Megite, a "memetracker" that finds hot news topics from around the blogosphere, are launching their 2.0 version today. The changes include a "new blog auto discovery algorithm" that makes the site more inclusive and doesn't favor the so-called "A-list" bloggers as much. There's also a better ranking algorithm - they ascertain how important a story is based on the blog's importance, the relevance of the content, linkage, recency and more. The clustering algorithm has changed too, meaning that Megite can more accurately group posts that talk about the same topic. Other changes include more frequent updates, new topic areas like Entertainment, Science, Food, Fashion and Chinese and a "River of News" view that aims to display a continual flow of interesting content.
There are plenty of services that track the importance of stories based on linking behavior and other factors (see Rise of the Memetrackers), but all of them have their flaws - most seem to be dominated by a small set of regular blogs that are favored by the algorithm, for instance. Meanwhile, human voting sites like Digg have been notoriously gamed - Digg's algorithm was recently changed so that the top users didn't have such overwhelming power. But that resulted in a backlash, and it seems that there's been some backpeddling to appease them. Meanwhile, there have been at least two minor acquisitions in this space recently - the Digg-like site Reddit was snapped up by Conde Nast, while the unpopular Blogniscient service was snapped up by an even less buzzworthy site: TopTenSources.