Author: Robert Hyatt
Date: 08:52:18 03/30/01
Go up one level in this thread
On March 28, 2001 at 23:52:35, Chessfun wrote: >On March 28, 2001 at 14:12:49, Robert Hyatt wrote: > >>On March 28, 2001 at 00:56:14, James T. Walker wrote: >> >>>On March 27, 2001 at 17:11:48, Ed Schröder wrote: >>> >>>>On March 27, 2001 at 16:21:09, James T. Walker wrote: >>>> >>>>>On March 27, 2001 at 12:11:02, Ed Schröder wrote: >>>>> >>>>>>On March 27, 2001 at 11:12:13, James T. Walker wrote: >>>>>> >>>>>>>On March 27, 2001 at 10:52:43, Ed Schröder wrote: >>>>>>> >>>>>>>>On March 27, 2001 at 10:34:29, James T. Walker wrote: >>>>>>>> >>>>>>>>>As a non-programmer I have to say this is one of the silliest threads I have >>>>>>>>>ever read. It brings up a few questions to programmers. >>>>>>>>> >>>>>>>>>1. Why would you let another program decide for you when/how to save a game or >>>>>>>>>any other info you think is important?? >>>>>>>>>2. Does being desiginated as "Slave" prohibit you from saving information when >>>>>>>>>terminating/starting a game?? >>>>>>>>> >>>>>>>>>I notice when playing two chessbase programs vs each other in auto232 mode they >>>>>>>>>both save the game WITHOUT the box being checked. This seems reasonable to me. >>>>>>>>>Where's the beef? >>>>>>>>>Jim >>>>>>>> >>>>>>>>Being the "slave" you are dependant from the commands send by the "master". >>>>>>>> >>>>>>>>This includes the "save game", "new game" and "force move" commands. >>>>>>>> >>>>>>>>The option "save opponent game" is there because the "save game" may cause >>>>>>>>troubles on the remote program. If it does you turn it off. >>>>>>>> >>>>>>>>Ed >>>>>>> >>>>>>>Hello Ed, >>>>>>>As you know I'm not a programmer and know nothing about it but what do you mean >>>>>>>by "dependant"?? Sure you need to be told "lets start a new game/It's your >>>>>>>move" but seems to me letting your opponent control if you save info or not is >>>>>>>crazy. How can the "master" stop you from saving info before starting a new >>>>>>>game? >>>>>>>Jim >>>>>> >>>>>>"Dependant" is the key-word of the auto232 protocol. The slave in principle >>>>>>has nothing to do and the master is in full control sending commands (mainly >>>>>>moves) to the slave. That's why the expression master-slave is a good chosen >>>>>>one. >>>>>> >>>>>>Of course you can add your own stuff such as controlling "save game" yourself >>>>>>for instance when the slave retrieves the "new game" command from the master >>>>>>you could decide to do a "save game" being the slave. However if the "save >>>>>>opponent game" box is checked on the remote PC the game is saved twice which >>>>>>isn't desirable also. >>>>>> >>>>>>Ed >>>>> >>>>>Hello Ed, >>>>>Well naturally that's what I was thinking. Also if you decide to save the info >>>>>at the "new game" command you could also decide to ignore any "save game" >>>>>commands from the master. I just don't see any sense in putting any more >>>>>control of your program than necessarey into the hands of the opponent. You >>>>>should be able to make your program do anything you want. And I see no reason >>>>>to complain about someone else's program sending or not sending "save game" >>>>>commands. >>>>>Jim >>>> >>>>When was I complaining? >>>> >>>>I was only explaining. >>>> >>>>Ed >>> >>>Hello Ed, >>>No, Not YOu. I was just responding to the thread at the end. You just happened >>>to be the last poster. Vincent was the one starting this thread and he's such a >>>great programmer I can't understand why he would have a problem with auto232 >>>causing his program problems. >>>Best Regards, >>>Jim >> >> >>I have said this before, but I have not said it recently, so here goes: >> >>"auto232 is a piece of trash". >> >>Nothing else to be said. When a protocol has built-in timing dependencies >>that get fried when they are not met, such a protocol is trash. At one point >>Crafty worked perfectly. Then someone got a faster CPU. I had to add a delay >>to not move _too_ quickly else auto232 would miss the move and the game would >>hang. If I probe endgame databases too hard, the interrupts somehow cause >>auto232 to hang. >> >>To have to have a function "Delay()" in your code, and to have to have a >>command "delay N" where N is in milliseconds, is terrible. But when you then >>have to tell users "you have to find N for yourself. Try the default and if >>it hangs, try other values until it doesn't" makes my software engineering >>skin crawl. Think about how many different values there are for up to a one >>second delay. :( >> >>And a user has to experiment to find the right one? And then he upgrades >>something (faster processor, faster disks, more memory, new operating system, >>faster/slower version of the chess engine) and then he has to go Easter-egg >>hunting again trying to find the right delay value? >> >>trash, trash, trash. Can't say it enough. :) > > >Interesting post as I know Jim as well as I have had problems running >Tiger at auto 232. And from reading Vincent's original post on the subject >of Tiger autoplaying seems he don't have any problems at all. > >I always felt the problem was something to do with timing receipt of the >move, as it seems it often hangs when the other program makes an instant >reply. > >Sarah. There are some timing constraints that are not known. If you run on a 1.2ghz machine you might see problems that the same program on a 800mhz machine won't produce. IE faster hardware sends characters/moves faster, etc. If you use a newer/better I/O library it might drive the serial UART better and screw up the auto232 driver with data that comes in too quickly. It really is gross...
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.