auto232 question (again)
Robert Hyatt
I've been getting more and more excited about the possibility that the auto232
issue is solved. Crafty's been playing lots of successful games, with tablebases,
without, with opening book, without, etc... And all of a sudden, a "hole" shows
up. Crafty is failing when it gets the "SP" command. The reason it fails is that
I have not implemented an SP command. :)
What is this supposed to do?
Bob
mclane
hy...@crafty.cis.uab.edu (Robert Hyatt) wrote:
>Bob
Hm. Maybe sp is german for SPIEL = play/go/move ?
Why are you having these problems in implementing the
autoplayer-device, meanwhile all others have implemented it ?
Maybe you should ask somebody who knows, maybe John Stanback, or
Vincent or Joerg Burwitz, the Diogenes programmer... :-)
Remi Coulom
Robert Hyatt wrote:
>
> I've been getting more and more excited about the possibility that the auto232
> issue is solved. Crafty's been playing lots of successful games, with tablebases,
> without, with opening book, without, etc... And all of a sudden, a "hole" shows
> up. Crafty is failing when it gets the "SP" command. The reason it fails is that
> I have not implemented an SP command. :)
>
> What is this supposed to do?
>
> Bob
Here is all the english documentation I have about auto232. I used it to write my
interface for The Crazy Bishop. This comes from Moritz Berger, and was sent to me
by one of TCB's users.
| 2a) Moves: From-Field To-Field Return e.g. "e2e4<Carriage Return>"
| 2b) New Game: "AN<Carriage Return>"
| 2c) Turn Board: "BD<Carriage Return>"
| 2d) Take Back: "ZU<Carriage Return>"
| 2e) Make Move/Stop computation: "BLANK"
| 2f) Start Computation: "SP<Carriage Return>"
| 2g) Memo-Mode on/off "MZ<Carriage Return>"
| 2h) Printer on/off "DR<Carriage Return>"
| 2i) Store-Game "SR<Carriage Return>"
| "Filename<Carriage Return>"
|
| This is my code to output moves for both white and black.
|
| if (player == 'w') strcpy(txt1,""); else strcpy(txt1,"\t");
| sprintf(txt2, " %c%c-%c%c", s[0],s[1],s[2],s[3]);
| if (s[4] == '=') sprintf(txt3,"/%c",s[5]); else strcpy(txt3,"");
| sprintf(txt,"%s%s%s\n",txt1,txt2,txt3);
| fprintf(stdprn,txt);
| fflush(stdprn);
|
| Where the string "s[]" contains the long algebraic move in the format
| "e7e8=Q". This code converts it to "e7e8/Q" <CR> and spits it to the
| printer. I don't know if the tab is needed before black's move.
|
| I poll for input every 50 mS. If I'm pondering and a move comes in, I
| break pondering, make the move and echo it back to the printer in less
| than a millisecond I would guess.
| I think it would wait 50 mS after receiving a move to send it's move to
| the printer.
|
I hope this helps you. If anyone has a more formal definition of the protocol in
english, I would welcome it.
You could also take a look at the source code I have released on TCB's page, but I
doubt it could be of any help. I do not even know if it works, since I have had no
report yet from the user that requested that I add an auto232 interface.
Remi
-------------------------------------------------------------------------------------
I need more new text that quoted to post, so I add this :
I hope this helps you. If anyone has a more formal definition of the protocol in
english, I would welcome it.
You could also take a look at the source code I have released on TCB's page, but I
doubt it could be of any help. I do not even know if it works, since I have had no
report yet from the user that requested that I add an auto232 interface.
Remi
I hope this helps you. If anyone has a more formal definition of the protocol in
english, I would welcome it.
You could also take a look at the source code I have released on TCB's page, but I
doubt it could be of any help. I do not even know if it works, since I have had no
report yet from the user that requested that I add an auto232 interface.
Remi
I hope this helps you. If anyone has a more formal definition of the protocol in
english, I would welcome it.
You could also take a look at the source code I have released on TCB's page, but I
doubt it could be of any help. I do not even know if it works, since I have had no
report yet from the user that requested that I add an auto232 interface.
Remi
I hope this helps you. If anyone has a more formal definition of the protocol in
english, I would welcome it.
You could also take a look at the source code I have released on TCB's page, but I
doubt it could be of any help. I do not even know if it works, since I have had no
report yet from the user that requested that I add an auto232 interface.
