Re: Transactions: good or bad?

"Costin Cozianu" <c_cozianu_at_hotmail.com> wrote in message news:bcss2i$n58su$1_at_ID-152540.news.dfncis.de...
> Bob Badour wrote:
> > "Mikito Harakiri" <mikharakiri_at_ywho.com> wrote in message
> > news:_K4Ia.18$t13.158_at_news.oracle.com...
> >
> >>"Bob Badour" <bbadour_at_golden.net> wrote in message
> >>news:Li5Ia.137$vn2.18732148_at_mantis.golden.net...
> >> > Where is your proof that all chess games halt? After all, your
> >
> > assertion
> >
> >>>that chess games are finite depends on just such a proof.
> >>
> >>I beleve there is a chess rule that if position is repeated, then the
game
> >>is draw. It is quite obvious that the chess game is finite, then.
> >
> >
> > I seem to recall the rule is actually a little more complex than that.
Many
> > games repeat positions without anyone insisting on a draw. There are
other
> > halting rules in chess as well: for instance, there are limits on the
number
> > of moves allowed to achieve checkmate once one player has only a king
before
> > the game is a draw, but a player with only a king has an incentive to
seek a
> > draw (it's the best possible outcome for that player.)
> >
> > I concede that, given sufficient memory of past states, one can always
> > establish an arbitrary rule to halt a finite state machine--including an
> > automated theorem prover--but that just makes a bigger finite state
machine
> > in some class of finite state machines that have a halting proof.

>

Because I am in fact the competent one in this exchange.

> What do you mean a halting proof for a finite state machine ? That's
> complete non-sense. Go back and read on the Theory of Computation.

I apologize for exceeding your ability to comprehend; however, my statements were completely sensible. I would suggest you go read the same text; however, I see no reason to expect you would profit from doing so.

Instead, I will give you a guided tour of the dictionary. Start at "finite state machine" by clicking the following link: http://dictionary.reference.com/search?q=finite%20state%20machine

If you read the definition carefully, you will see the definition points out that FSMs are used in computability theory where "computability theory" is a hyperlink. Clicking on the link will take you to: http://dictionary.reference.com/search?q=computability%20theory

You will see that the definition for "computability theory" mentions the "halting problem" as a hyperlink which will take you to: http://dictionary.reference.com/search?q=halting%20problem

If you really want to pick a nit like the "singe ridicule" you are, you could chastise me for requesting a proof that all chess games halt instead of asking for a "termination analysis" of chess games that results in "definitely terminates." From the definition of "halting problem", you can easily navigate the link to "termination analysis": http://dictionary.reference.com/search?q=termination%20analysis

> Let me explain this to you more gently:

>

That is my understanding as well. As it is my understanding that any given finite state machine is just a particular program, and finite state machines are simply a class of programs.

> This is a general problem: how to write an algorithm that will relaibly
> decide about any particularly given program on any particularly given
> input. Halting theorem: there's no such program.

That is my understanding as well. There are classes of programs for which we can prove termination, but applying any such proof to programs outside the class will at best yield a "don't know" result or at worst cause the proof to spin out of control. See "termination analysis" link above.

> Finite state automata (aka finite state machines): they have no bearing
> wehatsoever and no relation with the halting problem, chess programs and
> what we discuss here.

Really? Then you are claiming that a chess game between two computers is something other than a finite state machine. You must be using some idiosychratic definition of the term because such a chess game is a finite state machine.

> They have a finite set of states, and they perform
> transition from one state to another.

You mean like a chess game has a finite set of states and a series of moves that transition it from one state to another? Or like a real, physical computer has a finite set of states and transitions from one state to another?

> They are given a finite input
> string

Why finite? What, other than time, prevents a finite state machine from receiving an infinite input string? It's the finite number of states that give it the name.

>, they perform the transitions corresponding to the letters

By letters, do you mean symbols like some symbolic representation of chess moves? And do such strings include an empty string over an empty set of letters?

> and if
> they land on a terminal state when they finnished the input that means
> they "recognized" the string as part of the language they define. They
> always halt

I assure you I can construct a finite state machine that does not halt and does not require a non-empty imput string. A simple two-state oscillator will do.

>, and it makes no sense to debate the halting problem for

I am not confused about anything. I know exactly what I am doing: I am wasting my time on a jerk who simply refuses to acknowledge he was wrong when he responded to Alfredo's purely factual observation with philosophical mumbo-jumbo and superstitious claptrap.

On occasion you say some intelligent things, so I am hesitant to add you to my twit-filter.

> A chess program has to decide what is the next best move and play it.

A chess program is not a chess game. As a move decider, one must apply a chess program repeatedly to make even half a game. A turing machine only has to decide whether to move to the left or to the right, but this simpler decision tree fails to escape the clutches of the halting problem either. The halting problem limits the human brain as much as it limits any other machine.

You have yet to even comprehend what is required of you to support your earlier assertions. Your assertion that all chess games are finite requires that all chess games halt--not just that some algorithm to choose each next move halts. Where is your proof that all chess games have a finite number of moves?

> In
> order to do that all chess programs perform a variation of the min max
> aklgorithms. My valuation of the position after my move is the minimum
> of the opponent's valuations on all his possible move, and I get to9
> choose the move that maximizes my valuation. That's min. - max. for you.

I understand what min-max is you ridiculous ape. Do you understand what a chess game is, yet?

> As you might see if you're not busy trolling the valuation function is a
> recursively defined function over the space of all positions, with some
> well defined positions where it's valued win or drawn by rules of chess
> (mate, stalemate, insufficient material, etc). Also when you meet a
> previously analyzed position that hasn't been assigned a value yety, you
> value the branch as a draw for the purpose of min. or max. step.

Nothing prevents two chess programs from taking each other down a path that loops back to some position on the path. Where is your proof that all chess games halt?

> Since the domain of the min. max. algorithm is finite, the algorithm
> cannot suffer from the halting problem unless it is written by an
> incompetent or a troll like you.

A turing machine evaluates a much simpler function at each move, and it fails to elude the grip of the halting problem. Do you understand what a chess game is, yet? Where is your proof that all chess games halt? Your earlier assertions depend on just such a proof.

> That's crass ignorance

When it comes to the crass, I prefer the ignorant to intellectual cowards and to those deluded by their own intellectual dishonesty. Ignorance, at least, has a cure. Received on Thu Jun 19 2003 - 22:12:37 CEST