Ron Hudson wrote:
...
If the desert island has vines you could make a rope
computer, as
described in a
Scientific American, I don't remember the issue.
It was an april fool's issue, I believe. If memory serves, the island
was called APRLFUL or something like that. The article showed how, with
clever arrangements pullies, toggles, springs and such, inverters and
simple logic gates could be produced. Entertaining enough.
The problem with that "computer" is that the logic gates have no gain.
Each gate has no power supply other than the mechanical power of the
input signals (ropes getting pulled). For instance, say you pull an
input rope of a NAND gate one foot (OK, let's keep it metric) ten inches
:-) and the output rope moves nine inches, so the gain is 0.9. Put N of
these gates in series and the output signal is (0.9**N) of the input
signal. After half a dozen gates, very little is left at the output.
The rope gates had a gain much less than 0.9 -- probably 0.5.
It is the same reason why computers aren't built of of just diodes and
resistors. You can have a gate or two in series, but then the signal
needs to be reconditioned before being applied to the next stage.