Authored by: Anonymous on Sunday, May 12 2013 @ 12:05 AM EDT |
Of course a slide rule maintains a state. Have you ever used one? What do you
think that slide with a hairline is for?[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Sunday, May 12 2013 @ 08:25 PM EDT |
Admittedly it has only two registers, the slide index, and the cursor.
But I suppose an n-dimensional slide rule could have a go at AI.
[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Monday, May 13 2013 @ 12:44 AM EDT |
Correct. Which is why what software does on this state
machine is always trivial and never novel.[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Monday, May 13 2013 @ 01:02 AM EDT |
Consult a computer scientist if you need a precise review of this, but
There are only about three equivalence classes for deterministic
computational models...
No stack = DFA (state machine, Deterministic Finite Automata). Can
parse regular expressions but can't remember "nested context" because
it
only supports a finite number of current states. NFA's are equivalent to
DFA's in expressiveness.
1 stack = PDA (Push-Down Automata). Can parse any CFL (context-free
language) but can't do a lot of more general computations.
2 stacks = Turing machine. Can compute any computable function.
Many variations have been explored (using 3 stacks, or random access, or
multiple tape heads, etc.) and none of them have turned out to be capable
of calculations which the Turing machine could not already express.
Its a funny little thing: You've got DFA's at one end of the spectrum, and
Turing machines capable of any general-purpose computation at the other
end. In-between is this distinct class of automata with the limitation to
one stack. What a difference it makes![ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Monday, May 13 2013 @ 12:24 PM EDT |
Folks forget that. [ Reply to This | Parent | # ]
|
|