Chess programming
fermath
developers. I don't know what is better: To build an engine from scratch or
take a simple engine as starting point. If so, what engine?. It sems to me
that taking advanced engines as crafty or phalanx is a bad choice, because
they are relatively complete and only minor improvements or changes can be
done.
I was thinking on engines like Pierre, by Alex Bobby (I don't know if the
source can be used for my purposes).
Pierre's homepage:
Thanx
Fermath
Dann Corbit
news:aeage9$mr5$1...@peque.uv.es...
Suggestion:
1. Read papers on chess programming. The net is full of them. A web
search for "The Yellow Bishop" will turn up a boatload of them.
2. Look at what others have done. There was a question on the Winboard
Forum about it:
http://f11.parsimony.net/forum16635/messages/29018.htm
that shows lots of examples.
3. Play around with modifying other people's chess engines.
4. Write your own from scratch.
As an alternative (if you are going to modify someone else's engine) you
will need to get permission, unless it is GPL. If it is GPL, then you
legally must publish all the changes you make.
I think the outline above will prove the most satisfying.
You will also want to investigate the computer chess club, the winboard
forum, the crafty mailing list and the chess engines mailing list.
--
C-FAQ: http://www.eskimo.com/~scs/C-faq/top.html
"The C-FAQ Book" ISBN 0-201-84519-9
C.A.P. FAQ: ftp://cap.connx.com/pub/Chess%20Analysis%20Project%20FAQ.htm
Simon Waters
>
> 3. Play around with modifying other people's chess engines.
>
> will need to get permission, unless it is GPL. If it is GPL, then you
> legally must publish all the changes you make.
You are of course welcome to modify GNU Chess which is under the
GNU GPL.
It plays a good game of chess (well it beats all it's
programmers fairly reliably), but there is plenty of scope for
polish and features, particularly surrounding preserving and
reusing information from one position in a later position, which
should get your mind around the meatier issues of chess
programming.
At the moment it is kind of optimised for finding the best move
in a specific position, but as soon as a move is made all the
information discovered is pretty much thrown away!
> You will also want to investigate the computer chess club, the winboard
> forum, the crafty mailing list and the chess engines mailing list.
Hear-hear.
Lukas Geyer
> As an alternative (if you are going to modify someone else's engine) you
> will need to get permission, unless it is GPL. If it is GPL, then you
> legally must publish all the changes you make.
There are other licenses granting the right to modify sources,
e.g. BSD or Public Domain (which is not really a license). Concerning
chess software, the crafty license permits modifications, subject to
some restrictions. The GPL does not contain any requirements that you
publish your changes, or am I mistaken? (Or did I misinterpret you?)
Cheers, Lukas
Dann Corbit
news:87it4lc...@lgeyermac.math.lsa.umich.edu...
The GPL requires that you publish any changes you make. It does not say
anything about exemptions from that principle.
http://www.gnu.org/copyleft/gpl.html
http://www.wikipedia.com/wiki/viral+copyright+license
Lukas Geyer
> The GPL requires that you publish any changes you make. It does not say
> anything about exemptions from that principle.
>
> http://www.gnu.org/copyleft/gpl.html
>
> http://www.wikipedia.com/wiki/viral+copyright+license
The GPL says that if you distribute modified versions or derived work,
then you have to do it under the same conditions, i.e. GPL which in
particular means making the source available. However, nobody forces
you to redistribute your versions of the program. If, for example, you
take gnuchess, hack your private copy to make it beat crafty (I am
waiting for that, though it's unrealistic... :) ), you can run it on
FICS or any other site without showing anyone the changes you made. If
you begin to distribute it (and be it only binary versions), you have
to license it under GPL and make the source available.
If you disagree with this interpretation, please quote the specific
section from the GPL which contradicts it.
Cheers, Lukas
P.S.: The "viral" character of the GPL has in my opinion been one of
the main reasons for the strength of the free software movement. (Yes,
I know, this is off-topic here, but better than some of these other
off-topic threads with personal insults...)
Gian-Carlo Pascutto
> "Dann Corbit" <dco...@connx.com> writes:
>
> If you disagree with this interpretation, please quote the specific
> section from the GPL which contradicts it.
