Taking the karma hit to clarify why I, at least, downvoted this comment. When you say the Turing model is outdated, you seem to be assuming that the model was originally intended as a physical model of how actual computers do (or should) work. But that was never its purpose. It was supposed to be a mathematical model that captures the intuitive notion of an effective procedure. All the talk of tapes and tape heads is just meant to aid understanding, and maybe that part is outdated, but the actual definition of a Turing machine itself can be given purely mathematically, without any assumptions about physical instantiation. Saying the Turing model is outdated would only make sense if there were good reason to doubt the Church-Turing thesis, and there isn’t.
I seriously disagree, incidentally. For example, it has some pretty specific assumptions about instantiation—it will be the sort of computer they had seventy some odd years ago. Because the single pool of memory, the single processor, the scan-one mechanism of traversal, all of these are assumptions which have serious effects on the very field the mathematical model was devised to consider, computability.
(And I can point out one good reason to doubt the Church-Turing thesis. A Turing machine is incapable of replicating the nondeterministic behavior of concurrent processing. I’ve never yet seen a process which depended on this nondeterminism, but it is nonetheless possible to devise one. Maybe to generate random numbers, I’m not sure.)
Taking the karma hit to clarify why I, at least, downvoted this comment. When you say the Turing model is outdated, you seem to be assuming that the model was originally intended as a physical model of how actual computers do (or should) work. But that was never its purpose. It was supposed to be a mathematical model that captures the intuitive notion of an effective procedure. All the talk of tapes and tape heads is just meant to aid understanding, and maybe that part is outdated, but the actual definition of a Turing machine itself can be given purely mathematically, without any assumptions about physical instantiation. Saying the Turing model is outdated would only make sense if there were good reason to doubt the Church-Turing thesis, and there isn’t.
Thank you.
I seriously disagree, incidentally. For example, it has some pretty specific assumptions about instantiation—it will be the sort of computer they had seventy some odd years ago. Because the single pool of memory, the single processor, the scan-one mechanism of traversal, all of these are assumptions which have serious effects on the very field the mathematical model was devised to consider, computability.
(And I can point out one good reason to doubt the Church-Turing thesis. A Turing machine is incapable of replicating the nondeterministic behavior of concurrent processing. I’ve never yet seen a process which depended on this nondeterminism, but it is nonetheless possible to devise one. Maybe to generate random numbers, I’m not sure.)