Friday, 26 June 2009

Octopus gonna fuck you up diva boi

fo' real

Modular organisation

Another useful quote from Linked, this time on modular organisation in complex networks, a particular focus of mine these days:

Hartwell and colleagues jumped into the middle of this debate by proposing that cells sustain a multitude of functions - i.e., multitask - thanks to a discrete modular organisation. According to this view, the network behind the cell is fragmented into groups of diverse molecules, or modules, each module being responsible for a different cellular function. As modules are connected via a few links to other modules, the network behind the cell is similar to Granovetter's circle of friends, where those within the same circle know each other well and communication with other circles is maintained by a few weak ties...Modularity is a defining feature of most complex systems. Indeed, departmentalisation allows large companies to create relatively secluded groups of employees who work together to solve specific tasks; the Web is fragmented into heavily interlinked communities of Webpages whose creators share common interests; modularity in our intellectual and professional interests allows Amazon to offer book recommendations inspired by the reading patters of people within a comparable intellectual module; a modular computer design allows us to replace the old bulky screen with a flat panel display without redesigning the whole computer. Yet, a modular architecture is at odds with everything we have learned so far about complex networks. Most networks, from the cell to the World Wide Web, are scale free, held together by a few hubs. By virtue of the many links hubs possess, they must be in contact with nodes from numerous modules. Therefore modules cannot be that isolated after all, resulting in a fundamental conflict between the known scale-free architecture and the modular hypothesis.

So the obvious question for me is how to model a modular architecture for da wissum, so it can do specialised tasks more effectively. Answers (with diagrams) on a postcard.

Sunday, 14 June 2009


Currently reading Linked, a book about networks, mainly an introduction to the work behind all that hoo-ha. Interesting and easy to read, and starting to provide some useful nuggets. Like this

The surprising discovery of power laws in the Web forced us to acknowledge the hubs. The slowly decaying power law distribution accomodates such highly linked anomalies in a natural way. It predicts that each scale-free network will have several large hubs that will fundamentally define the network's topology. The finding that most networks of conceptual importance, ranging from the World Wide Web to the network within the cell, are scale-free gave legitimacy to hubs. We would come to see that they determine the structural stability, dynamic behavious, robustness, and error and attack tolerance of real networks. They stand as proof of the highly important organising principles that govern network evolution.