An idea for a new WCCC format - what do you think?

[Tord Romstad]

Designing a new protocol sounds like a really bad idea.

The existing ICS protocol is actually very suitable; The handling of it only gets very complicated if you want to use all the features meant for human interaction. There is no reason for an engine to do that, if the only purpose is to play an automated game, it can pretty much ignore anything it receives.

Perhaps it could be usable, if someone can write a specification designed for a chess program author's point of view, and not for a human chess player. Reading lots of ICC help files with the goal of finding out just what you have to do to enable your program to play automatically is neither easy nor fun. I've tried.

Unless we can get either a better description of how to interface to the ICC or some sort of simple server designed especially for automated computer tournaments, I almost certainly wouldn't bother to join such a tournament.

Your position on the Glaurung / XBoard hybrid does seem a bit extreme, though, and pretty much precludes any form of internet play. I am all for outlawing use of code from others that affects move decisions (such as EGTB or book code). But if software written by others is not allowed to relay your moves, you will of course also have to write your own TCP/IP stack, (preferably using an alternative protocol, of course), your own network-card driver, re-program your modem, etc.

No, that's not the same thing. I wouldn't mind linking to and calling a few library functions for communicating with a chess server, but XBoard does far too much: It replaces half of my chess program (the entire GUI). I can't participate in an official tournament and rip out half of my program and replace it with something I didn't write.

Of course, informal, fun tournaments are another matter. I hope to be able to participate in the next tournament on your server. Will there be one next weekend?

Tord

[Tord Romstad]

I don't follow this at all. Years ago Bruce Moreland and I started at the same time to develop a custom ICS (now ICC) interface. I had one working in two days of part-time coding. It is about 700 lines of C, uses the computer-interface "style 12" and works well.

In my life, two days of part-time coding and 700 lines of C is a lot. Besides, when considering that you and Bruce are both incomparably more qualified than me for the task, I would need ten times longer time than you. I could do it, but it's not worth the trouble. Life is too short, and there are always other tournaments.

But one does not have to write ttheir own interface. I have not seen a commercial gui that won't connect to ICC. And then there is winboard/xboard which also does this perfectly...

As I said before, I can't use xboard or some commercial interface, because I didn't write them. When I participate in a public tournament, I use my own program, and not a hybrid program. Using the Glaurung engine with the XBoard GUI would be just as bad to me as using the Glaurung GUI with the Crafty engine. The user interface is a just as important part of the chess program as the engine.

This is the argument Levy and company have always used. "It is too hard." But it really isn't. ICC is the perfect choice. It works and has been working for 15+ years so that the server code is robust, and there is no development nor support needed.

And still, even after these 15+ years, only a handful of programs support it.

Xboard works perfectly and the protocol is not that complicated, particularly if one sticks to protocol version 1 which is perfectly acceptable to play games. Polyglot seems to be working as well to allow UCI programs to connect thru xboard/winboard...

See above. XBoard is not an option to me, and adding Polyglot to the mix makes it even worse. I didn't write Polyglot either.

It might take a _little_ work, but the benefits are huge compared to sticking with manual move relaying which is ridiculous for a "computer chess event" (emphasis on "computer").

I was not arguing about manual play vs automated play (I do prefer manual play, because I find it far more fun, but that's a completely different discussion), but about using the ICC or some more specialised server software optimized for facilitating automated computer tournaments. ICC was made for humans, not for computers, and it shows.

Tord

[IWB]

Hello

1. It be handled over a internet chess server such as ICC, Playchess
or something else, This could lead to some level of sponsorship
via a fee to the server to show it.

I personaly dont like the idea of playing a world championship on a remote server is a good idea. As cheating is easy you will have a suspicious situation as soon as an "outsider" wins some games vs the big players. Fakeing an engine output or nodes is much easier than to write a top class chess engine. I think an on site event is preventing cheating much better than an remote server thing (nothing is 100% of course and cheating was, is, and will be always possible and happen, but you should not make it that easy!

Bye
Ingo

PS: Just think about an "inbeween" program which is randomly multipling nodes within a certain frame and cutting or mixing PVs - some say it is already excisting!

You do realize this already happens? "people" attend the event, but they use "remote" computers. So it is already easy, and playing on a server does not make it any easier than it already is.

Sure I "know" that it can happen (maybe even happened), but usually people agree that a fraud eye to eye is more difficult than an anonym thing over the internet! I do not agree with your second sentence and can only repeat that in my opinion an internet tourney is much more in danger of cheating than an event where people are facing each other!

