Norm Schmidt .....and the computer chess

Discussion of anything and everything relating to chess playing software and machines.

Moderators: Harvey Williamson, bob, hgm

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
bob
Posts: 20918
Joined: Mon Feb 27, 2006 6:30 pm
Location: Birmingham, AL

Re: For Clones we dont need private police but Good logic

Post by bob » Wed Sep 02, 2009 7:32 pm

Uri Blass wrote:
bob wrote:
Bill Rogers wrote:Hi fellas
I was always under the impressiion that the software laws in this country stated that it took a complete 10% change in the software to declare it a new program. That would mean if a piece of software contained 30,000 lines of code to be able to claim a new program with sould take 3,000 new lines of code added to the original.
I worked for a software house for 15 years and although it had nothing to do with chess they operated under those rules as they claimed to know the Federal regulations for such things.
If I am wrong or there has been some changes in the Federal Rules then maybe someone can point me to the offical location where they are disputed.
Bill
According to the GPL, if you keep _one_ original line from a copied source, this is not considered original and is a violation of the GPL unless the modified program is also released under the GPL.
I do not think that it is logical that if you copy only 1 original line fron a source then this is a violation of the GPL.

The GPL may say it but one line that is the same(if you do not find more than it) can clearly be a case when the programmer even did not read the code.

It is more logical to say that you are not allowed to copy significant part when the court will decide if part that is the same considered to be a signigicant part.

It is not logical to accuse a programmer for breaking the GPL only because the program has the line i++; that is also in some program that is protected by the GPL.

Uri
Your parser is broken. What you wrote is not what I said. I said that if you copy code from a GPL program, then as long as you keep one line of that copied code in your program, you must comply with the GPL. I did not say that if you copy one line, your program must be GPL.

User avatar
Graham Banks
Posts: 34544
Joined: Sun Feb 26, 2006 9:52 am
Location: Auckland, NZ

Re: For Clones we dont need private police but Good logic

Post by Graham Banks » Wed Sep 02, 2009 7:33 pm

Le Fou numerique wrote:Hi,

What surprises me is that as we talk about law and not morality.

Will users have no qualms about using a clone?

It seems when the rating is high enough ...

Regards,
Patrick
Morality and legality are two separate issues. The law is there to protect society, whereas morality is an individual thing.
Whether one likes it or not, derivative chess engines that comply with the GPL are legal. Individuals can choose to use them or ignore them, but neither side has the right to tell the other that they're wrong.
Having said that, I don't want to see a glut of derivatives either.

Cheers,
Graham.
gbanksnz at gmail.com

User avatar
Olivier Deville
Posts: 937
Joined: Wed Mar 08, 2006 8:13 pm
Location: Aurec, France
Contact:

Re: For Clones we dont need private police but Good logic

Post by Olivier Deville » Wed Sep 02, 2009 7:34 pm

Hi Rolf

I just don't like cheaters.

Xyclops (that he even tried to sell) and later Crimson are more than enough to ban Norman Schmidt from any tournament, no matter whether his future productions are legal or not.

Olivier

bob
Posts: 20918
Joined: Mon Feb 27, 2006 6:30 pm
Location: Birmingham, AL

Re: For Clones we dont need private police but Good logic

Post by bob » Wed Sep 02, 2009 7:35 pm

hgm wrote:Uri is right, of course: like with any crime, copyright violtions can only be punished in court if it can be proven beyond reasonable doubt that you committed the crime, i.e. indeed copied the work. Identical lines in the work you alegedly copied from are not really evidence if they are common idiom of the programming language. As most programs consist for a large part of idiom, there can be a lot of coincidental matches with a high probability, making it more difficult to pass the "reasonable doubt" threshold. One line would almost certainly not be enough, (unless it happens to ontain program name and version number. :lol: )
If it is inconsequential, the question would be "why was it copied in the first place?" Which would mean it probably would not be copied. But a line like this:

# define AttacksBishop(square, occ) *(magic_bishop_indices[square]+((((occ)&magic_bishop_mask[square])*magic_bishop[square])>>magic_bishop_shift[square]))

Is not very likely to be created by another programmer, and it is a key part of magic move generation.

