Joefish wrote:A great centenary, but there's always something bugging me about people building their own 'Turing Machines' that need a microcontroller to operate them. Surely the point is to do it with solid-state electronics or a mechanical system? Requiring a CPU and RAM to demonstrate the fundamental principles of the most basic processing device imaginable seems to me to miss the point by quite a wide mark!
Well, we're waiting ... where is your discrete electronics or mechanical version? A microcontroller uses solid-state electronics, BTW - you probably meant discrete electronics (individual components). Solid-state refers to the fact that the components use crystalline semiconductor materials, instead of inert gas-filled or evacuated components (i.e., vacuum tubes or, as our UK cousins say, "valves").
As for the microcontrollers, if you read closely, there is nothing in Turing's writings about how his machines were to be powered or the logic was to be carried out - just the results (e.g., reading, marking, and erasing symbols on a tape). As an operator, maintainer, and presenter of the only Babbage Difference Engine in the world that's still regularly operated (at the Computer History Museum in Mountain View, California - its older twin brother at the Science Museum in London is rarely operated, nowadays), I would love it if there were hand-cranked completely mechanical Turing machines. Now I know what I want to do during my dottering days of retirement, if the "economy" (or, more precisely, lack thereof these days) ever recovers sufficiently to ensure there will be any porridge left to keep us alive without having to dig ditches, much less star on a "reality" TV show doing all manner of embarrassing things.
Note that the symbols don't have to be ones and zeroes - they can be any kind of symbols, and they don't even have to be letters of an alphabet - they could be geometric shapes, such as triangles, squares, circles, etc. (think APL of the 1960s). The importance of Turing machines isn't the mechanics that drive them, it's the manipulation of symbols that's such a powerful idea, as we can marvel at today (and, IMNSHO, should do much more). Think about the awe-inspiring complex 3-D graphics that the Pi can produce as just another form of symbols, and the true power of Turing's ideas starts to become more apparent.
I'm sure Turing must have imagined large-scale computing in the form of high-performance computing systems with very large numbers of processing elements, but, I'm also certain he would be absolutely flabbergasted at what we can carry around in our pockets at continually-diminishing prices. They're no longer just supercomputers, since they're composed of upwards of millions of miniscule computers that are each immensely computationally powerful in their own right (witness the GPU in the Pi).
We're looking forward to seeing your discrete-component or mechanical Turing machine ... but, we won't be holding our breath