Book work

We met for the first proper day's work back at the beginning of January at the Evans Studio farmstead in the Forest Of Dean. After dealing with the practicalities - setting up a work schedule, responsibilities and objectives - we started in on the theory and boy is there plenty of that. We are looking at non-linear dynamics within pervasive media systems, specifically swarms. It sounds a bit dry but it is about enabling intuitive cooperation amongst people and has loads of potential in gaming, social action, low carbon economies - all sorts of areas.

Simon J is the project partner who has done the backgorund reading on the area and he had a go at explaining some of the ideas in Manual De Landa's book One Thousand Years of Non Linear History. Simon J is keen on De Landa, whose work tries to cross breed concepts from the Natural Sciences with those more familiarly associated with the Social Sciences - for instance the notion of historical processes in geology, rock strata forming over time, and geological processes within history, social classes forming through sedimentation. Of course, this approach is not without its critics but it does offer a usefully formulaic approach to social dynamics, suggesting that the embodiment of these processes within computer systems would be possible. Again, it raises the familiar question of computers not becoming more human-like, but humans being invited (compelled) to become more machine like.

Alongside this general reading has been a more specific research path that we have followed over the last month. There are a number of applications, demos and toys online that use swarm dynamics to generate graphics, such as levitatedTentacle1 and Swarm Box We are analysing the code of these programs to find out how other programmers have have understood swarming and realised it in code.

An early objective of our research has been to identify the fundamental components of swarming. We started out with an idea that three areas would be relevant:

space/proximity - the geographic dimension.
communication - between agents
persistence - time.

We then realised that persistence (ie. time) was really a property of the other two components and not something in itself. So we arrive at:

Communication is optional or compelled/controlled, but never absent or forbidden.
Geography - size of game world has as yet undefined relationship with number of participants.

But isn't geography really another way of discussing communication? Swarm behaviour involves a range of relationships between agents; chemical messages between ants, for instance, or spatial awareness between fish in a shoal. In many cases communication is indirect, the fish merely try and stay within a certain distance of a given number of others. The key principle, however, is the communication of information, in the fish example it is a visual clue at offers the relevant information. What matters is how far the given form of communication can travel, so geography (or extent of swarm space) is a property of communication. We can see this with Seekers - changing the call radius (range of communication) and number of seekers (agents) has a powerful affect on swarm efficiency and stability.

Clearly, modern technologies of communication will have an impact on the potential of swarm applications in social groupings, enabling swarm like behaviour over extensive geographic space. Pervasive media will enable highly dynamic groupings out in the world. It is in this intersection of ubiquitous communication and physical space that we want to work. We are both very keen on creating networked experiences in varied geographic space, to get games away from the desktop or console in the way the Blast Theory have so successfully done.

So there we have it: communication. Does this insight help much? It is a huge area and without an exact philosophical definition, impossible to explore. However, we have decided to accept a very broad definition of what human communication is and to focus not on the content but it's effect. A swarm is the manifestation of emergent behaviour from within the relationships of a group of individual agents. The behaviour is not dictated from any top down control but emerges out of communication, direct or indirect, between agents. Our goal is to establish what is needed to create emergent cooperation between people and to use these insights to build a pervasive media tool for other people to create games and experiences based on cooperation between people.