Discovery Informatics???

While I was glancing through e-news, I came across a new term which is 'Discovery Informatics'. I know 'Discovery' and 'Informatics', but what is this Discovery Informatics? I din't even get a small idea of what it is actually about. Hence I began my research on this topic and thought I would present it in a nutshell here.

Discovery Informatics is a rational approach to discover new materials and formulations. It's a combo of computer and Mathematical Science which extracts new information from huge streams of data. It is different from what modern computers do. A modern computer which can hold terabytes of data is built to do what is told. But DI suggests what more can it do.

Theory is always abstract. An example always extend its meaning. In this context of DI, here is an example. Imagine a computer that can read torrents of data pops up one day and says 'There's an interesting correlation between late night diaper sales and late night pop sales in convenience and grocery stores. Would you like to investigate this further?'

Another one: Imagine reading the entire Library of Congress (the largest library) each day in a search for subtle patterns that could lead to new discoveries. Except it's more difficult than that. Since many of the best discoveries come from unanticipated observations, you must be able to spot these patterns on the fly, sometimes without even knowing what you're looking for.

To put up simply, when you search something in any search engine, the results provided will be related to your search. But DI is something that would suggest you, what else can you look for in relation to your search?

The field of Informatics has impressive rainbows to chase - This is what scientists think.

Concepts of DI still looks hazy to me and it will, unless it takes a great leap in this techno arena and make people like me understand it better.