On escaping copyright by rewriting everything

[Rolf]

Uri,

Look at the rest of the evaluation (minus Rybka's material eval, which is in place of Fruit's). Do you see any other differences?

I will add that the way Fruit evaluates endgames is very unique, and quite recognizable. If I were to "clone" Fruit, I would definitely delete that. Not that that proves anything...

So you can look at the evaluation of Rybka. I see. <g>

He can look at the evaluation of strelka that is based on rybka1
but even if you look at the evaluation of strelka there are clear differences between it and fruit's evaluation.

Uri

But then even I can conclude that his proposition is premature to say the least.

[Rolf]

Uri,

Look at the rest of the evaluation (minus Rybka's material eval, which is in place of Fruit's). Do you see any other differences?

I will add that the way Fruit evaluates endgames is very unique, and quite recognizable. If I were to "clone" Fruit, I would definitely delete that. Not that that proves anything...

I remember that there are clear differences in the way that rybka evaluate passed pawns and using ideas of fruit is clearly legal.

For example ideas like
mobility evaluation that is a linear function of the number of moves is
something that everyone is allowed to use and average between middle game and end game is also ideas that everyone is allowed to use.

Based on your words I get the impression that you are against using ideas from other programs and that you can never have evaluator for simple endgames like KQ vs KP because fruit has copyright to this idea(you do not say it directly but you say that the way that fruit evaluate endgames is unique and imply that other programmers should not learn from it if they do not want to get into trouble).

Note that I remember that old chessmaster6000 had similiar evaluation to fruit in the case of KQ vs KP and chessmaster6000 came clearly earlier to fruit.

Uri

You argue with "taking ideas" but Bob informed me that he wouldnt talk about ideas but only code. Could you please cut this node? Slowly I get the impression that the whole campaign is meant to hold the topic as such in public to always implicitely accuse Vas of a wrongdoing until he might appear here nervously and state something that could then be used against him like this "claim" that Strelka is "his" program. I have said for several times that that this was a strategic claim to avoid further complications. IMO it was NOT that he wanted to say, yes, I checked Strelka and then I declare this...it's my Rybka 1. But the ones who always react sensibly in their own defense when someone puts something into their mouths, put everything into Vas' mouth for the benefits of the campaign.

[Terry McCracken]

Uri,

Look at the rest of the evaluation (minus Rybka's material eval, which is in place of Fruit's). Do you see any other differences?

I will add that the way Fruit evaluates endgames is very unique, and quite recognizable. If I were to "clone" Fruit, I would definitely delete that. Not that that proves anything...

I remember that there are clear differences in the way that rybka evaluate passed pawns and using ideas of fruit is clearly legal.

For example ideas like
mobility evaluation that is a linear function of the number of moves is
something that everyone is allowed to use and average between middle game and end game is also ideas that everyone is allowed to use.

Based on your words I get the impression that you are against using ideas from other programs and that you can never have evaluator for simple endgames like KQ vs KP because fruit has copyright to this idea(you do not say it directly but you say that the way that fruit evaluate endgames is unique and imply that other programmers should not learn from it if they do not want to get into trouble).

Note that I remember that old chessmaster6000 had similiar evaluation to fruit in the case of KQ vs KP and chessmaster6000 came clearly earlier to fruit.

Uri

You argue with "taking ideas" but Bob informed me that he wouldnt talk about ideas but only code. Could you please cut this node? Slowly I get the impression that the whole campaign is meant to hold the topic as such in public to always implicitely accuse Vas of a wrongdoing until he might appear here nervously and state something that could then be used against him like this "claim" that Strelka is "his" program. I have said for several times that that this was a strategic claim to avoid further complications. IMO it was NOT that he wanted to say, yes, I checked Strelka and then I declare this...it's my Rybka 1. But the ones who always react sensibly in their own defense when someone puts something into their mouths, put everything into Vas' mouth for the benefits of the campaign.

0%

[GenoM]

<...>
I can add that that chess players reject the idea that rybka is based on fruit for the simple fact that
rybka does not have some productive knowledge that fruit has about simple endgames and if Vas started by modifying fruit to make it stronger you could expect him to keep productive knoweldge.<...>

Uri

Are you sure this knowledge is productive? May be Vas thinks different way.

