Slow Chess 2.96 source released

[Jon Kreuzer]

At the bottom of this page you can find the source code to Slow Chess 2.96.
http://www.3dkingdoms.com/chess/slow.htm

For info about it:
http://www.3dkingdoms.com/chess/Readme296.txt

If anyone wants to make sure it compiles okay for them, make a faster compile, report bugs in the engine, or share their thoughts, they're welcome to reply here. Thoughts on source releases in general, anything else, (or is 560KB of bitbases too much to be part of the main download? Engame tables help play ELO-wise very little for being so huge, but it's still less than a meg download total, so I thought it was okay.)

I have no idea about strength of 2.96, since I only tried about 20 test games to see if it seemed to work. It is the successor to 2.94. v2.94 scored ~100 ELO weaker than Slow Chess Blitz in a direct match of 50 blitz games. (*not* the downloadable v0.4, I don't know how strong that one is, it seems to be stronger than 2.94, so I assume stronger than 2.96 too.) Confused yet? I better stop now.

-Jon

[Peter Eizenhammer]

Hello Jon,

great news indeed! A very strong engine (quite the opposite of what the name implies) in a new version and source code added - wow
Many thanks!

Peter

[Robert Allgeuer]

Thanks Jon,

to summarize the situation as I understand it:

- There are now two streams of SlowChess development: 2.9x and SlowChess Blitz.
- SlowChess Blitz is the stronger of the two
- 2.96 is an improved 2.94 and is now open source.
- SlowChess Blitz 0.4 is the most recent released Blitz version and also most probably the strongest released SlowChess to date.
- There is a new distribution of SlowChess Blitz 0.4 available on your website. The engine in this distribution is identical to the old Blitz download from autumn 2004, but the book has slightly changed (it is a tiny bit bigger).
- There is another unreleased and newer SlowChess Blitz version that is - I assume - stronger than SlowChess Blitz 0.4, but more importantly some 100 points stronger than 2.96.

Is this correct?

What is actually the difference between the two SlowChess streams? Better king safety I remember vaguely ...

Thanks
Robert

[Jon Kreuzer]

Peter: Thanks

Robert: You seem to have a good understanding of the situation. Something minor: the test match was actually between the unreleased SlowChess Blitz and 2.94, (both using the newest book, which as you mentioned has been updated in the SlowChess Blitz 0.4 download also.)

I think despite the results the unreleased Blitz is probably less than 100 ELO stronger than 2.94. I also played a few short test matches (nothing scientific) against Yace, Crafty Classic, Delfi, Ruffian, and Fritz 8, and Blitz ended up about 100 ELO stronger than my guess for 2.94 in those too. I'm still unconvinced that it's even close to that much stronger.

The biggest difference is midgame eval, (newest Blitz has mobility including some square control, more complicated king safety, different passed pawn scoring) although the search is also more selective, has better threat detection & extension, and certain types of Q-search checks. Every part has been fooled around with somewhat though. On the downside: It has a very low NPS, and sometimes misses moves due to pruning, and for all its evaluation, only on rare occasions does a good job at controlling the board.

There will eventually be released version of Slow Chess Blitz, but I'm in no hurry, I just like to use it to test ideas out every now and then.

[Uri Blass]

I wonder what is the motivation that push so many programmers to release source code of strong programs.

What reason do you have to release source code and reduce the chances to make money in the future from your program by helping the opponents(even if these chances are small)?

I can easily see what you can lose from releasing your program in term of earning money.

I fail to see what you can earn from it.

I talked with some people in Israel about it and inspite of the fact that they agree that the chances that I can earn money from movei are very small(because chess programs are not the direction to make money and if you are not number 1 you cannot live from it), they are against releasing the source code because there is no reason to reduce the small chances that I have to make money.

Uri

[Reinhard Scharnagl]

To Uri,

there have been a lot of times where I discussed just that problem, also in German language fora. I mainly earned a big bad echo.

There is no need to publish source code especially of strong chess programs, an author's ego would not grow by that. Instead there always is a possibility to discuss ideas without providing huge chapters of source code, by what one's intellectual ability would be underlined much better.

There is no need to distribute strong chess programs for free. Let's keep it private and avoid dumping or handle it as shareware. Don't help to feed the illusion that good chess programs are worthless.

Regards, Reinhard.

[Jon Kreuzer]

Uri:
Slow Chess was originally open source, and I felt is was time to have a something newer than the 2.82 source, but still not current. I think helping other chess programmers(opponents) is okay. Also it is possible someone might do something interesting with the source: compile a faster version, make a linux port, academic use for an idea test, etc.

I haven't released what is probably the strongest Slow Chess version (in source or .exe form.) I'm not pursuing any money from chess programming, though I have considered making Slow Chess shareware if it turns out stronger than I expect it to. I probably won't, but there is plenty of time to consider as I'm not planning any releases any time soon. I'm working at a very leisurely pace, since chess isn't my main interest.

