New programming paradigms are necessary to deal with the complexity of semi- or fully-autonomous real-world systems. In this section we evaluate the consequences of the PACE approach involving the omega machine for novel IT technology. We shall attempt to address the following questions:
•What does the omega machine architecture entail for IT?
•How does it differ from conventional (distributed) control theory?
•What is the potential impact for programming self-organizing systems?
•What is complementation and how can we use this to program systems?
•What is the potential impact of this approach on ICT?
The conventional approach to dealing with information processing in real world physical systems (such as robotics or embedded systems) involves model development, simulation, rational sequential design of behavior, implementation, testing, debugging, reinterpretation and iteration of this cycle. In the omega machine approach to designing complex systems, we employ microscopic feedback loops involving (ideally "universal") sensory and actuator systems, respectively capable of monitoring and influencing a broad spectrum of system behavior. As with the eye or a video camera, parallel spatial receptor fields provide relatively generalsensoryinformation about a material system: particularly if each local receptor is spectrally-resolved (color-sensitive, e.g. fluorescence, Raman, IR, NMR imaging). General spatial actuation systems are less familiar, but the skin-covering camouflage of camelions provides an example. In PACE, we have used a fine-grained 2D network of electrodes in a similar spatial manner. Microelectrodes provide a complex but rather general mode of localactuationof chemical systems: with effects ranging from electrophoresis and electroosmosis to electrochemically induced reactions, local heating and electrowetting depending on the local potentials, electrode arrangements, channel geometry and other conditions.
The local feedback system produces an electronically (or more precisely electro-optically) extended physico-chemical system, in which part of the spatio-dynamical behavior resides in the properties of the electronic feedback system. This is akin to chemical reactor control theory, but here we have a myriad of variant fine-grained control loops each at the same spatiotemporal scale as the artificial cells that we wish to see emerge. This qualitative difference allows evolutionary optimization to be performed in the extended space of hybrid chemical-electronic couples. It is this that enables us to speak of an electronic genome for the local chemistry: the properties of successful sensory-actuator-feedback control loops can be inherited along with the organizational forms and material compositions of the chemical systems.