Just to remind you. An old answer to your old [unchanged till now as we can see] opinion:

Code: Select all

Re: Is Rybbka a fruit clone? [NT] no
Author: Ryan B.
Date: 11:43:30 12/06/05

On December 06, 2005 at 08:01:10, Uri Blass wrote:

>On December 06, 2005 at 02:58:27, Ziad Haddad wrote:
>
>Rybka has serious problems in some simple endgames that fruit does not have.
>Rybka also has 64 bit version and fruit is not bitboard program.
>
>Uri

I do not think Rybka is a clone but if you turn distance pruning off in Fruit 2.1 you will seem some similar endgame bugs.  I think this is a coincidence though. 
Also is you look at the way Fabian made makemove if Fruit it would not be hard to switch to bitboards and keep a lot the same elsewhere.  There are other flags like crashes where Fruit would give an error and similar engine
output but I think both coincidence and possibly even rewriting some bugs from a
good reference in his own code are the likely reasons for this.

Ryan

[Rolf]

<...>
I can add that that chess players reject the idea that rybka is based on fruit for the simple fact that
rybka does not have some productive knowledge that fruit has about simple endgames and if Vas started by modifying fruit to make it stronger you could expect him to keep productive knoweldge.<...>

Uri

Are you sure this knowledge is productive? May be Vas thinks different way.

Just to remind you. An old answer to your old [unchanged till now as we can see] opinion:

Code: Select all

Re: Is Rybbka a fruit clone? [NT] no
Author: Ryan B.
Date: 11:43:30 12/06/05

On December 06, 2005 at 08:01:10, Uri Blass wrote:

>On December 06, 2005 at 02:58:27, Ziad Haddad wrote:
>
>Rybka has serious problems in some simple endgames that fruit does not have.
>Rybka also has 64 bit version and fruit is not bitboard program.
>
>Uri

I do not think Rybka is a clone but if you turn distance pruning off in Fruit 2.1 you will seem some similar endgame bugs.  I think this is a coincidence though. 
Also is you look at the way Fabian made makemove if Fruit it would not be hard to switch to bitboards and keep a lot the same elsewhere.  There are other flags like crashes where Fruit would give an error and similar engine
output but I think both coincidence and possibly even rewriting some bugs from a
good reference in his own code are the likely reasons for this.

Ryan

If you want to prove something scientifically you must make assumptions in conditionalis, if you then want to be certain something must seem similar and on top of that you deactivate the chess content and then the two chess programs give both their output for black jack. In that case you have proven their identity as the same chess players!

[GenoM]

hi Rolf
Are you Uri second?

[Uri Blass]

hi Rolf
Are you Uri second?

I never met Rolf and I never asked him for help.
He is free to express his opinion but we only express our opinion.

The fact that we have a similiar opinion about something does not mean that he is my second.

I do not see a reason for the question.

Uri

[GenoM]

hi Rolf
Are you Uri second?

I never met Rolf and I never asked him for help.
He is free to express his opinion but we only express our opinion.

The fact that we have a similiar opinion about something does not mean that he is my second.

I do not see a reason for the question.

Uri

I do not see an answer to my question to you.

[Zach Wegner]

I remember that there are clear differences in the way that rybka evaluate passed pawns and using ideas of fruit is clearly legal.

If by "differences" you mean weights, then you would be correct. Well, sort of. Perhaps you should take a closer look.

[Golem]

Just a naive question (not directly related to the original post but that does not seems to be a problem) :

Where can I find rybka 1.0 (beta) source code ?
I was able to find the strelka code and someone says some time ago in this forum that this code is 99% similar to the rybka source code. He says that strelka is a decompiled rybka 1.0 with a 1% added code from Jury Osipov.
If this is true, can someone provide the strelka code without the added code from Jury so that we can compile this code and confirm it is the decompiled code of rybka (the same eval must occur at the same depth).

I think that if you want to prove that rybka and fruit has similarities, the simplest way seems to compare the source code of rybka with the code of fruit, so were is the decompiled c++ source code of rybka ?