Re: Transactions: good or bad?

>>Bob's claim however was a straw-man even in the absence of those rules,
>>because a chess playing game can *prove* if a position is techincally
>>winnable by white, or by black, or it is draw (therefore effectively
>>calculating a proof in the formal system established by the rule of
>>chess), simply by exhaustively applying the min.-max. recursive
>>algorithm over the *finitely many* positions of chess.

That's complete BS. A game is not a computer. It doesn't serve the same purpose in Game Theory as computers do in Theory of Computations.

You proved more than enough that you don't know about either.

>>Given that they search in a game with an infinite space, programs have
>>to decide whether evaluating a particular branch will be successful,
>>without actually  going on the branch (because branches are not
>>guaranteed to be finite). It is there where they hit the halting problem.

Complete BS. Humans are not TM.

Humans may just pick a branch that they consider, and more likely than not they will be successful.

No fucking human on the planet ever lost any sleep over fearing whether his mind will gom in an infiinite loop. Humans are obviously oblivious to such considerations.

>>Evaluation branches in chess are always finite, even if games may be
>>infinite,

Clueless again.

>

>>because the minute you encounter a previous position you no
>>longer need to explore the branch.

That's complete BS. A chess playing program does not have the mission to play an infinite game. It has the mission to play the next best move in each and every position. That's elementary.

In order to *choose the next best move* (repeat after me: *the next best move*), it doesn't have to evaluate infinite games, therefore it doesn't suffer the halting problem (and that's all I claimed: chess playing program don't suffer from the halting problem).

In case you still don't get it you can always think of a program that will ask for draw by repetition, but this is a user friendly convenience, it has no bearing whatsoever on theoretical aspects of chess playing.

In no way is a fact that they can decide the next best move a sign that they are intelligent. Claiming that is the same thing as claiming that a computer is intelligent because it can perform billions of addition per second.

>>In practice it has been proven that performing an exhaustive search over
>>the space bounded by a certain depth of analysis (let's say 20 moves,
>>well beyond the reach of a human player) is statistically more than
>>enough to approximate the perfect evaluation function in quite many
>>position. Still human players get to beat chess programs simply by their
>>better intuition.

You don't know what I have inside my "machine". If you have the leats of mathematical background, you would know that you cannot speculate about objects that you cannot properly define.

Otherwise, here's this mathematical object for you: let there be a set of people (a village) and a barber in the village, and the barber shaves all the people who don't shave themselves.

Now you can have a lot of non-sense about the barber, and you might ask who shaves the barber.
>
>

>>The same way, mathematicians rely on intuition to know where to look for
>>a proof, just like a 6 year old gets to recognize hand written
>>character "automagically" (including taking the context of the phrase
>>into account, the language) while there's no scientific theory to allow
>>us to conclude that computers can do the same.

No magic, pure facts. Mathematician solve problems. The 6 year old recognizes characters. Computers don't.

>
>

>>Yet, there's this philosophical belief of some people, that we can
>>encode this "intutition" factor as a computable function. That's insofar
>>completely unsupported.

"Intutition" is whatever makes people play good chess, kids recognize characters, mathematicians recognize a path to follow towards a mathematical results. It exists because it has effects, we don't know what that is, but we know it's not a computer.

My claim is that we have no clue whether this is or is not a computable function, therefore reproduceable by a computer. All the evidence we have suggests that it is not.

Your claim is that you know it is a computable function, and although you and Alfredo have no real clue about theory of computation, fundamentals of mathematics or proof theory you took the liberty to call names on one of the best expert in these fields whose work is cited by virtually all the literature on the subject.

You must be kidding yourselves, if you have the illusion that by googling for random content while exhibiting so much basic incompetence on the subject you've got any clue, or in your case any claim to "intellectual honesty" for that mater. You're just an amateur troll. Received on Thu Jun 19 2003 - 21:34:16 CEST