Reinhard:

I do agree to some extent that people may not realize the work that goes into strong chess engines with so many out there freely available. But I think it's a small market, so since I've spent time on a program and want people to get some use from it, it's probably best to give it away for free. Also I published existing source instead any in-depth tutorials because I think the interest in chess programming tutorials (especially one written by me) is too small to be worth the effort to write it. I may write some more 3D-programming tutorials as I believe there's greater interest in them.

[Uri Blass]

Jon,I see that you did not release code for interface but only for the engine.

I wonder if there is some good code for interface that I can use.

I thought about the idea of writing a program to teach me and other people to play better but I gave up that idea because of the fact that I have no idea how to write an interface.

Maybe it is a simple task for an experienced programmer but I even do not know how to start.

There is one program with source code for interface that I know(hotbabe chess) but unfortunatley the source is written in a computer language that I know nothing about(Eiffel)

I wonder if there is a source for interface that is written in C or maybe it is not a good idea to use C for that purpose and I should learn another language.

Uri

[Jim Ablett]

http://slibo.sourceforge.net/

Written for Linux in C++, but source code available.

Jim.

[Tord Romstad]

Jon,

Thanks a lot for publishing the source code! The "open source team" has become much stronger over the last half year, and I hope to see (and to contribute to) this development continue.

It looks like the old (and to me completely incomprehensible) open source flame war is about to start again. There was a similar, very long thread when I released the source code of Glaurung.

I wonder what is the motivation that push so many programmers to release source code of strong programs.

What reason do you have to release source code and reduce the chances to make money in the future from your program by helping the opponents(even if these chances are small)?

I have no desire to ever sell my program. On the contrary, I would feel ashamed of even trying to do so. Most of the programs I use every day, including those which are used in the development of my program, are free and open source. They are also of exceptionally high quality, vastly superior to any software I will ever be able to create myself. It would be perfectly legal for me to use these tools to produce commercial software, but how could I do so without feeling greedy?

There is also the issue of increased responsibilities when selling a program. As long as my program is free, there is no need to spend more time than I like on documentation, minor bug fixes and support. If I sold my program, there would be no fun left in chess programming at all.

The phrase "helping my opponents" has no meaning to me. I have no opponents, just a lot of people who happen to share this strange hobby. Sharing code and ideas and helping each other improve is part of what makes chess programming fun to do.

There is no need to publish source code especially of strong chess programs, an author's ego would not grow by that.

Of course there is no need to publish the source code of chess programs. In fact, there is no real need for chess programs at all. Some of us think they are fun to create or play with, that's all. Open source engines are more useful and convenient than other engines. SlowChess, for instance, was completely useless to me until yesterday. It requires Windows, which I cannot use. Now, when I have the source code, I should be able to make it work on my Macintosh. I am very happy to have found another strong sparring partner for my engine.

There are also many other ways in which open source engines can be useful. Some people might have some interesting new search ideas, but lack the time or motivation to write a complete chess engine on their own. It is good that they have a variety of strong open source chess engines to experiment with. Some people might like an existing chess engine, but would like to add some new features. Do you want to see how well Fruit plays FRC? Do you want to see how Crafty would perform with checks in the qsearch? Would you like to run Glaurung on your mobile phone? In all cases, you can just grab the source code and do the necessary changes yourself. You can't do this with List, Aristarch or Ruffian, for obvious reasons.

Open source also has advantages for the author, because it makes it easier to receive help from other people. A good example was when I released Glaurung 0.2.0 recently. It turned out that my new opening book code, which worked perfectly in Linux and Mac OS X, didn't work in Windows. The problem would have been difficult or impossible to solve for me, but thanks to the generous help of numerous Windows programmers I had a new version with the problem fixed ready less than a day later. This couldn't have happened unless my program was open source.

The only possible negative effect I can see of publishing my source code is the risk of cloning. I don't think this is a very serious problem. A Glaurung clone would be farily easy to spot, I think.

Instead there always is a possibility to discuss ideas without providing huge chapters of source code, by what one's intellectual ability would be underlined much better.

These are not mutually exclusive. If anything, having the complete source code available facilitates discussion of ideas.

There is no need to distribute strong chess programs for free. Let's keep it private and avoid dumping or handle it as shareware. Don't help to feed the illusion that good chess programs are worthless.

There is no such illusion. Free does not imply worthless. Most of the things I value in life cannot be bought for money.

Tord

[Uri Blass]

Jon,

Thanks a lot for publishing the source code! The "open source team" has become much stronger over the last half year, and I hope to see (and to contribute to) this development continue.

It looks like the old (and to me completely incomprehensible) open source flame war is about to start again. There was a similar, very long thread when I released the source code of Glaurung.

