Monday, September 3, 2012

How is information distributed through Clique Space(TM)?

This question is a bit obvious, but it is one which I had no specific answer for when I put my patent together.

When I conceived the Clique Space(TM) data model, I took as an article of some speculative faith, the premise that I could work out the technical considerations of how this model would be used to realise the mechanism that actually allows one Agent Device to exchange information with one or more others, as well as how one Agent Device exchanges Clique Space Element information for external device state information. I had hypothesised that mechanisms would emerge as a result of the deliberations necessary to get from the concept recorded in the patent, to its implementation recorded in code.

I envisaged that perhaps some degree of similarity between Clique Spaces and biological nervous systems would emerge, and indeed, this appears to have happened in at least one example: the emergence of the logical synapse necessary to open a channel for exchange between two Agent Devices. Although the synapse is reasonably stable, my current deliberations will have an impact on this mechanism because the changing form of the information exchange will also change the way Agent Devices engage and disengage, forming and disbanding bipartite Cliques which represent these synapses. Although I had thought the synapse might emerge as far back as when I put the patent together, this hypothetical mechanism lacked the implementation detail that exists currently; it is anticipated that some of this current detail will have been discarded and replaced before one can confidently say the synapse mechanism exhibits its final form.

The wider deliberation of the mechanism needed to get collections of Agent Devices, being members of a Clique, to communicate the changing state of their Elements amongst each other have brought me to yet other limits of my intellectual capacity as it was when I put the patent together. I feel that the implementation has evolved to a point where, since stabilising the implementation of the properties mechanism discussed in my last post - an act in itself which has pointed me the way to the promise of an answer - I now appear prepared to start replacing at least some of this faith with some reason.

At least at a purely logical level, I think my progression from this point may provide some interesting insight into how a biological nervous system recruits and coordinates its individual neurons in the performance of some act instigated by and for the organism as a whole. Such a mechanism would exhibit hierarchical and executive command and control structures, but would also simultaneously appear to self-assemble based on some form of consensus between neurons (Agent Devices) in accordance with the varying capacity of each neuron to participate. Hence, it would be very hard in any Clique Space (indeed, as it might be demonstrated in any neural system) to point to a single Agent Device as being ultimately responsible for the conduct of any Clique Space as a whole.

Whether synthetic or biological, any such system would definitely require some systematic form of message propagation. It is to this mechanism that my deliberations appear once more to turn. I have some general ideas which I hope to test. For instance, as I explained in a previous post, I see the necessity for Agent Devices to have two Cliques model a single collaboration between a collection of devices, whether these devices are Agent Devices being members of a Clique space, or a collection of devices participating in a collaboration defined by a Media Profile type external to the Clique Space implementation.

The first Clique primarily models what is actually going on in the collaboration; it gives Clique Space a sense of the collaboration's reason for existence, and to the characteristics of its member Participants in accordance to the ability of the Clique's medium to report the collaboration's activity. The second Clique is used to model which Agent Devices have a stake in knowing what is going on in the first Clique by virtue of the fact that any member Participants of this Agent Collaboration's Clique share a combination of the following three relationships with the first Clique: 1. that these Agent Devices are connected to the devices who's collaborative activity is being modelled as participants of the first Clique, or 2. that these Agent Devices are connected to a View or persistence mechanism-enabled devices viewing or persisting activity of the first Clique, and optionally, may be capable of controlling the first Clique, or 3. that these Agent Devices may be acting as relay devices, permitting propagation of Element state from Agent Devices of the first two types located in one physical location to those located in another, but for which the Agent Devices at the first location share no synapses with those at the second; perhaps a bit like a corpus callosum.

This dual-Clique scheme seems a reasonable way to propagate messages amongst relevant Agent Devices, but this mechanism does not include View or persistence devices, because they are not Agent Devices, and hence lack the mechanics necessary to participate in an Agent Collaboration. These type of external devices require another protocol to receive and process Element state information, even if these devices use the same message transport vehicle - the component transmitter, discussed in the previous entry - used in the Agent Collaboration.

It might be interesting at this point to consider how it is a designed intent that the Media Profile Element fits the bill for encoding the protocol for any collaboration type so an Agent Device can receive, process, and control device state. The information content generated by any device may be augmented with Clique Space information in the given device's media in accordance with the Media Profile through which a Connection has been established. The general concept of the Media Profile is used to describe the parameters that a specific protocol will accept. Specifically, the protocols which describe the operational parameters of 1. the Agent Collaboration, and 2. the administrator client, are exposed through Media Profiles like any other protocols. Hence, the Agent Collaboration's protocol is exposed inside the mechanism that implements this protocol by an Agent Collaboration Media Profile so the Agent Collaboration's activity, modelled as a Clique, can be Viewed, persisted, or controlled by an administrator client or any external device which possesses such a capability.

Hence, the Agent Devices have their Agent Collaboration Media Profile. This Media Profile provides the mechanism that realises the second Clique necessary for the dual-Clique propagation scheme. A Media Profile for the administrator client will be fashioned that facilitates the component exchange scheme between an Agent Device and a View or persistence enabled external device. Both the administrator client's component exchange and the Agent Device's dual-Clique schemes will use the common component propagation and exchange mechanism - the component transmitter.

The component transmitter mechanism currently looks like the ideal candidate for the administrator client component exchange scheme as it is for the Agent Device's dual-clique scheme. This is the intention for the design of the component transmitter mechanism; this mechanism replaces the failed Element precis mechanism for the Agent Collaboration, and a promising, but demonstrably inadequate Element projection mechanism for the administrator client. Both the failure of the Element precis and the inadequacy of the Element projection mechanisms prompted me to look to the component transmitter mechanism as a sufficient and flexible alternative candidate for a solution to the question of message propagation.


  1. If you think this blog entry is interesting, then what have you got to lose by getting in contact with its author? You'll find my email address the the "About Me" link from this page. You can post a comment.