Author: Robert Hyatt
Date: 06:53:16 06/22/02
Go up one level in this thread
On June 21, 2002 at 23:37:03, Keith Evans wrote: >On June 21, 2002 at 23:16:53, Robert Hyatt wrote: > >>On June 21, 2002 at 21:21:48, Keith Evans wrote: >> >>>On June 21, 2002 at 14:13:17, Robert Hyatt wrote: >>> >>>>On June 21, 2002 at 12:05:12, Jeroen Noomen wrote: >>>> >>>>>On June 21, 2002 at 11:45:40, Miguel A. Ballicora wrote: >>>>> >>>>>So you find it OK that any program that play under a specific GUI >>>>>can use ANY book that is available with that GUI? Without permission >>>>>of the author? >>>>> >>>>>You think that it is fair that such programs can have a big advantage >>>>>over others by simply copying and/or merging books together that are >>>>>not their property? >>>>> >>>>>Jeroen >>>>> >>>>> >>>> >>>>No. But neither to I find it fair that I might have to play against >>>>different programs using the _same_ book _with_ permission. I believe >>>>in one author per program, _period_. No shared books. No shared code. >>>>Each person picks a team to associate with, and that is that... >>> >>>Just curious - do you consider Eugene's tablebase access code to be a special >>>case of shared code, or would you object to people sharing that? >>> >>>-Keith >> >> >>I consider it a "special case" that is OK. Why? Because the tables are >>finite and fixed in their content. They contain perfect mate-in-N or draw >>scores for a pre-defined set of positions. If I construct a valid set of >>tables and you construct a valid set, the content will be identical. Since >>that is the case, I don't object to everyone using them. They aren't >>"custom-made" to contain suggested lines of play, for example. They would >>be exactly the same for everybody if they share, or if they "roll their own." >> >>I would object if someone had done the tables (Eugene, say) and then he only >>let a select few people use them in tournaments. Because then he would be >>exerting influence on the tournament result by favoring those participants. As >>it is, he favors _nobody_ since the commercial and amateur engines can all use >>the tables equally. >> >>That is my complaint about "shared books". It is simply not a fair way to >>participate... based on the one-program/one-author type of rule the ICCA has >>always used... > >I have another question regarding shared code. Let's say that Slater or someone >else makes progress with some sort of chess coprocessor. Maybe a Hsu style chip >or a bitboard accelerator for example. And that same person uses it to >accelerate an existing chess program such as Crafty - what would your feelings >be about this? That would still be "crafty" IMHO. IE just because someone runs it on a monster multiple-cpu machine doesn't change what it really is. Neither would a chess chip. > >I would think that something like a bitboard processor used to accelerate Crafty >would present a problem because it would really not change the nature of Crafty, >it would be like running it on a faster CPU. So that person would be best off >cooperating with you and jointly entering an accelerated Crafty. But how about >something along the lines of a Deep Blue chip which would include it's own >unique move ordering, evaluation,... If the creator of such a chip used the >Crafty source as the basis for a chess program and Crafty perhaps even searched >the first n plies in software, would you allow it to compete in a tournament >with Crafty? Or would you try to encourage the creators of such a beast to do it >completely from scratch? Maybe this would have to be decided on a case by case >basis. > >Regards, I agree with your analysis. However, it is pretty unlikely, unless someone were actually working directly with me. There are many ways to do bitmaps, and it is likely that designing special-purpose hardware would dictate some different design approaches than what I used, because of hardware issues that require attention. I think the intent of the rules is to allow a single program to compute only once in any given tournament... >Keith
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.