I wonder what is the motivation that push so many programmers to release source code of strong programs.

What reason do you have to release source code and reduce the chances to make money in the future from your program by helping the opponents(even if these chances are small)?

I have no desire to ever sell my program. On the contrary, I would feel ashamed of even trying to do so. Most of the programs I use every day, including those which are used in the development of my program, are free and open source. They are also of exceptionally high quality, vastly superior to any software I will ever be able to create myself. It would be perfectly legal for me to use these tools to produce commercial software, but how could I do so without feeling greedy?

There is also the issue of increased responsibilities when selling a program. As long as my program is free, there is no need to spend more time than I like on documentation, minor bug fixes and support. If I sold my program, there would be no fun left in chess programming at all.

The phrase "helping my opponents" has no meaning to me. I have no opponents, just a lot of people who happen to share this strange hobby. Sharing code and ideas and helping each other improve is part of what makes chess programming fun to do.

Tord

I do not understand a reason that you should feel bad about selling something that you worked on it and even if you spend only few hours per week on it then the total time is still significant time.

If the purpose is getting help in specific problems then you can send the source to selected people and I do not see the advantage of making it free.

I also find that programs like Junior have bugs that the programmers do not fix for a long time.

You can be sure that Amir Ban knows about similiar problems for years(I told him about them some years ago in the time of Junior4.xx or Junior5.xx) but he does not care to fix them

see the following messages

http://www.talkchess.com/forums/1/message.html?411159
http://www.talkchess.com/forums/1/message.html?411361

Uri

[Reinhard Scharnagl]

Hi Tord,

There is no need to publish source code especially of strong chess programs, an author's ego would not grow by that.

Of course there is no need to publish the source code of chess programs. In fact, there is no real need for chess programs at all. Some of us think they are fun to create or play with, that's all. Open source engines are more useful and convenient than other engines.

I am not talking about engines but on source code. There is no need to publish chess program's source code, if you regard computer chess competitions a sort of brain sport of the programmers. Open source chess programs invite for cheating especially when sources are not given to a well known person but to anonymus public.

Instead there always is a possibility to discuss ideas without providing huge chapters of source code, by what one's intellectual ability would be underlined much better.

These are not mutually exclusive. If anything, having the complete source code available facilitates discussion of ideas.

That is nearly nonsense. Writing chess programs is independent from the selected programming language used. Thus it is possible to talk on chess ideas without having a readable code. And if you finally would be convinced in special situations that sources must be at hands, there is no need to spread it over the whole anonymus world. Sending it to your discussion partner already would do.

There is no need to distribute strong chess programs for free. Let's keep it private and avoid dumping or handle it as shareware. Don't help to feed the illusion that good chess programs are worthless.

There is no such illusion. Free does not imply worthless. Most of the things I value in life cannot be bought for money.

Because chess programs could be baught without any moralic misfeeling this is a different case where my arguments hold.

Regards, Reinhard.

[Rozens]

To Uri
Maybe you just misunderstand the idea with this forum. Some people like to share ther knowledge of programming, and others only think about making money...if you consider the last, you properly should read the story below, maybe it can solve your over all thinking money for ever..!!!

This is an easy way to make money from the post office.
Buy an Helium filled floating party balloon and write your address on it.
Go to the post office and say you want to post the balloon.
Tie the balloon onto the weighing scale and ask for the postage cost.
The floating balloon will make the scale read below zero.
The post office would thus pay you money for posting it.
You don?t even have to carry the balloon home as the post office will deliver it to your address !

Go for it Uri...and dont tell your friends....!!!

May god be with you!
Leif Rozens

I wonder what is the motivation that push so many programmers to release source code of strong programs.

What reason do you have to release source code and reduce the chances to make money in the future from your program by helping the opponents(even if these chances are small)?

I can easily see what you can lose from releasing your program in term of earning money.

I fail to see what you can earn from it.

I talked with some people in Israel about it and inspite of the fact that they agree that the chances that I can earn money from movei are very small(because chess programs are not the direction to make money and if you are not number 1 you cannot live from it), they are against releasing the source code because there is no reason to reduce the small chances that I have to make money.

Uri

[Reinhard Scharnagl]

Hi Leif,

... Some people like to share ther knowledge of programming, ...

the problem is not sharing knowledge but to share sources. Of course it could be that there are forum members who are unable to share knowledge and instead would prefer to share sources. And because moreover some of them are not able to address that sources to individuals then, they spread it as a bundle of source code, in extreme instantly compilable, too.

... others only think about making money ...

I wonder where you will find arguments for to use the word "only". Is it illegal or amoralic to earn some money in exchange for one's working results? What is your problem with that?

Reinhard.

