I like the idea of Hypervisors. They provide a framework for virtual operating systems like XP and MacOS running side by side with a kind of dumb ignorance of each other's existence. An example is the recent marriage touted by Apple users who wistfully viewed the enormous amount of XP software. Constructs like shared objects exist that allow communication between the hypervisor and the virtual machines it creates, such as shared disk files or client-server connections.
Presumably, if a fully OS independent manner of Hypervision is implemented, then all kinds of operating systems could run concurrently, even new ones that haven't been thought of before. This would be a good place for Artificial Intelligence because there can be a kind of "God" Operating System that imposes all the rules and quirks upon the lowly "Beasts" (XP and MacOS) who are God's creations. Many years ago, there were VMs or virtual machines, which multiplexed machine code translations so that say, an IBM360 and a Honeywell Mainframe might be emulated in the same room full of computer equipment. Given a flexible enough VM base language (the microcode), any other machine could be implemented as a higher level instruction set.
Even without VM, there could be just a virtual partition within a native architecture, so that any Intel CPU based operating system could run concurrent with another. At any rate the possibilities are interesting -- if not somehow very expensive in either hardware costs or performance hits.
XP, MacOS, Linux could all run with OP (a whimsical Optical Processor) all at once, to perform many diverse intercommunications and specialized functions that are best done in their separate worlds. The results can stay parallel and only observed at a higher level by humans, or the overall effect could be for each node's strengths and weaknesses supplemented by the others.
Optical Processing (only a theoretical system) can be used recursively as a kind of dream-state in which slightly re-arranged stimulus-response images immediately become seeds of possible "ideas". Trees and arrays of hypervisor objects could perform as temporary trials, flushed when they get hung, or remembered more strongly whenever they stumble upon suitable, even trivial, solutions to complex problems.
Arrays of nodes are easy to construct with data passed though IN and OUT pipes, with the INs being inbranched through guard processes and the OUTs being outbranched by replicator processes. The channels to the pipes can be interspersed with hypervisor "thoughts", or little modifications to the data to give a more global presence to otherwise private interactions between logical nodes, presuming that the hypervisor has access to these pathways. In the human brain, separate neural nodes can be effected by chemical diffusion, i.e. adrenaline, to have such global effects on otherwise private channels.
Evolution Software (which can alter data in little ways and pass on good things and kill bad things) can then optimize the massive amounts of parametric flux.
The beauty of the system would be the manner in which viruses or otherwise malicious code be found as the side effects of failure in the hypervised nodes. Diseased nodes are just killed off, and presumably adjacent nodes would be spared. Viruses might still exist, but their necessary adaptation to heavy kill-offs would make them rare.
Certainly in this manner, at least to a level I'm not sure is all that comfortably shallow, it seems that agents of our own consciousness (and unconsciousness) could successfully be mapped into nodes of a giant hypervisor system. But then, whose mind would be the model? The Good Brain or the A.B. Normal brain.
Presumably, if a fully OS independent manner of Hypervision is implemented, then all kinds of operating systems could run concurrently, even new ones that haven't been thought of before. This would be a good place for Artificial Intelligence because there can be a kind of "God" Operating System that imposes all the rules and quirks upon the lowly "Beasts" (XP and MacOS) who are God's creations. Many years ago, there were VMs or virtual machines, which multiplexed machine code translations so that say, an IBM360 and a Honeywell Mainframe might be emulated in the same room full of computer equipment. Given a flexible enough VM base language (the microcode), any other machine could be implemented as a higher level instruction set.
Even without VM, there could be just a virtual partition within a native architecture, so that any Intel CPU based operating system could run concurrent with another. At any rate the possibilities are interesting -- if not somehow very expensive in either hardware costs or performance hits.
XP, MacOS, Linux could all run with OP (a whimsical Optical Processor) all at once, to perform many diverse intercommunications and specialized functions that are best done in their separate worlds. The results can stay parallel and only observed at a higher level by humans, or the overall effect could be for each node's strengths and weaknesses supplemented by the others.
Optical Processing (only a theoretical system) can be used recursively as a kind of dream-state in which slightly re-arranged stimulus-response images immediately become seeds of possible "ideas". Trees and arrays of hypervisor objects could perform as temporary trials, flushed when they get hung, or remembered more strongly whenever they stumble upon suitable, even trivial, solutions to complex problems.
Arrays of nodes are easy to construct with data passed though IN and OUT pipes, with the INs being inbranched through guard processes and the OUTs being outbranched by replicator processes. The channels to the pipes can be interspersed with hypervisor "thoughts", or little modifications to the data to give a more global presence to otherwise private interactions between logical nodes, presuming that the hypervisor has access to these pathways. In the human brain, separate neural nodes can be effected by chemical diffusion, i.e. adrenaline, to have such global effects on otherwise private channels.
Evolution Software (which can alter data in little ways and pass on good things and kill bad things) can then optimize the massive amounts of parametric flux.
The beauty of the system would be the manner in which viruses or otherwise malicious code be found as the side effects of failure in the hypervised nodes. Diseased nodes are just killed off, and presumably adjacent nodes would be spared. Viruses might still exist, but their necessary adaptation to heavy kill-offs would make them rare.
Certainly in this manner, at least to a level I'm not sure is all that comfortably shallow, it seems that agents of our own consciousness (and unconsciousness) could successfully be mapped into nodes of a giant hypervisor system. But then, whose mind would be the model? The Good Brain or the A.B. Normal brain.
No comments:
Post a Comment