Author: Dan Homan
Date: 18:19:09 09/10/98
Go up one level in this thread
On September 10, 1998 at 13:11:24, Robert Hyatt wrote: >On September 09, 1998 at 21:26:18, Keith Ian Price wrote: > >>On September 09, 1998 at 19:32:18, Moritz Berger wrote: >> >>>On September 09, 1998 at 18:59:21, Danniel Corbit wrote: >>> >>>>I would like to add some other features to the proposal besides auto232: >>>> >>>>1. How about a network protocol for Chess message passing? >>>> >>>> >>>>2. How about allowing connections to FICS and ICC? >>>> >>>>Computer products could even *automagically* register themselves as computers >>>>and cut down on cheating. >>>> >>>>3. How about an automatic email interface for long time controls like KKUP2? >>>> >>>> >>>>Serial port seems a little bit archaic. Same for parallel port. How many >>>>people cannot afford a net card and yet can afford multiple PC's? >>> >>>The big reason why no different standard has emerged so far is the compatibility >>>with old (mostly DOS based) chess programs, namely >>> >>>- Chess Genius 2,3,4,5 >>>- M-Chess PRO 4,5,6,7 >>>- Hiarcs 3,4,5,6 >>>- REBEL 6,7,8,9 >>>- Fritz 3,(5) >>>- Junior 5(Fritz.32 engine) >>>- Nimzo 3,3.5,98 >>>- Shredder 1,2 >>>- Kallisto >>>- CSTal >>>- Crafty (DOS, up to release 14, hopefully soon in 15.x) >>>- Comet >>>etc. >>> >>>Moritz >> >>I don't buy that as a valid excuse for not adapting a new standard. You could >>keep the already written Auto232 capacity for backward compatiblity, while >>adding the new features to get people (like me) to upgrade. I will buy any chess >>engine so long as it is as strong as Rebel Decade, and has these features of >>ICC/FICS compatiblity, e-mail compatibility, and network compatibility. Even >>though I said that I would not pay $100 for Crafty, I would if it had a UI and >>these features. Computer chess strength is secondary to me. Chess strength can't >>be determined by how a program plays against other programs, only strength in >>computer chess. Features like this are what I would like. Chess System Tal is >>much more fun to play than any of the supposedly stronger programs, but if the >>Windows version has only a Windows interface and a stronger engine, I won't be >>buying it, since it won't be any more fun for me to use than the one I have >>already. I am probably in the minority with this view, but there it is... >> >>kp > > > >I agree totally. the Auto232 standard is gross. Completely gross. There are >many better ways to accomplish this task. First problem is that the message >format is rediculous, with the original auto232 interface not supporting *real* >chess since it didn't allow underpromotion. Then there were the timing issues >that resulted in hangs when a program moved too quickly. Cryptic move format >requiring a tab here, no tab there, etc... > >None of it made any sense from a software engineering point of view. I would >be more than happy to sit down with a group and work out a standard >communication interface that is easy to implement, easy to parse, and easy to >understand how it is supposed to work. > >We ought to be able to also provide some basic software that will let this work >on both unix and windows boxes (IE I can do the unix part myself, and we can >take that to make a "auto232" library that anyone using unix can call). I have >been trying to study the windows auto232 interface, but it is a nightmare, >still, because it uses the old auto232 message format with two levels of parsing >(which makes little sense). IE I send a somewhat cryptic message to the driver >(cryptic because of a byzantine format) that the driver then modifies and sends >to the other driver over the interface, which has to modify that to send it to >the engine, which has to modify that to interpret what the devil it means. > >That is not necessary. And there is *no* sense in thinking "windoze" only for >this interface, because it can work linux to windows, and linux to linux, as >well as windows to windows, if done correctly. > >Anyone interested? Shareware/Freeware guys want to take the lead here and do >this right, once and for all? I like this idea very much. It seems as though the Xboard protocol is becoming rather standard as an interface communication. Xboard naturally supports ICS protocol, so why not make the ICS protocol the standard? I know it is not ideal, but its current wide use makes it convienent and new programs that match the standard can play on ICC or FICS with no modification. The FICS software is freely available and supports timestamping. It does require an intermediate computer to play the role as judge, but this can be an old 486 and can serve as the judge for many simultaneous matches between computers. The TCP/IP protocol is very wide-spread so matches between many types of computers are possible. The FICS software does have many features that are not necessary for comp-comp testing in your basement or over the internet, but these features don't need to be included in the offical protocol. We could define a board style (Xboard and Robofics use 12) and other essential commands as 'must have' and the rest can be at the descretion of the programmer. Not a new standard, but perhaps a convienent one. - Dan
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.