Author: Ed Schröder
Date: 00:03:05 03/27/01
Go up one level in this thread
On March 27, 2001 at 02:21:40, Lex Loep wrote: >On March 26, 2001 at 17:10:56, Ed Schröder wrote: > >>On March 26, 2001 at 13:56:06, Vincent Diepeveen wrote: >> >>>Hello People, >>> >>>Why design a protocol for auto232 player? >>> >>>That is the basic question. >>> >>>My interpretation is that this protocol needs to >>>be followed to play games at the auto232 player then. >>> >>>The protocol as designed by Chrilly and Stefan is >>>having a number of commands. >>> >>>The most important is that one is called 'slave' and the >>>other is 'master'. Now being master says shit about whether >>>you play better chess, but it says something about what >>>your function is within the protocol. >>> >>>If you are master, then your function is to start the game >>>and afterwards ship your opponent the command to save the game. >>> >>>Let's first discuss the chesspartner interface which is used for >>>Gambit Tiger. Gambit Tiger is giving very little problems on the >>>auto232 player, let's start mentionning that. It doesn't have >>>big demands to play a game. It's happy very soon. No need to >>>have a machine with zillions of megabytes of RAM, no need to >>>have 7.5 GB of EGTBs on the harddisk before it start playing. >>> >>>It plays no problems there. >>> >>>however, WHY does it have an UNMARKED checkbox by default >>>to let the other guy save the game when Tiger is Master. >>> >>>This means the opponent is by default NOT ALLOWED to save the >>>game. >>> >>>This is very unfair behaviour. >>> >>>It's like playing a grandmaster for the first time, then >>>ship the grandmaster to a clinic. They operate him and he has >>>lost all memories about the game! >>> >>>Of course you can avoid this by difficult programming. So saving >>>the game during the game already. Learning during the game etcetera. >>> >>>BUT WHAT IS THE PURPOSE OF THE PROTOCOL THEN? >>> >>>In my opinion it is UNFAIR to by default leave this checkbox unmarked. >>> >>>In the default settings it must be marked! >> >> >>It is no problem to set the option "marked" as default. I will forward >>the topic to Lex. Maybe he has his reasons, I don't know. >> > >The options you set are remembered from session to session, thus >once you have set the option on, it stays that way as default. > >Lex Yes, that is true. But there are some programs that rely on "learning" during "save game" and so I think it is justified to set "save opponent game" to "on" as default setting which was the original complaint of Vincent. Of course such a change should disturb the quality and stability of auto232 in general. That decision is up to you. Ed >>One plausible reason is that people don't have an interest at all to >>have the games saved twice. Of course you as a programmer want to have >>the games saved in your own format. >> >>But then chess programmers are not in the majority concerning the volume >>of end-users who only care about a wellknown format such as having the >>games in PGN. >> >>So I think you are in the minority here, buy hey the option is there >>and that is the main important thing. >> >>Ed >> >> >> >> >>>The same applies to chessbase products but even worse. >>> >>>First it is very worried about having the right openings book to >>>auto232 play. Secondly it wanted more hashtables and at least a >>>machine with 128mb RAM. Further it wants all EGTBs installed on harddisk. >>> >>>Only after all those criteria are met, then finally fritz wants to >>>auto232 play. >>> >>>The first box you see then it already has by DEFAULT UNMARKED a markbox >>>which will ship a 'save game' command to the opponent after the game. >>> >>>This is pretty unfair! >>> >>>So it wants itself the BEST POSSIBLE conditions, like at least 128mb RAM, >>>a lot of EGTB installed. Hundreds of megabytes of harddisk for a big >>>openingsbook etcetera. All those criteria it wants in order to not even >>>by default give the opponent a 'save game' command after the game, >>>DESPITE THAT THIS IS THE PROTOCOL! >>> >>>Now people can legally complain that their protocol looks like Chrilly/Stefans >>>protocol, but that it is not the same, and that the only differences >>>are that by default chessbase ships some extra commands in order to >>>recognize whether on the other side is also a chessbase program and that >>>the other thing is to by default leave the 'save game' for the opponent >>>is unmarked. >>> >>>all legal crap. JUST GIVE THE SAVE GAME COMMAND by default. >>> >>>That chessbase wants their own main product to win the auto232 matches >>>somehow by shipping commands to other chessbase interfaces to get certain >>>things done, that is completely their own responsibility and decision. >>>Quite logical decision actually. I would want Fritz to win too if it was >>>my main product. I'm not here to speak for how chessbase must run their >>>company. That is their own business. >>> >>>But i'm here for those who want a fair match between non-chessbase products >>>and a chessbase product, as well as chesspartner-tiger, >>>which in future also is going to lose from Fritz as i understood. >>> >>>I understand that programs not learning are greatly influenced by >>>this default unmark trick. >>> >>>You can produce your own PGNs and only those can get interpreted, whereas >>>opponent is NOT allowed to show as slave the pgn, except if that >>>programmer works around this. >>> >>>Much easier as everyone doing a hell of a lot of effort is simply to >>>give everyone that 'save game' command.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.