Making Sense of Emergent Patterns in Networks
Valdis Krebs, T N T - The Network Thinker, November 7, 2013

This is a really good and very visual way of representing the use of social network analysis to identify communities - or 'clusters' in networks. What the article makes really clear is that there is no one way to do this, and that different clustering algorithms will identify different numbers, and different types, of cluster. There is no one 'right' way to do this; it depends on which features you think are important, and what sort of application you have in mind.

[Link] [Comment]

Facebook open sources its SQL-on-Hadoop engine, and the web rejoices
Derrick Harris, GigaOm, November 7, 2013

It is well known that one of Google's secrets was that it was designed to run on a large cluster of computers, and not one great big computer. In the last few years frameworks have been built to support massive parallel computing in clusters. One well-known framework is Apache's Hadoop, "designed to scale up from single servers to thousands of machines, each offering local computation and storage." It's how online services can start small and build to accommodate demand. Now the news is that Facebook has made its SQL database engine called Presto available as open source, which will provide significant speed advantages over previously available Hadoop database engines.

[Link] [Comment]