Author: Robert Hyatt
Date: 10:19:05 11/07/01
Go up one level in this thread
On November 07, 2001 at 12:59:01, Miguel A. Ballicora wrote: >On November 07, 2001 at 12:15:11, Robert Hyatt wrote: > > >>The best solution is a robotic interface. Everybody can still show up and >>watch, but once the game starts, the humans are _totally_ removed from the >>procedure. Moves are relayed automatically thru a third party (like a local >>chess server, for example, since now all programs are able to play on ICC). >>No human hands in the mix, no outside intervention, no errors, no move >>takebacks, no arguments about "hey, you made the wrong move, I entered it, >>and now you take it back and change it. I lost all my pondering stuff..." >> >>It would certainly solve a lot of problems without introducing any new ones. >>No lost time from operators. No clock synchronization since the server would >>maintain the time electronically. Seems like a reasonable idea. A simple >>laptop with linux and the FICS code would do the trick. A NIC for the laptop, >>a switch/hub for everyone to connect to, and a NIC for each competitor and the >>world would be much simpler for the TD. > >Devil's advocate says: >1) How would you deal with a crash or communication problems? The rules cover this. If you think the machine has crashed, you call the TD over and tell him. He will then allow you to do whatever you need to do to ask "are you alive and well" (ie pop up a new dos window in windows, for example). If the machine has crashed, you reboot (the clock is not running for two of these failures) restart the engine, set everything back to just as it was at the failure (same time controls, clock time, contempt factor, etc.) and then the game continues. If it is a communication problem, the rules say you may stop your clock immediately and tell the TD. The clock stays stopped until communications are restored. In my way of automating one of these events there would be _no_ comm failures as the chess server and all clients would be directly connected to each other at the tournament site. >2) Computer will be forced to follow a protocol. Why a standalone computer >or a experimental supercomputer should be forced to follow a protocol? (that >they might not be able to follow perhaps). Cray Blitz could follow any protocol. You can run xboard on a Cray, for example. Or you can run it on a linux box and run the engine on a Cray as far as that goes. Stand-alone computers are a thing of the past basically and have not been competing in WMCCC events. >3) More debugging to be done in something that has nothing to do with chess >(the protocol and the connection) This isn't wasted. Now you have an interface to _real_ chess servers that can greatly enhance your testing and debugging anyway. >4) Automated things are great until they fail once. At that point, the mess is >very ugly when you have contending parts. I don't think there would be any failures. FICS code is known to work. As does xboard/winboard. The number of failures would definitely be far smaller than the number of arguments caused by operator inattention. > >Regards, >Miguel
This page took 0.01 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.