Google now has about 8% of the web under its indexing umbrella - that's a lot of pages. Now we're looking for a needle in a huge stack of needles. Enter the "Micro-Sphere" - a collection of blogs about one thing.

This opens the door for millions of silos of domain expertise. Bud Gibson pioneered this concept in learning last year with the introduction of the University of Michigan's Learning Blogosphere. It's a micro-sphere of student blogs aggregated into one searchable (and subscribable) experience. Marcomblog is a similar undertaking - also driven at the educational level. I applaud the use of blogs for this purpose - nothing better than to see the real-world intersecting with students (both parties will learn a bunch from these types of exercises).

At a technical level, there are two challenges on the horizon.

1. If I'm right, there will be an explosion of micro-sphere's. The good news is that finding information in a micro-sphere will be easier and likely more successful. The bad news - we still need to find the right micro-spheres. There's help on the way - clustering.
2. Aggregating blogs, regardless of the products used to create them, is a technical requirement. We cannot assume that authors participating in a blog-based micro-sphere will use the products or UI chosen by the micro-sphere developer. This sounds simple at the outset, but there are lots of little devils in this detail - most of which were overcome in the Learning Blogosphere demonstration. Imagine aggregating 100 different blogs whose RSS configurations each vary and whose character encodings may differ slightly. Imagine if the micro-sphere had to be secure.