Blogs have become an important medium for people to publish opinions and ideas on the web. Bloggers with interest and expertise in specific domains (e.g., politics, or technology) often create and maintain blogs to publish news, opinions and ideas about those domains. In this paper, we present Spectrum, a novel blog search system that enables users to search for different points of view related to a topic from the blogosphere. Given a topic, Spectrum retrieves blog posts from bloggers with interests and expertise in various domains, enabling users to browse and compare the opinions related to different aspects of the topic. To identify bloggers in a domain category, we propose a two-layer classification model that predicts bloggers’ interests based on short snippets of posts by the blogger and posts citing the blogger. The model characterizes the recurrent interests of bloggers and the importance of the bloggers in the domain. Experiments were conducted on a list of bloggers collected from blog directories, with their snippets collected from Google Blog Search. Categorization of bloggers’ interests achieves precision of 88.4% and recall of 84.5% by micro-averaging over all the categories, outperforming a baseline algorithm which directly classifies the bloggers’ snippets. We further apply this multi-perspective blog search to explore the ecological relationship between news and blogs. The system aggregates recent popular news stories and then automatically aggregates different points of view about those news stories in the blogosphere.