RemiI hope this helps you. If anyone has a more formal definition of the protocol in
english, I would welcome it.
You could also take a look at the source code I have released on TCB's page, but I
doubt it could be of any help. I do not even know if it works, since I have had no
report yet from the user that requested that I add an auto232 interface.
Remi
Robert Hyatt
mclane (mcl...@prima.ruhr.de) wrote:
: hy...@crafty.cis.uab.edu (Robert Hyatt) wrote:
: >I've been getting more and more excited about the possibility that the auto232
: >issue is solved. Crafty's been playing lots of successful games, with tablebases,
: >without, with opening book, without, etc... And all of a sudden, a "hole" shows
: >up. Crafty is failing when it gets the "SP" command. The reason it fails is that
: >I have not implemented an SP command. :)
: >What is this supposed to do?
: >Bob
: Hm. Maybe sp is german for SPIEL = play/go/move ?
: Why are you having these problems in implementing the
: autoplayer-device, meanwhile all others have implemented it ?
: Maybe you should ask somebody who knows, maybe John Stanback, or
: Vincent or Joerg Burwitz, the Diogenes programmer... :-)
I'm having interesting difficulties because there is little (if any
at all) written documentation on the timing requirements. I'm not sure
that there is *any* program that works perfectly with this system. All
of the feedback I'm getting point to difficulties in other programs as
well, where the testers have to frick around with different options on
the auto232 driver to make it work for particular processors. I have been
sent games where most programs have hung at some point or another, where
another program refuses to start a new game, and so forth. So it is not
*only* Crafty that't the problem...
In fact, the problems are not in the chess engines at all... you have
to look at the common denominator to see what's wrong. Don't forget, Crafty
and a couple of other programs work *perfectly* with xboard to play automatically
on chess servers. This is a harder interface problem than auto232, because
there is more functionality.. and a lot fewer quirks.
Crafty will never work perfectly until auto232 is corrected in some places,
for example, only allowing promotion to queen... that will cost a game here
and there for no reason...
The best bet would be a new protocol that eliminates *all* of the oddball
timing inconsistancies that exist in auto232. This business of responding
not too fast and not too slow is rediculous with the wide range of machines
we have today...
If the commercial programmers would make a linux version available, I'll
personally write an interface for linux that will solve all of this nonsense
once and for all. However, DOS is a doomed platform and it's not worth
fooling with there... and I don't do DOS anyway... :) I do do Unix of
course. :) :)
Robert Hyatt
Remi Coulom (Remi....@ensimag.imag.fr) wrote:
: Robert Hyatt wrote:
: >
: > I've been getting more and more excited about the possibility that the auto232
: > issue is solved. Crafty's been playing lots of successful games, with tablebases,
: > without, with opening book, without, etc... And all of a sudden, a "hole" shows
: > up. Crafty is failing when it gets the "SP" command. The reason it fails is that
: > I have not implemented an SP command. :)
: >
: > What is this supposed to do?
: >
: > Bob
: Here is all the english documentation I have about auto232. I used it to write my
: Remi
: Remi
another screwup on my part. the SP command was implemented. It is handled in main.c
at the start of a game. Crafty is set to not "ponder" on the first move so that it can
read commands like "GO" (SP) and handle them, but the "new" command does not clear this
flag again so on the second game, it starts pondering immediately, and falls on it's
own sword because it can't handle "go" or "SP"...
on my list to fix...
note this will break crafty using auto232 when it plays white. When playing black, it
will work perfectly... so it seems...
Bob
Gambit-Soft
Robert Hyatt wrote:
> once and for all. However, DOS is a doomed platform and it's not worth
> fooling with there... and I don't do DOS anyway... :) I do do Unix of
> course. :) :)
Hello, Chrilly Donninger, the programmer of AUTO232 seems to have an
email address now. Maybe Andreas Mader knows this address.
This is the first step. The second would be that Donninger
answers the email he receives.
I am told that he works in a private office every day, where
he has no phone and no electronic gadgets besides his NIMZO computer
and maybe a coffee machine.
But who knows, maybe times are changing.
His opinion about Deep Blue, by the way, is:
I don't know where I want to go, but I want to go there fast!
