I am not sure if I understand the question. I guess you might be asking why I don't hang around in CCC anymore. I was watching the piranhas savagely attacking Albert Silver, and I said to myself, "I don't belong here."
If the question is about something else, you may have to spell it out for me.
Admin Admin
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 12:44 am
Yes, you understood the intent of my post well, your department from CCC. But why should a few people (in my case one) take away the fun and passion of our hobby? That was what I was trying to communicate.
Mclane
Posts : 3009 Join date : 2020-11-17 Age : 57 Location : United States of Europe, Germany, Ruhr area
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 12:50 am
I am one of the piranhas attacking him. IMO he deserves it, I like to Support programmers. But he is not. He took something that is for free and modified it for money.
Dann Corbit
Posts : 189 Join date : 2020-11-26
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 1:28 am
Mclane wrote:
I am one of the piranhas attacking him. IMO he deserves it, I like to Support programmers. But he is not. He took something that is for free and modified it for money.
All income is created in that way in some fashion or another. Albert broke no rules, Albert broke no laws. Anyone else was free to do the same thing. I believe that many commercial programs have done much worse, taken things in violation of the law.
The spirit of the GPL says that you can make money on GPL software. Some examples: Commercial Linux variants MySQL Those products make millions of dollars, yet the people who wrote them wrote them for free and get nothing from it.
we have this: "The terms and conditions of the GPL must be made available to anybody receiving a copy of the work that has a GPL applied to it ("the licensee"). Any licensee who adheres to the terms and conditions is given permission to modify the work, as well as to copy and redistribute the work or any derivative version. The licensee is allowed to charge a fee for this service or do this free of charge. This latter point distinguishes the GPL from software licenses that prohibit commercial redistribution. The FSF argues that free software should not place restrictions on commercial use,[47] and the GPL explicitly states that GPL works may be sold at any price."
And further: "Software under the GPL may be run for all purposes, including commercial purposes and even as a tool for creating proprietary software, such as when using GPL-licensed compilers.[48] Users or companies who distribute GPL-licensed works (e.g. software), may charge a fee for copies or give them free of charge. This distinguishes the GPL from shareware software licenses that allow copying for personal use but prohibit the commercial distribution or proprietary licenses where copying is prohibited by copyright law. The FSF argues that freedom-respecting free software should also not restrict commercial use and distribution (including redistribution):[47]"
Now, I am not saying you have to like what Albert did. I am only saying that he has done nothing wrong or illegal. It is fine to not like it if you choose that. But it is wrong to attack him when he has done nothing worthy of attack.
Dann Corbit
Posts : 189 Join date : 2020-11-26
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 1:39 am
Admin wrote:
Yes, you understood the intent of my post well, your department from CCC. But why should a few people (in my case one) take away the fun and passion of our hobby? That was what I was trying to communicate.
I have not abandoned computer chess, only CCC.
I have been as busy as ever in my computer chess fun. The project with dorszcz probably represents a thousand hours of effort. We may have analyzed 1000 positions at long time control with many engines to obtain the final set. For some of the positions, we analyzed carefully dozens of variant possibilities to ensure correctness. I also have my own projects ongoing.
My joy in computer chess is not diminished, but I see an atmosphere of persecution based on whim in CCC and that I cannot tolerate.
Admin and TheSelfImprover like this post
Dann Corbit
Posts : 189 Join date : 2020-11-26
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 2:09 am
Let me explain what I mean by this: "It is fine to not like it if you choose that. But it is wrong to attack him when he has done nothing worthy of attack."
I do not like what Vas did in taking ideas from Fruit to create Rybka. However, Vas did nothing illegal or immoral. That is why I have defended him, even though I do not like what he did. I separate my personal preferences from what society has chosen as right and wrong. I do not try to impose my personal preferences on other people because my preferences have no weight of law.
We read up above "the GPL explicitly states that GPL works may be sold at any price." I think we would all agree that this is exactly what Albert did. He sold GPL software for money. The GPL license agreement explicitly says "That is fine to do that." and even encourages it.
So, you do not have law or morality on your side if you have hatred for what Albert did. But that does not even mean that you are wrong. Only that Albert has correctly followed all of the rules and to hate him for that has no basis in rules explicitly written for conduct using the GPL license.
Albert did exactly what the license agreement says is totally fine to do.
If it annoys you, and/or if it annoys the SF programmers, they should have chosen another license because there are many license types that forbid commercial sales without explicit written permission.
Dann Corbit
Posts : 189 Join date : 2020-11-26
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 10:20 am
I think I should clarify exactly what I do not like about what Vas did. After all, my signature on CCC used to say, "Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning. But sharing ideas is an even greater virtue. We have another word for this. It is called teaching."
If I have been actively encouraging taking ideas why should I be upset if Vas took ideas? The thing that bothers me is not that Vas took an idea from Fruit. And it is not that Vas took some good ideas from Fruit. The thing that bothers me is that Vas took ALL the good ideas from Fruit. Now, in doing this he did add his own innovations, and especially in the search he added many new things. However, especially concerning the evaluation, he took all of the good ideas. That, specifically, is what bothers me about Rybka 1.0.