Have a nice weekend
Ingo

[bob]

Designing a new protocol sounds like a really bad idea.

The existing ICS protocol is actually very suitable; The handling of it only gets very complicated if you want to use all the features meant for human interaction. There is no reason for an engine to do that, if the only purpose is to play an automated game, it can pretty much ignore anything it receives.

Perhaps it could be usable, if someone can write a specification designed for a chess program author's point of view, and not for a human chess player. Reading lots of ICC help files with the goal of finding out just what you have to do to enable your program to play automatically is neither easy nor fun. I've tried.

Have you tried "help style12"??? Style 12 is the "machine-readable version that ICC sends after each move is made... Includes most everything from name, to rating, to time control, time left, move played, etc...

Unless we can get either a better description of how to interface to the ICC or some sort of simple server designed especially for automated computer tournaments, I almost certainly wouldn't bother to join such a tournament.

Your position on the Glaurung / XBoard hybrid does seem a bit extreme, though, and pretty much precludes any form of internet play. I am all for outlawing use of code from others that affects move decisions (such as EGTB or book code). But if software written by others is not allowed to relay your moves, you will of course also have to write your own TCP/IP stack, (preferably using an alternative protocol, of course), your own network-card driver, re-program your modem, etc.

No, that's not the same thing. I wouldn't mind linking to and calling a few library functions for communicating with a chess server, but XBoard does far too much: It replaces half of my chess program (the entire GUI). I can't participate in an official tournament and rip out half of my program and replace it with something I didn't write.

Of course, informal, fun tournaments are another matter. I hope to be able to participate in the next tournament on your server. Will there be one next weekend?

Tord

[bob]

I don't follow this at all. Years ago Bruce Moreland and I started at the same time to develop a custom ICS (now ICC) interface. I had one working in two days of part-time coding. It is about 700 lines of C, uses the computer-interface "style 12" and works well.

In my life, two days of part-time coding and 700 lines of C is a lot. Besides, when considering that you and Bruce are both incomparably more qualified than me for the task, I would need ten times longer time than you. I could do it, but it's not worth the trouble. Life is too short, and there are always other tournaments.

I can send you the 700 lines of C I wrote. No GUI stuff, just parsing the style12 output and sending a few commands to Crafty (setting remaining time for both sides, move played, etc)...

But one does not have to write ttheir own interface. I have not seen a commercial gui that won't connect to ICC. And then there is winboard/xboard which also does this perfectly...

As I said before, I can't use xboard or some commercial interface, because I didn't write them. When I participate in a public tournament, I use my own program, and not a hybrid program. Using the Glaurung engine with the XBoard GUI would be just as bad to me as using the Glaurung GUI with the Crafty engine. The user interface is a just as important part of the chess program as the engine.

I'm not sure "why"... A normal GUI does not play chess, it is a "User Interface" that sits between the engine and the chess server. Yes some go farther than that. And I don't consider them to be "legal entities for chess events either since they choose opening moves, handle book learning, egtb at the root, and such. But xboard???

This is the argument Levy and company have always used. "It is too hard." But it really isn't. ICC is the perfect choice. It works and has been working for 15+ years so that the server code is robust, and there is no development nor support needed.

And still, even after these 15+ years, only a handful of programs support it.

Au Contrare'... There are hardly any programs that _don't_ support it. Look at the list of participants for the CCT events which have been fully automated for years using ICS. There's not a single commercial program that has not played in these events at some time or another. And most amateurs also support xboard or UCI, either of which gives them access to ICS to play.

Xboard works perfectly and the protocol is not that complicated, particularly if one sticks to protocol version 1 which is perfectly acceptable to play games. Polyglot seems to be working as well to allow UCI programs to connect thru xboard/winboard...

See above. XBoard is not an option to me, and adding Polyglot to the mix makes it even worse. I didn't write Polyglot either.

It might take a _little_ work, but the benefits are huge compared to sticking with manual move relaying which is ridiculous for a "computer chess event" (emphasis on "computer").

I was not arguing about manual play vs automated play (I do prefer manual play, because I find it far more fun, but that's a completely different discussion), but about using the ICC or some more specialised server software optimized for facilitating automated computer tournaments. ICC was made for humans, not for computers, and it shows.

