NM 4.1- 4.2 – Networks

Blog Entries 4.1, 4.2:

Chris Anderson’s article in Wired, entitled ‘The Long Tail’, makes an interesting case for online distribution. Anderson effectively outlines how the music industry has been afflicted by a case of single-loop learning, or failing to examine the ways in which consumers are moving away from the old forms, aided by digital distribution to participate in what Anderson calls ‘The Long Tail’. TLT is about consumers being able to embrace obscurity as they choose, rather than having taste dictated by a culture of ‘hits’ or blockbusters. Using web technology to host obscure and niche artists or films allows anyone, anywhere with an interest to participate. Anderson proposes and interesting pricing system where the most popular items, are priced most highly while having lower prices down the tail to try and democratise the market, allowing niche artists to flourish and consumers to exercise more choice.

In Watt’s article, ‘Six Degrees’, he talks about the difficulty of designing whole networks that will be resistant to collapse, citing certain mass power-failures in American history as examples. The secret and difficulty, he says, is in considering not individual elements, but multiple facets of the same problem, or rather, different combinations of factors and how they might effect a given system or network. Using the human brain as an example Watt explains the problem with trying to use individual nodes to understand the whole. One neuron is a protein but somehow, one-trillion neurons make a conscious being. The question of where consciousness arises is of course still beyond modern science, although some have theories.

Similar to the human brain (an organism which we now know, almost never stops growing and runs contrary to the conventional wisdom of when I was growing up that proposed the finite neuron theory) Watt suggests we move beyond the idea of the network as ‘Pure structure, with its components fixed in time’, to a more fluid understanding of what a network can be and how it can operate. Watt says we are moving towards a ‘continually evolving and self-continuing system’, in terms of how we understand networks but also adds that the enormous complexity generated by this possibility is still being overcome.

In Barbarasi’s ‘80/20’ rule, he cites Vilfredo Pareto, the Italian scientist who came up with the idea that 20percent of the world’s people control 80percent of the landmass. The genesis for this thought was observing his peas and realising that 80percent of the peas were produced by 20percent of the plants. To account for this in the non-plant-based world, Barabasi discusses what is called a ‘Power Law’, which is a rule that describes how small numbers of things or actors can account for, or produce large amounts of output. For example, 20percent of consumers create 80percent of complaints. In the world of the internet, he describes how a random or democratic system does not exist, instead most internet traffic is funnelled through incredibly large hubs. In other words, the more popular something is, the more popular it gets because of what he calls a ‘feedback loop’. This kind of network, one that defies bell-curves in favour of power-laws (where there can be infinite discrepancy between large and small) is called a scale-free network. Meaning that scale does not apply or rather that there are no outliers but instead the network in characterised by clustering which allows many parts of a cluster or a node to communicate very efficiently with one another. The brain is apparently characterised by this same idea of ‘short pathways and clusters’.

This too explains the old adage and title of Barabasi’s other article, ‘Rich Get richer’. It is slightly disturbing for me to think that Global Income distribution might be determined by science and network theory rather than greed, but I suppose the outcome is still the same regardless. What we all know intuitively is now borne out in fact. The power of money allows making more money easier. This kind of network, one where there can be infinite discrepancy between large and small, is called a scale-free network. Meaning that scale does not apply and that unlike a bell curve which defines certain kinds of distribution like height or IQ, there are no outliers.