[Rozens]

Hello Reinhard

Uri has to speak for himself....I got no problems for earning money, but gentle peoples still excist, and thats why the chess programmers gives away ther source codes...they dont think about benefit...the chess programming is a kind of hobby for them, they would like to share with other programmers.

Enough for now!

Hi Leif,

... Some people like to share ther knowledge of programming, ...

the problem is not sharing knowledge but to share sources. Of course it could be that there are forum members who are unable to share knowledge and instead would prefer to share sources. And because moreover some of them are not able to address that sources to individuals then, they spread it as a bundle of source code, in extreme instantly compilable, too.

... others only think about making money ...

I wonder where you will find arguments for to use the word "only". Is it illegal or amoralic to earn some money in exchange for one's working results? What is your problem with that?

Reinhard.

[Volker Pittlik]

Gentlemen,

this isn't the first discussions about the pros and cons of open source projects and probably it will not be the last.

Personally I agree more with Jon's and Tord's point of view. But as a moderator it seems to me the argumentation has become a little bit repetitive. It seems to me both sides have valid arguments. At the moment I can't see any progress.

I don't see any reason to put oil into the fire unnecessarly by using personal attacks or assumptions. To avoid that kind of "discussions" seems to me a pretty nice "feature" of this forum. I'd like to ask everyone to keep this "feature" alive.

Regards

Volker

[Peter Fendrich]

The opposite to open source is patents.
There is a shift going on about open source even by giants such as IBM, Sun and others. They have both participated in open source projects before but this is one step further. IBM had berfore the most rigid view of patents of the whole pack

IBM
http://www.ibm.com/news/us/en/2005/01/patents.html

Sun
http://www.sun.com/smi/Press/sunflash/2005-01/sunflash.20050125.2.html

Remember Linux and Linus Torvalds?
http://www.techworld.com/applications/news/index.cfm?NewsID=3059

I haven't released Terra as open source but if any open source author want to have my source - no problems. However, I don't think it is of much value for anyone

/Peter

[Dann Corbit]

Both ideas are equally valid.

If you work for 1000 hours on something, you have the right to:
1. Keep it
2. Give it away
3. Show what you did
4. Hide what you did
5. Talk about what you did
6. All of the above
7. None of the above

I have spent at least $1000 on chess programs and have no regrets about it. I have contributed code to at least 100 chess prorgrams and have no regrets about it. If every free chess program author wanted to stop giving away his program I have no regrets about it -- that is their right. If the strongest programs in the world wanted to publish their source code, then that is fine too.

If people choose to share then we benefit. We can see what they did and learn about it. We can spend many enjoyable hours tracing through the search to see all the intricate cleverness. In every open source program I ever looked at (including the simplest ones like TSCP) I definitely learned something worthwhile.

On the other hand, if an author wants to create a professional program and sell it, then I also benefit from that. Typically, the analysis is better. If I want to know what is most likely the best move in some position, I am going to have Tiger, Shredder, DeepSjeng, Ruffian and other tough guys attack it.

I have respect for all approaches and think people should be able to do whatever they want to with it.

For everything there are costs and benefits. For professional programs the costs and benefits are clearly obvious. We also know exactly what the costs and benefits of the open source programs are -- we have seen them discussed here many times.

So everything has its value. My opinion is that we should enjoy them all.
It is more fruitful to look at the half-full cups instead of the half empty ones. And all of them are half full. Some have kool-aid and others champaigne. But sometimes I would reach for the kool-aid, others the champaigne. Variety is the spice of life.

[Jon Kreuzer]

Uri:
The basics of the Slow Chess interface code are about the same as in the 2.82 source, so you can try looking at that. I'm not sure if I'd recommend it, it's probably difficult code to understand, more so than it needs to be. It's reasonable to use C for writing a GUI (it's what I've always used.) It can take a long time to learn enough to really be able to easily do what you want though.
Winboard is an open source C interface (I haven't done more than glance at the source, so I don't know how useful it would be.)

Tord: I have a suspicion that it might be tough to port Slow Chess since it uses Windows thread functions. Maybe it isn't much of a problem. Someday I might try changing this for the open source version.

[Dann Corbit]

Uri:
The basics of the Slow Chess interface code are about the same as in the 2.82 source, so you can try looking at that. I'm not sure if I'd recommend it, it's probably difficult code to understand, more so than it needs to be. It's reasonable to use C for writing a GUI (it's what I've always used.) It can take a long time to learn enough to really be able to easily do what you want though.
Winboard is an open source C interface (I haven't done more than glance at the source, so I don't know how useful it would be.)

Tord: I have a suspicion that it might be tough to port Slow Chess since it uses Windows thread functions. Maybe it isn't much of a problem. Someday I might try changing this for the open source version.

You can find POSIX threads that work under Windows here:
http://sources.redhat.com/pthreads-win32/

I have found it very easy to port to this interface from Windows.