Bert
mclane
gambi...@t-online.de (Gambit-Soft) wrote:
>Hello, Chrilly Donninger, the programmer of AUTO232 seems to have an
>email address now. Maybe Andreas Mader knows this address.
Unbelievable. An austrian with electricity !!
Who has told them how to handle fire ?
>This is the first step. The second would be that Donninger
>answers the email he receives.
>I am told that he works in a private office every day, where
>he has no phone and no electronic gadgets besides his NIMZO computer
>and maybe a coffee machine.
No phone - no electronics, only NIMZO and coffee.
As I said....
Coffee machine ???
The last machine that was invented in austria was the turk, and in
this "chess-machine" there was no machine, but a small human beeing.
So machine is a little overestimated.
>But who knows, maybe times are changing.
>His opinion about Deep Blue, by the way, is:
>I don't know where I want to go, but I want to go there fast!
The only thing Deep Blue can find out very fast is, how much money
they have to spend on the electricity-bill after the event.
I hope this time the operator is able to input the moves correlating
to the chess-rules. BTW: do these people play chess ??? Or do they
know somebody who knows somebody who has heard of how to play ?
>Bert
Robert Hyatt
mclane (mcl...@prima.ruhr.de) wrote:
: gambi...@t-online.de (Gambit-Soft) wrote:
I suspect GM Joel Benjamin qualifies???
mclane
hy...@crafty.cis.uab.edu (Robert Hyatt) wrote:
>I suspect GM Joel Benjamin qualifies???
Depends on what he is doing and if he is understanded.
If he is operating the coffee-machine the whole day, than I guess this
will be not enough....
Vincent Diepeveen
Well they hired Kasparov for that part of the game.. :)
Bye the way is it true that Nh3?? in game 3 is so terrible bad?
I was looking to the game, with Diep analysing on the background,
and saw a real beginners move: Nh3??
Who in the world wants to put his knight on that field volunteerly
clearly getting a terrible pin over f3 by the bishop.
Why did Kasparov play that move, did i miss something?
It is in my eyes a real beginners move would not even have thought
about it.
It seems Kasparov is just doing something, and waits until he is having
a difficult position, then wakes up, and plays some grandmaster moves.
What a terrible openingsplay by Kasparov. Especially that first game.
Of course a computer cannot handle closed positions. All aegon tournaments
already told us that.
We know that already for many years. Most Grandmasters are however walking
openingsbooks (they have 24 hours a day to study it), and just look how
Kosashvili won with 6 out of 6 the Aegon tournament.
He did it by playing like he does normaly, and simply playing good moves
after opening.
The result: immediately after having a good position in opening (derived
from the better openingstheorem he knows), he got
an even better position, as the computer programs played here or there
worse moves, which was enough to win the games.
This is the way in which i would expect Kasparov to win from DB also!
That would be a great way to win from DB, where it seems that
now Kasparov is just playing moves, and sees what happens.
Vincent Diepeveen
--
+----------------------------------------------------+
| Vincent Diepeveen email: vdie...@cs.ruu.nl |
| http://www.students.cs.ruu.nl/~vdiepeve/ |
+----------------------------------------------------+
Komputer Korner
Vincent Diepeveen wrote:
snipped
> The result: immediately after having a good position in opening (derived
> from the better openingstheorem he knows), he got
> an even better position, as the computer programs played here or there
> worse moves, which was enough to win the games.
>
> This is the way in which i would expect Kasparov to win from DB also!
> That would be a great way to win from DB, where it seems that
> now Kasparov is just playing moves, and sees what happens.
>
> Vincent Diepeveen
>
> --
> +----------------------------------------------------+
> | Vincent Diepeveen email: vdie...@cs.ruu.nl |
> | http://www.students.cs.ruu.nl/~vdiepeve/ |
> +----------------------------------------------------+
You are right Vincent, I have been saying that from the beginning.
I think that Kasparov should play right into Deep Blue's strengths.
Out calculate it. The difference of course is that Kasparov will be
doing it from a better position. Game 1 of the last match is not
a good example because Kasparov tried it from a worse position. All
Kasparov has to do is play normal solid chess and wait for an opening.
As it is right now, he is playing abnormal chess and waiting for an
opening. Sometimes that works and sometimes that doesn't whereas
by playing normal solid chess it will work all the time. Gary believed
the hype from the Deep Blue team and is playing into their hands.
--
Komputer Korner
The inkompetent komputer.