The color of information
Here's a scenario. Imagine you are looking at a company document full of unstructured information. Throughout the document, information pertaining to legal matters, industrial manufacturing, pharmaceuticals, etc. is spread in a jumbled fashion. You're only interested in one aspect, but it's incredibly difficult to find the information you want without slogging through everything.

With the technology available today, you can tell the computer exactly what you want to know and the perspectives you need to understand from the documents.

The system will then return the most strongly correlated documents. It won't stop there, however. It'll also automatically reorder the content in the document along the common knowledge threads. It'll highlight in colors that you specify--say yellow for the language that tied the documents together by common knowledge, green highlighter for the legal knowledge thread, blue highlighter for the industrial manufacturing knowledge thread, and orange highlighter for the pharmaceutical knowledge thread.

The results can then be returned to you as a Web site. You can drill into a hierarchy of hyperlinked documents, interconnected by common context, each with yellow, green, blue, and orange highlighted phrases. These will draw your attention to the salient themes in the text, just as we are taught in speed-reading to read only meaningful phrases. Think of that, and you will get a sense of the power and utility such technology can bring in helping you unravel and leverage the power of knowledge trapped inside the opacity of unstructured documents.

Enhancing current search tools
Such powerful technology does not displace full text searches, but augments them. If you know what you want to know, then simply ask the full text engine to go out and find it. However, technology that can use color-coded phrases or themes to represent a knowledge thread is powerful stuff. It can change the level of the playing field. Consider the last time you did a full text search on the Internet and 5,000 items were returned.

Imagine how much easier it would've been if you'd opened one of the documents and asked the full text engine, "What does it say?" and a speed-read view of the document was quickly displayed, highlighting requested knowledge threads. You would quickly understand what the document was about because you knew the subject-area the knowledge thread represented and could follow the color codes.

It gets better, though. From there, you could ask the system to reduce the 5,000 items to a manageable sum, say the top 100 documents associated by common content. You could ask it to reflect the knowledge threads in the same highlighted colors as a selected query document. Using different query documents to re-organize the knowledge threads in the documents, you could better understand the knowledge represented.