You are 100% right in your correction.
> P.S.: The "viral" character of the GPL has in my opinion been one of
> the main reasons for the strength of the free software movement. (Yes,
> I know, this is off-topic here, but better than some of these other
> off-topic threads with personal insults...)
I used to be a GPL advocate, thinking it was a great idea, publishing
all my stuff under GPL and so fort. Nowadays though, I've gone 180°
on that.
GPL software is only 'free' if you are using the GPL yourself. It's
like saying there's free speech but only letting those talk who
agree with you.
I can't reasonably get myself to consider that 'free' anymore. If
I publish something I want to share, I will use BSD-like licenses
in the future.
--
GCP
Simon Waters
>
> GPL software is only 'free' if you are using the GPL yourself. It's
> like saying there's free speech but only letting those talk who
> agree with you.
>
> I can't reasonably get myself to consider that 'free' anymore.
The response usually given is that without the clause governing
redistribution in the GNU GPL v2 you would be competing with
versions of the same program that had been improved and kept
private, for avoiding which I think the GNU GPL v2 is
appropriate.
I personally think that many odd little programs or snippets may
be better under a BSD licence.
Perhaps the interesting Chess related licence discussion was on
a proposed new Chess server. Since these are ASP type programs,
redistribution doesn't come into the equation, and I proposed
the Affero GPL for the project.
This licence insists that changes to the server code are made
available for download.
Since the project voted for GNU GPL v2, the question I now have
to answer is do I want to contribute to a project where anyone
can run an enhanced server, and hoard the code? This is one of
the few areas I've got involved where code hoarding has been a
substantial issue already in the field, and in this case the
original source was under the GNU GPL.
As always the answer is to think about the licence your software
uses, and see if it meets the goals and purposes you want from
your efforts. It isn't a one size fits all issue.
Jyrki Alakuijala
>
>I was thinking on engines like Pierre, by Alex Bobby (I don't know if the
>source can be used for my purposes).
>
want to look at chess programming in python.
I have had a lot of fun implementing a small python/tkinter based
computer chess program. It has its own graphical UI and the engine in
some 4000 lines of 100% pure python code.
http://www.kolumbus.fi/jyrki.alakuijala/pychess.html
You need to install python 2.0 or newer and tcl/tk to run it. You can
get these
from www.python.org.
fermath
> Dann Corbit wrote:
> >
> > 3. Play around with modifying other people's chess engines.
> >
> > As an alternative (if you are going to modify someone else's engine) you
> > will need to get permission, unless it is GPL. If it is GPL, then you
> > legally must publish all the changes you make.
>
I have done these steps (until step 3 :-)). I think you're right. It's time
to begin my own engine from scratch.
> You are of course welcome to modify GNU Chess which is under the
> GNU GPL.
>
> It plays a good game of chess (well it beats all it's
> programmers fairly reliably), but there is plenty of scope for
> polish and features, particularly surrounding preserving and
> reusing information from one position in a later position, which
> should get your mind around the meatier issues of chess
> programming.
>
> At the moment it is kind of optimised for finding the best move
> in a specific position, but as soon as a move is made all the
> information discovered is pretty much thrown away!
>
I think that these engines (GNUChess, Crafty, etc.) are too hard for me.
The source code is very long ... perhaps it will seem easier to me within
some time ... Anyway, Thank you.
> > You will also want to investigate the computer chess club, the winboard
> > forum, the crafty mailing list and the chess engines mailing list.
>
> Hear-hear.
Well, thank you. I'm happy because my post began a discussion about
licensing ...
Best regards
fermath
Dann Corbit
news:LCKO8.34114$ud....@afrodite.telenet-ops.be...
> "Lukas Geyer" <lge...@umich.edu> wrote in message
> news:87660lc...@lgeyermac.math.lsa.umich.edu...
> > "Dann Corbit" <dco...@connx.com> writes:
> >
> > If you disagree with this interpretation, please quote the specific
> > section from the GPL which contradicts it.
>
> You are 100% right in your correction.
If you only run it yourself, and never give it to anyone else, it might be
possible to interpret the contract in that way. If so, it is totally
useless, wouldn't you say? At any rate, my reading says that you must give
out the source changes if you modify it, but I am not a lawyer. I think the
problem is that people know what GPL is supposed to mean and interpret the
language with that in view. However, when read literally, it has a
different meaning. At least to me. At any rate, I don't much care. I use
and enjoy GPL software, but I prefer other license types.
> > P.S.: The "viral" character of the GPL has in my opinion been one of
> > the main reasons for the strength of the free software movement. (Yes,
> > I know, this is off-topic here, but better than some of these other
> > off-topic threads with personal insults...)
>
> I used to be a GPL advocate, thinking it was a great idea, publishing
> all my stuff under GPL and so fort. Nowadays though, I've gone 180°
> on that.
>
> GPL software is only 'free' if you are using the GPL yourself. It's
> like saying there's free speech but only letting those talk who
> agree with you.
>
> I can't reasonably get myself to consider that 'free' anymore. If
> I publish something I want to share, I will use BSD-like licenses
> in the future.
The problem I see with GPL is that it totally prevents commercial usage.
That is fine, of course. I do think that there are excellent GPL packages,
like GCC. I broadly prefer the *really* free licenses like:
1. Public Domain
2. BSD
3. ACE
4. Imatix
etc.
However, if someone wants to write GPL software, I think it is fine and
dandy. For my purposes, it turns it into a toy that can't be used for
anything serious, but there is nothing wrong with that either.
An example: The GSL -- I would love to use it but I simply cannot, so it is
nothing more than a toy that sits on my disk and that I fiddle with from
time to time.
Gian-Carlo Pascutto
news:aelgo...@enews1.newsguy.com...
> At any rate, my reading says that you must give
> out the source changes if you modify it, but I am not a lawyer. I think the
> problem is that people know what GPL is supposed to mean and interpret the
> language with that in view. However, when read literally, it has a
> different meaning. At least to me.
I'm telling you what the lawyers told me: You cannot force someone
to publish his changes if he does not distribute the program.
> The problem I see with GPL is that it totally prevents commercial usage.
> That is fine, of course. I do think that there are excellent GPL packages,
> like GCC. I broadly prefer the *really* free licenses like:
> 1. Public Domain
> 2. BSD
> 3. ACE
> 4. Imatix
> etc.
>
> However, if someone wants to write GPL software, I think it is fine and
> dandy. For my purposes, it turns it into a toy that can't be used for
> anything serious, but there is nothing wrong with that either.
>
> An example: The GSL -- I would love to use it but I simply cannot, so it is
> nothing more than a toy that sits on my disk and that I fiddle with from
> time to time.
My sentiments are similar. I still work on *cough* 'Free' *gargle*
software and I even distribute software under the GPL, but I
don't think it's _truly_ free.
--
GCP
Dann Corbit
> "Dann Corbit" <dco...@connx.com> wrote in message
> news:aelgo...@enews1.newsguy.com...
>
> > At any rate, my reading says that you must give
> > out the source changes if you modify it, but I am not a lawyer. I think
the
> > problem is that people know what GPL is supposed to mean and interpret
the
> > language with that in view. However, when read literally, it has a
> > different meaning. At least to me.
>
> I'm telling you what the lawyers told me: You cannot force someone
> to publish his changes if he does not distribute the program.
Laws vary from country to country.
Imagine this scenario:
Programmer Joe downloads GnuChess and makes something marvelous out of it.
He uses the binary to win the WMCCC.
He does or does not have to disclose the source code changes?
Simon Waters
>
> Imagine this scenario:
>
> Programmer Joe downloads GnuChess and makes something marvelous out of it.
>
> He uses the binary to win the WMCCC.
>
> He does or does not have to disclose the source code changes?
This is covered in the FAQ for the GNU GPL
http://www.gnu.org/licenses/gpl-faq.html
See "Does the GPL require that source code of modified versions
be posted to the public?"
Joe would be at liberty to keep his changes secret, however
should he distribute a copy of the program to me (or anyone
else) then anyone could demand the full modified source code.
In this instance such hoarding of improvements would be against
the spirit of the project, but not against the licence.
A world beating chess program only you can run, that you can't
sell or give away isn't going to be very much fun. Such a
program isn't going to advance human achievement, improve other
chess programs, or make you rich (unless you are taking side
bets).
This has already happened with the GPL'ed source to a well known
Chess server. The Affero GPL mentioned already attempts to
address this issue with server software.
Simon
PS: The same would be true if I downloaded Sjeng and modified it
to be world beating program (at those chess variants it isn't
already world beating at ;), since you already seem to have
contributed to an engine under the GNU GPL.
Dann Corbit
news:3D0E723E...@wretched.demon.co.uk...
> Dann Corbit wrote:
> >
> > Imagine this scenario:
> >
> > Programmer Joe downloads GnuChess and makes something marvelous out of
it.
> >
> > He uses the binary to win the WMCCC.
> >
> > He does or does not have to disclose the source code changes?
>
> This is covered in the FAQ for the GNU GPL
>
> http://www.gnu.org/licenses/gpl-faq.html
>
> See "Does the GPL require that source code of modified versions
> be posted to the public?"
>
> Joe would be at liberty to keep his changes secret, however
> should he distribute a copy of the program to me (or anyone
> else) then anyone could demand the full modified source code.
>
> In this instance such hoarding of improvements would be against
> the spirit of the project, but not against the licence.
>
> A world beating chess program only you can run, that you can't
> sell or give away isn't going to be very much fun. Such a
> program isn't going to advance human achievement, improve other
> chess programs, or make you rich (unless you are taking side
> bets).
This {nearly} has already happened, though not with a GPL program. It was
with a copyrighted open source chess program (Crafty). It was cloned w/o
permission and renamed 'Bionic':
http://home.wanadoo.nl/thstorm/docc98.html
The program placed 5th but only 1.5 points behind the champion.
> This has already happened with the GPL'ed source to a well known
> Chess server. The Affero GPL mentioned already attempts to
> address this issue with server software.
>
> Simon
>
> PS: The same would be true if I downloaded Sjeng and modified it
> to be world beating program (at those chess variants it isn't
> already world beating at ;), since you already seem to have
> contributed to an engine under the GNU GPL.
My assistance with Sjeng has been very minor (or were you talking to GCP).
I have had contributions to a large number of GPL projects, and also public
domain, ACE license, etc.
Metastabl
> news:LCKO8.34114$ud....@afrodite.telenet-ops.be...
> > "Lukas Geyer" <lge...@umich.edu> wrote in message
> > news:87660lc...@lgeyermac.math.lsa.umich.edu...
> > > "Dann Corbit" <dco...@connx.com> writes:
> > >
> > > If you disagree with this interpretation, please quote the specific
> > > section from the GPL which contradicts it.
> >
> > You are 100% right in your correction.
>
> If you only run it yourself, and never give it to anyone else, it might be
> possible to interpret the contract in that way. If so, it is totally
> useless, wouldn't you say? At any rate, my reading says that you must give
Doesn't Bruce Moreland basically do this with Ferret? I know that it's
not generally available anyway.
So it's possible that there would be other chess engine authors who
would consider a similar model - write an engine and play it on ICC
(or whatever) and never distribute it to anybody. I'm not sure what
Bruce's motivations are, but it might be a way for somebody to get
started programming engines and not getting involved in a lot of
support issues. Also if they come up with ideas while tinkering with
GPL'ed code, they can save those ideas for later when they're starting
their own engine which wouldn't necessarily have to be GPLed.
Regards,
Keith
Dann Corbit
news:d782eac7.02061...@posting.google.com...
> "Dann Corbit" <dco...@connx.com> wrote in message
news:<aelgo...@enews1.newsguy.com>...
> > "Gian-Carlo Pascutto" <nat...@hotmail.com> wrote in message
> > news:LCKO8.34114$ud....@afrodite.telenet-ops.be...
> > > "Lukas Geyer" <lge...@umich.edu> wrote in message
> > > news:87660lc...@lgeyermac.math.lsa.umich.edu...
> > > > "Dann Corbit" <dco...@connx.com> writes:
> > > >
> > > > If you disagree with this interpretation, please quote the specific
> > > > section from the GPL which contradicts it.
> > >
> > > You are 100% right in your correction.
> >
> > If you only run it yourself, and never give it to anyone else, it might
be
> > possible to interpret the contract in that way. If so, it is totally
> > useless, wouldn't you say? At any rate, my reading says that you must
give
>
> Doesn't Bruce Moreland basically do this with Ferret? I know that it's
> not generally available anyway.
No. Ferret is 100% Bruce Moreland's creation. In other words, it's not a
clone.
> So it's possible that there would be other chess engine authors who
> would consider a similar model - write an engine and play it on ICC
> (or whatever) and never distribute it to anybody. I'm not sure what
> Bruce's motivations are, but it might be a way for somebody to get
> started programming engines and not getting involved in a lot of
> support issues. Also if they come up with ideas while tinkering with
> GPL'ed code, they can save those ideas for later when they're starting
> their own engine which wouldn't necessarily have to be GPLed.
Perhaps you are thinking of Bruce Moreland's program Gerbil (which [I am
guessing] is basically the outline of Ferret with all the really clever
stuff removed). Gerbil comes with source code and shows how to write a
chess program.
Gian-Carlo Pascutto
> > Doesn't Bruce Moreland basically do this with Ferret? I know that it's
> > not generally available anyway.
>
> No. Ferret is 100% Bruce Moreland's creation. In other words, it's not a
> clone.
He wasn't talking about clones. He was talking about keeping the
program totally to yourself.
--
GCP
Gian-Carlo Pascutto
>
> > I'm telling you what the lawyers told me: You cannot force someone
> > to publish his changes if he does not distribute the program.
>
>
> Laws vary from country to country.
Copyright laws tend to be quite well standardized. Except the
ludicrous stuff like the DMCA of course.
There are some notable exceptions though. (I think it's in Brasil
that a license written in English has no value)
> Imagine this scenario:
>
> Programmer Joe downloads GnuChess and makes something marvelous out of it.
>
> He uses the binary to win the WMCCC.
>
> He does or does not have to disclose the source code changes?
Yes, but not due to the GPL. He'll have to because of tournament
rules. (or rather: can be forced to)
--
GCP
Gian-Carlo Pascutto
> "Simon Waters" <Si...@wretched.demon.co.uk> wrote in message
> news:3D0E723E...@wretched.demon.co.uk...
> This {nearly} has already happened, though not with a GPL program. It was
> with a copyrighted open source chess program (Crafty). It was cloned w/o
> permission and renamed 'Bionic':
>
> http://home.wanadoo.nl/thstorm/docc98.html
>
> The program placed 5th but only 1.5 points behind the champion.
It wasn't cloned without premission. The permission was given
in the license.
The problem of Bionic was that
a) they did not ask the author before joining a tournament (they
did ask the organizers)
b) they never published their changes
b) is an infringement in Crafty's license, which states that you
must give back improvements.
--
GCP
Metastabl
> "Metastabl" <silen...@hotmail.com> wrote in message
> news:d782eac7.02061...@posting.google.com...
> > "Dann Corbit" <dco...@connx.com> wrote in message
> news:<aelgo...@enews1.newsguy.com>...
> > > "Gian-Carlo Pascutto" <nat...@hotmail.com> wrote in message
> > > news:LCKO8.34114$ud....@afrodite.telenet-ops.be...
> > > > "Lukas Geyer" <lge...@umich.edu> wrote in message
> > > > news:87660lc...@lgeyermac.math.lsa.umich.edu...
> > > > > "Dann Corbit" <dco...@connx.com> writes:
> > > > >
> > > > > If you disagree with this interpretation, please quote the specific
> > > > > section from the GPL which contradicts it.
> > > >
> > > > You are 100% right in your correction.
> > >
> > > If you only run it yourself, and never give it to anyone else, it might
> be
> > > possible to interpret the contract in that way. If so, it is totally
> > > useless, wouldn't you say? At any rate, my reading says that you must
> give
> >
> > Doesn't Bruce Moreland basically do this with Ferret? I know that it's
> > not generally available anyway.
>
> No. Ferret is 100% Bruce Moreland's creation. In other words, it's not a
> clone.
I was responding specifically to your statement "If you only run it
yourself, and never give it to anyone else, it might be possible to
interpret the contract in that way. If so, it is totally useless,
wouldn't you say?"
I didn't mean to imply that Bruce Moreland was modifying GPL'ed code,
just that I believe that he only runs Ferret himself and that he finds
that a useful thing to do.
Regards,
Keith
Dann Corbit
> I was responding specifically to your statement "If you only run it
> yourself, and never give it to anyone else, it might be possible to
> interpret the contract in that way. If so, it is totally useless,
> wouldn't you say?"
There is some value in private code. But Bruce is not restricted at all in
what he does with Ferret. He has made a *choice* -- he has not been forced
to hide his code. He can later on open up the source code or sell the
program commercially or whatever else he might choose.
I suppose there may be some utility in making a binary only for yourself.
It has no appeal to me. I will find something with equivalent functionality
that I can share.
I don't mind GPL code, and have participated in many, many GPL projects with
contributions. But I do think that there are alternatives that are far
better.
IMO-YMMV.
Randall Jouett
Gian-Carlo Pascutto wrote:
>
> "Lukas Geyer" <lge...@umich.edu> wrote in message
> news:87660lc...@lgeyermac.math.lsa.umich.edu...
> > "Dann Corbit" <dco...@connx.com> writes:
> >
> > If you disagree with this interpretation, please quote the specific
> > section from the GPL which contradicts it.
>
> You are 100% right in your correction.
>
> > P.S.: The "viral" character of the GPL has in my opinion been one of
> > the main reasons for the strength of the free software movement. (Yes,
> > I know, this is off-topic here, but better than some of these other
> > off-topic threads with personal insults...)
>
> I used to be a GPL advocate, thinking it was a great idea, publishing
> all my stuff under GPL and so fort. Nowadays though, I've gone 180°
> on that.
Well, I have a question: If I decide to use the gcc compiler, do I have
to make my chess program's source code publicly available? I don't
have any problems with this, BTW, but I would like to know where I stand
on this particular situation.
TIA
Randall Jouett
nsr...@bellsouth.net
remove "ns" to send e-mail directly.
Randall Jouett
Gian-Carlo Pascutto wrote:
>
> "Lukas Geyer" <lge...@umich.edu> wrote in message
> news:87660lc...@lgeyermac.math.lsa.umich.edu...
> > "Dann Corbit" <dco...@connx.com> writes:
> >
> > If you disagree with this interpretation, please quote the specific
> > section from the GPL which contradicts it.
>
> You are 100% right in your correction.
>
> > P.S.: The "viral" character of the GPL has in my opinion been one of
> > the main reasons for the strength of the free software movement. (Yes,
> > I know, this is off-topic here, but better than some of these other
> > off-topic threads with personal insults...)
>
> I used to be a GPL advocate, thinking it was a great idea, publishing
> all my stuff under GPL and so fort. Nowadays though, I've gone 180°
> on that.
Well, I have a question: If I decide to use the gcc compiler, do I have
Simon Waters
>
> Well, I have a question: If I decide to use the gcc compiler, do I have
> to make my chess program's source code publicly available?
Of course not. It is a compiler, you are free to use GCC to
write any software you like with it under any licence, that is
what free software is about.
Gian-Carlo Pascutto
news:3D12E69A...@bellsouth.net...
> Hello, Gian-Carlo and Group...
>
>
> Well, I have a question: If I decide to use the gcc compiler, do I have
> to make my chess program's source code publicly available? I don't
> have any problems with this, BTW, but I would like to know where I stand
> on this particular situation.
No, that is not needed.
--
GCP
Kym
Programs produced from a GPL platform are NOT GPLed. The only time GPL
comes into play is when you use GPL code! Ie: You modify gcc and then don't
distribute the source of your modified gcc (if you distribute).
Free = Free to modify etc. NOT free price.
Another minor variation is library code, which is why there is a LGPL!
see www.gnu.org
"Simon Waters" <Si...@wretched.demon.co.uk> wrote in message
news:3D12F9E3...@wretched.demon.co.uk...