Tord

Actually, from my perspective, it was created for computers. It is perfect for that, Crafty plays chess 24/7 with no help from me required to play hundreds of games in a week... I don't see how it could be better for computers if one designed something specifically for computers. The basic required functionality is to interface between two chess players, whether one or both is human or computeris pretty well invisible to me.

[bob]

Hello

1. It be handled over a internet chess server such as ICC, Playchess
or something else, This could lead to some level of sponsorship
via a fee to the server to show it.

I personaly dont like the idea of playing a world championship on a remote server is a good idea. As cheating is easy you will have a suspicious situation as soon as an "outsider" wins some games vs the big players. Fakeing an engine output or nodes is much easier than to write a top class chess engine. I think an on site event is preventing cheating much better than an remote server thing (nothing is 100% of course and cheating was, is, and will be always possible and happen, but you should not make it that easy!

Bye
Ingo

PS: Just think about an "inbeween" program which is randomly multipling nodes within a certain frame and cutting or mixing PVs - some say it is already excisting!

You do realize this already happens? "people" attend the event, but they use "remote" computers. So it is already easy, and playing on a server does not make it any easier than it already is.

Sure I "know" that it can happen (maybe even happened), but usually people agree that a fraud eye to eye is more difficult than an anonym thing over the internet! I do not agree with your second sentence and can only repeat that in my opinion an internet tourney is much more in danger of cheating than an event where people are facing each other!

Have a nice weekend
Ingo

The opposing issue is that internet tournaments attract _lots_ of participation due to reduced time required and zero cost to the particpants. WCCC is almost dead. It is on life support at the moment and the prognosis is not good if drastic action is not taken...

[Vasik Rajlich]

Sure, that's all part of the fun.

Vas

[Vasik Rajlich]

Tord,

putting aside the minor practical issues - if you were to start from scratch and put together a completely new server & protocol, it would soon be as complex and messy as what we have now, the same way that Glaurung is now as messy as Gothmog. Those quirks in the ICS protocol are probably there to deal with issues you haven't even dreamed about yet.

I can't participate in an official tournament and rip out half of my program and replace it with something I didn't write.

Of course, informal, fun tournaments are another matter. I hope to be able to participate in the next tournament on your server. Will there be one next weekend?

Fun is what this is all about!

Vas

[Vasik Rajlich]

You do realize this already happens? "people" attend the event, but they use "remote" computers. So it is already easy, and playing on a server does not make it any easier than it already is.

Sure I "know" that it can happen (maybe even happened), but usually people agree that a fraud eye to eye is more difficult than an anonym thing over the internet! I do not agree with your second sentence and can only repeat that in my opinion an internet tourney is much more in danger of cheating than an event where people are facing each other!

Have a nice weekend
Ingo

Is there some evidence to suggest that this is true. For example, a past CCT winner who you think did something like this?

Formally, everything you can do in an internet tournament can also be done via remote play.

Vas

[Vasik Rajlich]

Hello

1. It be handled over a internet chess server such as ICC, Playchess
or something else, This could lead to some level of sponsorship
via a fee to the server to show it.

I personaly dont like the idea of playing a world championship on a remote server is a good idea. As cheating is easy you will have a suspicious situation as soon as an "outsider" wins some games vs the big players. Fakeing an engine output or nodes is much easier than to write a top class chess engine. I think an on site event is preventing cheating much better than an remote server thing (nothing is 100% of course and cheating was, is, and will be always possible and happen, but you should not make it that easy!

Bye
Ingo

PS: Just think about an "inbeween" program which is randomly multipling nodes within a certain frame and cutting or mixing PVs - some say it is already excisting!

You do realize this already happens? "people" attend the event, but they use "remote" computers. So it is already easy, and playing on a server does not make it any easier than it already is.

Sure I "know" that it can happen (maybe even happened), but usually people agree that a fraud eye to eye is more difficult than an anonym thing over the internet! I do not agree with your second sentence and can only repeat that in my opinion an internet tourney is much more in danger of cheating than an event where people are facing each other!

Have a nice weekend
Ingo

The opposing issue is that internet tournaments attract _lots_ of participation due to reduced time required and zero cost to the particpants. WCCC is almost dead. It is on life support at the moment and the prognosis is not good if drastic action is not taken...

It's quite evident that something needs to be done. The WCCC should be the highlight of the tournament calendar. It should feature widespread participation from all ranks, all games should be broadcast live, etc.

Vas