So we are off into the trivial discussions again. Normally someone would only copy something that is important. And that is not a simple *a=0; type line of code. We are talking about copying things that are actually worth copying.

bob
Posts: 20918
Joined: Mon Feb 27, 2006 6:30 pm
Location: Birmingham, AL

Re: For Clones we dont need private police but Good logic

Post by bob » Wed Sep 02, 2009 7:38 pm

Rolf wrote:
Alexander Schmidt wrote:
Bill Rogers wrote: point me to the legal area where it is quoted.
When useing a gpl package u agree to the gpl. I thought this should be easy to understand.

http://www.gnu.org/licenses/gpl-3.0.html

When you hack windows, change some code, do you realy think u can sell it as a new product?

Bill Gates would be very happy to get you to court :lol:
Therefore I say - also to Olivier - that with chess software it's always also about strength. How does Windows compete? How does strength have an impact? But in chess it's about strength, and only about strength.

Could you please tell me how something could be a clone if it's terribly stronger than the original? Ok, I understand if it's about legal rules, but from the viewpoint of a personality, an entity, how could something be a copy if it's in real someone else?

And Olivier is also concerned about morality. Fine, as a concept. So, we as user should care about morality. But strength has not so much impact allegedly?

I think, the whole clone concept sucks if 1) a copy is assumed if something is totally stronger and 2) if it's morally higher to use weaker software. That is counter intuitive and IMO classic nonsense. Finally it would hinder progress in computerchess.
If I were going to write another compiler from scratch, something I have done a few times over the past 41 years, I would _much_ rather copy gcc's source and then start from there. It would save me a year or two's work, so that I could start right off working on the optimizer and not have to deal with the parsing and code production parts of the compiler at all.

That's the issue here. It doesn't matter if my new compiler produces code 50% faster than gcc. It does contain gcc code in it, and this would be both a copyright violation as well as a GPL violation.

bob
Posts: 20918
Joined: Mon Feb 27, 2006 6:30 pm
Location: Birmingham, AL

Re: For Clones we dont need private police but Good logic

Post by bob » Wed Sep 02, 2009 7:40 pm

Graham Banks wrote:
Le Fou numerique wrote:Hi,

What surprises me is that as we talk about law and not morality.

Will users have no qualms about using a clone?

It seems when the rating is high enough ...

Regards,
Patrick
Morality and legality are two separate issues. The law is there to protect society, whereas morality is an individual thing.
Whether one likes it or not, derivative chess engines that comply with the GPL are legal. Individuals can choose to use them or ignore them, but neither side has the right to tell the other that they're wrong.
Having said that, I don't want to see a glut of derivatives either.

Cheers,
Graham.
I think this discussion is more about the case of copying a program, and then releasing it as your own work, with no mention of its being copied, and a complete failure to comply with the GPL and copyright laws on top of that.

User avatar
Graham Banks
Posts: 34544
Joined: Sun Feb 26, 2006 9:52 am
Location: Auckland, NZ

Re: For Clones we dont need private police but Good logic

Post by Graham Banks » Wed Sep 02, 2009 7:50 pm

bob wrote:
Graham Banks wrote:
Le Fou numerique wrote:Hi,

What surprises me is that as we talk about law and not morality.

Will users have no qualms about using a clone?

It seems when the rating is high enough ...

Regards,
Patrick
Morality and legality are two separate issues. The law is there to protect society, whereas morality is an individual thing.
Whether one likes it or not, derivative chess engines that comply with the GPL are legal. Individuals can choose to use them or ignore them, but neither side has the right to tell the other that they're wrong.
Having said that, I don't want to see a glut of derivatives either.

