Thursday, June 27, 2013

Implementation of the observer mechanism.

As I write this entry, I am taking some time away from coding the observer mechanism.

This observer mechanism, like any other mechanism that models and moderates the behaviour of devices within Clique Space is a medium. Hence, the observer mechanism must exhibit the structural components necessary for the representation of any medium in Clique Space: the abstract classes for the Media Profile, the Connection, and the Participant require implementations that can be instantiated.

However, the observer's medium is a medium that models a functional subset of the behaviour of the Agent Device, and hence, there are other structures that need to be implemented.

The Clique needs a subclass on the Agent Device: the observer Clique must contain only observer Participants. This Clique is importantly distinct from a Clique that models the way other media are modelled and controlled in that the observer Clique indicates which Agent Devices have knowledge of a specific component. Every observable component has an observer Clique when an Agent Device responsible for the creation and management of an observable component transmits knowledge of this component to another Agent Device. The Agent Device that created the observable component is the observer Clique's Owner, although in an observer Clique, like any other Clique, ownership can cede to another Agent Device, and so like any other Clique, the observer Clique is not anchored to any one location.

There is at least one minor issue with this model. If an observer Participant is a component (all of Clique Space's Elements are components), and if the observer Participant, in being a component other than a Clique, is observed, then how is that observer Participant observed? One cannot create another observer Clique because the creation of another observer Clique (Cliques - including the observer Clique - are components, but they are the only type of component that is not observable) would involve the creation of more observer Participants. This is an infinite regression, and so to avoid this, any specific observer Participant's observer Clique is the Clique which contains the observer Participant. This solution not only appears to be a convenient way of avoiding the infinite regress, but it appears to have a certain (though circular) logical necessity; in this mechanism, I feel echoes of the same necessarily visceral emotional base (the ego) that I seem to observe when I observe myself.

With this mechanism, component transmission can be tracked and controlled by a collection of Agent Devices. It appears that this mechanism is crucial to the cooperative activity of Agent Devices, and I believe this mechanism will have something to say about cognitive function in biological neural systems.

Anyway, implementation awaits...

1 comment: