, with some friendly assistance from XML (eXtensible Markup Language). After all, we've worked hard during this series of articles to create auto-classification. Now it's time to taste our cooking. First, however, let's review the importance of color-coding knowledge threads, so that we don't go off half baked.

Reverse-engineering the speed-reading process
Color highlighting trains our eyes to speed-read document text so we can focus on the "fall-line" of subject knowledge threads as they weave their way through the document. Meanwhile, we absorb the surrounding context through our peripheral vision. This permits us to anchor and associate an increasingly accurate meaning of the knowledge thread at a much faster rate than normal reading would allow.

With the technology I described last time, we can cross-associate the interrelated meaning of different knowledge threads via the visual color-coding that is interwoven throughout the document, creating a latticework of interlinked themes describing the knowledge relationships across document collections. This multidimensional contextual unfolding is a key underpinning of not just speed-reading, but also Power Learning, which this technology facilitates.

What it's good for
Color highlighting can be driven by automatically generated JavaScript, XML, and XSL (eXtensible Style Language) style sheets stuffed inside the header of HTML pages. As a result, by providing a library of JavaScript functions and XSL style sheets, we can drive the thematic XML tags to do almost anything we want.

For example, we can color highlight the tagged text by subject category, color the font of tagged text for printing, or magnify tagged text by two or three times for those who are visually impaired. We could also invoke a pager, send an email, or initiate a workflow event when a certain combination of themes fits a pattern of a specified strength. Additionally, we could send an event to another process that is listening for an array of key events. Java Beans are based on this concept and BML (Bean Mark-up Language) provides another scripting tool that can raise the power of this solution.