notes taken from evolvable computing article - evolvable computing by means of evolvable components - lukas sekanina:
"hardware looks like programs: it is a configuration. hence programs, hardware (configurations) and data have the same nature."
mentions Xilinx Virtex II Pro - 4 PPC procs + FPGA
reconfigurable neural networks - (shu & sutton)
cell matrix architecture (MAcias 1999) - cell matrix is internally configured by exchanging configuration data with its neighbours
"evolvable computing and some modern systems do not share computational scenario of a standard Turing machine and cannot be simulated on Turing machines (my note- interactivity and infinity → consequences for abstraction theories - that it is not a field of equivalence - here time enters the equation)"
"at each time point evolvable devices have a finite description - however when one observes their computation in time they represent infinite sequences of reactive devices computing non-uniformly"
"evolvable computing is beyond scope of an ordinary Turing machine. it does not violate the church-turing thesis because this thesis deals with a slightly different class of computations corresponding to the concept of algorithm. this class of computations is not typical for contemporary computational systems"