Cheers,
Graham.
I think this discussion is more about the case of copying a program, and then releasing it as your own work, with no mention of its being copied, and a complete failure to comply with the GPL and copyright laws on top of that.
I agree that what Norm did regarding Xyclops and Crimson was very wrong, and I can well understand the anger regarding it (I also wasted a lot of computer time testing Crimson), but there should be no issues with Cyclone which is GPL compliant.

Cheers,
Graham.
gbanksnz at gmail.com

User avatar
hgm
Posts: 25094
Joined: Fri Mar 10, 2006 9:06 am
Location: Amsterdam
Full name: H G Muller
Contact:

Re: For Clones we dont need private police but Good logic

Post by hgm » Wed Sep 02, 2009 7:59 pm

bob wrote:# define AttacksBishop(square, occ) *(magic_bishop_indices[square]+((((occ)&magic_bishop_mask[square])*magic_bishop[square])>>magic_bishop_shift[square]))
Funny you should take that as an example. Magic bitboard move generation is an algoritm that is in the public domain, and this expression is what everyone uses. Apart from the fact that parentheses are used abundanty (even where not strictly needed, but hey, even compilers get pedantic with warnings if you leave them out), what would be your evidence that this was a copy rather than an accidental resemblance? The variable names? Well, naming a shift "shift" and a bitmask "mask" is not very original, and neither is naming a variable that holds the square number "square". (Do you really expect many programs to opt for "circle" or "triangle" here?) That "bishop" is included because it is for bishop moves could be considered a bit obvious too. And everyone calls it "magic" bitboads, I have never seen it called "multiplicative" bitboards or "two-ray move generation".

If the style of naming variables in the rest of the program includes stringing together words by underscores (very common style in C programs) this is really very far from "beyond reasonable doubt"... You could not get a conviction based on this line. And the presence of an additional rook line like it would not help, because how that rook line would look is fully implied by how the bishop line looks, and thus offers no independent evidence.

User avatar
Olivier Deville
Posts: 937
Joined: Wed Mar 08, 2006 8:13 pm
Location: Aurec, France
Contact:

Re: For Clones we dont need private police but Good logic

Post by Olivier Deville » Wed Sep 02, 2009 8:11 pm

Graham Banks wrote:
Le Fou numerique wrote:Hi,

What surprises me is that as we talk about law and not morality.

Will users have no qualms about using a clone?

It seems when the rating is high enough ...

Regards,
Patrick
Morality and legality are two separate issues. The law is there to protect society, whereas morality is an individual thing.
Whether one likes it or not, derivative chess engines that comply with the GPL are legal. Individuals can choose to use them or ignore them, but neither side has the right to tell the other that they're wrong.
Having said that, I don't want to see a glut of derivatives either.

Cheers,
Graham.
Graham,

Are you happy you spent hours of CPU time testing Crimson (not following my advice) ?

In my opinion, Cyclone is a trojan horse. He offers you a "legal clone" or "derivative" but later he will try to cheat you once more.

But again, if you like to be cheated...

Olivier

User avatar
Graham Banks
Posts: 34544
Joined: Sun Feb 26, 2006 9:52 am
Location: Auckland, NZ

Re: For Clones we dont need private police but Good logic

Post by Graham Banks » Wed Sep 02, 2009 8:17 pm

Olivier Deville wrote: Are you happy you spent hours of CPU time testing Crimson (not following my advice) ?

In my opinion, Cyclone is a trojan horse. He offers you a "legal clone" or "derivative" but later he will try to cheat you once more.

But again, if you like to be cheated...

Olivier
No, I'm not happy about wasted computer testing time.

Yes, we should be angry about what Norm did wrong, but we should also be prepared to accept that he could learn from it and not repeat his indiscretions. Then again, you could be proven correct. Time will tell.

I value your friendship and advice Olivier, but that doesn't mean that we'll always agree on things. :wink:

Cheers,
Graham.
gbanksnz at gmail.com

Post Reply