I've lost my head of steam on development for the moment. This loss of momentum has been due to considerations which have appeared in respect of implementing the final core Clique Space's Media Profile: the aptly named Clique Space Media Profile. This was previously known as the collaborator Media Profile, but that was when I though implementation of it and another Media Profile would be necessary. However, this necessity appeared to have evaporated, and shortly after that realisation, I decided that the name I was reserving for this other (Clique Space) Media Profile was more appropriate for the collaborator Media Profile; hence, I dropped the name "collaborator" for the collaborator Media Profile in favour of "Clique Space".
So, to the mechanics of an Agent Collaboration I now find necessary to entertain.
Each Agent Device uses its Clique Space Media Profile to represent itself as a Participant in the Clique Space Clique - the Clique that models the the Agent Collaboration in which each Participant represents each Agent Device's membership of a given Clique Space. Pretty simple: the Clique Space Clique is a model of a given Agent Collaboration much in the way any other Clique is a model of any other media collaboration - reflected in another Agent Collaboration which represents all the Agent Device's which possess a connection to the devices which have Participants in the other media Clique. So, in the specific case of the Clique Space Agent Collaboration, the medium being modelled is Clique Space.
That's all great, and it is exactly what I envisaged in July (or August, or perhaps September) 2004. However, now that I have arrived at this stage of implementation, I have to do a bit of thinking about the mechanics of the Agent Collaboration, and how the Clique Space Media Profile's 1: Enabling and 2: Limiting Constraints will map to 1: Agent Device functionality and 2: functional limitations imposed by the administrator(s) of a particular Clique Space.
Putting the media mapping to one side for now, I have to consider how the Agent Devices are actually going to collaborate. Each Agent Device is merely a node in a neural network, and messages are merely propagated across an Agent Collaboration to all Agent Devices which are members of a Clique. In the case of the Clique Space Clique, the given Agent Collaboration represents all Agent Devices that are members of the given Clique Space.
Another side point that is very interesting about this concept. Clique Spaces can be "federated" in a way by allowing a particular Clique to contain Participants from different Clique Spaces. Coupled with the flexibility of the Media Profile hierarchy and the flexible way Enabling Constraints can be expressed, and in the way Limiting Constraints (properties of any Element, but especially Account Profiles and their associated hierarchy) can be expressed in a Participant, this provides a level of administrative control that I believe is capable of "infinite phenomenological flexibility" (uh... okay...) in that any one individual can have full control of who else can see an interact with any collection of one or more devices one may possess within the phenomenon of what it is to be an individual.
Anyway, back to the technical details of the Agent Collaboration. Messages (I've settled on the term "pulses") shall propagate through the Clique Space's Agent Collaboration (the Clique Space's medium) like ripples through a pond. Any Agent Device with sufficient Limiting Constraint affinity can create a pulse. There are a whole slew of details that need to be figured out: what if an Agent Device goes off-line? What happens when an Agent Collaboration representing another medium contains Participants from multiple Clique Spaces? How are decisions made in situations when one pulse contradicts another?
All these considerations may be to a greater or lesser degree, different to the media being modelled. However, the Clique Space's Agent Collaboration - the Clique Space's medium (modelled by the Clique Space's Media Profile) being the medium that regulates all other media which Clique Space will model, will set an ultimate baseline to the ability of a Clique Space's implementation to realise this idea of infinite phenomenological flexibility. It is the Agent Collaboration's pulse that remains the largest unknown to this reality; and after working out how Agent Devices engage each other to create synapses, and how administrator Client Devices connect to and receive projections of various Elements so they can be viewed and persisted as necessary by the individual users of Clique Space, the pulse message is now the subject of my full attention.
Another six months perhaps...
Another six months... heh... I should avoid making such statements - they never meet expectations.
ReplyDelete