I've got something in my head. I've been working on it in there since last Monday, and it'll have to remain in there until I file and serve my submissions in an application IBM's legal counsel have made in my case to have it struck out. Too bad there perhaps. Perhaps not. Perhaps I'll know on 19 April.
Back to my head and to the things therein.
I'm looking at what I perceive to be a solution that revisits deliberations around the same subject matter given in this entry in such a way that draws on the inherent abstract nature of the core data model. What is in my head is promising a lot.
In order to understand what is in my head, one needs to understand how every piece of information an Agent Device knows of (everything except Cliques) must be expressible as a component (well, more correctly, a transmissible component) before it can be propagated to other Agent Devices or projected to other V/PM-enabled devices. All components except the Sovereign's Clique Space are transmissible, and all transmissible components except the "observer" Participant are observable. Identifiers are not components, but do have a method named asQuantum that accepts no parameters.
The asQuantum method is also declared in the transmissible component interface so these type of components can be represented as a serialisable quantum which can be propagated or projected or persisted. All Elements implement the transmissible component interface, and delegate to the enclosed identifier's asQuantum method in the corresponding Element's asQuantum method.
The lovely observation my head has captured about the third category of component (the observable transmissible component) described above is that components of this type are indeed "observable". That is, these type of components have an observer - an "observer" Clique composed entirely of "observer" Participants. This Clique registers all devices (whether Agent Devices or any V/PM device - collectively known as "observer" devices) that are interested in a particular "observable transmissible" component.
I can't wait to start work on this. I reckon this mechanism is the final key to a demonstrable prototype. The thing about this mechanism is that it (or the something that truly has to be implemented) has been in my head since mid-2004. It is only now, however, that the implementation of everything else had to be done before I was sufficiently prepared to investigate this observer mechanism.
Maybe I'm just a dickhead. Eureka!