Now, by the time he got to Rybka 2.0, the eval is no longer mostly a collection of Fruit ideas because he had been actively working with Larry Kaufman to improve the evaluation.
So the specific versions targeted by the ICGA were actually compliant with the rules. But even so, I do see why Fabian was upset.
That, specifically, is what I did not like about what Vas had done.
Admin Admin
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 10:28 am
Dann Corbit wrote:
Admin wrote:
Yes, you understood the intent of my post well, your department from CCC. But why should a few people (in my case one) take away the fun and passion of our hobby? That was what I was trying to communicate.
I have not abandoned computer chess, only CCC.
I have been as busy as ever in my computer chess fun. The project with dorszcz probably represents a thousand hours of effort. We may have analyzed 1000 positions at long time control with many engines to obtain the final set. For some of the positions, we analyzed carefully dozens of variant possibilities to ensure correctness. I also have my own projects ongoing.
My joy in computer chess is not diminished, but I see an atmosphere of persecution based on whim in CCC and that I cannot tolerate.
Good to hear Dann.
Admin Admin
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 10:50 am
Dann Corbit wrote:
Let me explain what I mean by this: "It is fine to not like it if you choose that. But it is wrong to attack him when he has done nothing worthy of attack."
Cheating on your wife is not against the law, it is however unethical. What ChessBase did (not only Albert) while not against the law is unethical. Above that became all the boasting, ChessBase and Albert. That people become angry and communicate that in an orderly way is not only logical but needed also. However and here I agree with you that personal attacks all the way up to a witch hunt should be stopped it in its tracks immediately.
You were a moderator, did you try ?
Dann Corbit
Posts : 189 Join date : 2020-11-26
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 11:28 am
Admin wrote:
Dann Corbit wrote:
Let me explain what I mean by this: "It is fine to not like it if you choose that. But it is wrong to attack him when he has done nothing worthy of attack."
Cheating on your wife is not against the law, it is however unethical. What ChessBase did (not only Albert) while not against the law is unethical. Above that became all the boasting, ChessBase and Albert. That people become angry and communicate that in an orderly way is not only logical but needed also. However and here I agree with you that personal attacks all the way up to a witch hunt should be stopped it in its tracks immediately.
You were a moderator, did you try ?
What exactly, did Albert do that was unethical? He absolutely, clearly, and unmistakeably complied with the GPL license.
There were several innovations by Albert, the most important of which are: 1. Doubling the size of the net. 2. Using guided analysis, including other engines.
BOTH of these have been copied by the Stockfish team. It is not wrong to make money on GPL software. Albert's beaming pride is hardly a fault. Everyone does that.
I literally see nothing whatsoever wrong with what Albert did.
As for chessbase boasting, did you ever see (for instance) a software vendor say that their program was "WORLD CHAMPION" because it beat some weak sisters of the poor at the ICGA contest, with all of the strong programs staying home? Because I have seen that, each and every year, by whoever won the contest.
Every software vendor pushed the envelope of the truth to show how good their software is. I do not like that, but it is what it is.
To me, Albert is as pure as the driven snow, and the people attacking him are utterly, completely, hopelessly wrong.
Last edited by Dann Corbit on Fri Jul 09, 2021 11:46 am; edited 1 time in total
Dann Corbit
Posts : 189 Join date : 2020-11-26
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 11:32 am
Admin wrote:
However and here I agree with you that personal attacks all the way up to a witch hunt should be stopped it in its tracks immediately.
You were a moderator, did you try ?
I tried many times, and I cited the boorish behavior against Albert as my reason for resignation as moderator.
Admin Admin
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 11:54 am
Dann Corbit wrote:
Admin wrote:
However and here I agree with you that personal attacks all the way up to a witch hunt should be stopped it in its tracks immediately.
You were a moderator, did you try ?
I tried many times, and I cited the boorish behavior against Albert as my reason for resignation as moderator.
Bravo.
Dann Corbit
Posts : 189 Join date : 2020-11-26
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 11:59 am
Now I must head something off at the pass. Someone is going to object, "The stockfish team did not copy Albert's doubling of the net size. They used a completely different method."
Quite frankly, that is true. What I should have said is: "They attempted to copy Albert's doubling of the net idea, but instead of using his simple and elegant doubling of a template parameter size, they invented a horrific "Rube-Goldberg" tangled snare of trash that does not scale with time or GPUs, so that as soon as you increase the time above hyper-bullet time control, it does an EPIC faceplant. All of this was so that they would look like they were not simple copyists. Which, of course, would have been OK, since both projects are GPL"
Admin Admin
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 12:06 pm
Dann Corbit wrote:
Admin wrote:
Dann Corbit wrote:
Let me explain what I mean by this: "It is fine to not like it if you choose that. But it is wrong to attack him when he has done nothing worthy of attack."
Cheating on your wife is not against the law, it is however unethical. What ChessBase did (not only Albert) while not against the law is unethical. Above that became all the boasting, ChessBase and Albert. That people become angry and communicate that in an orderly way is not only logical but needed also. However and here I agree with you that personal attacks all the way up to a witch hunt should be stopped it in its tracks immediately.
You were a moderator, did you try ?
What exactly, did Albert do that was unethical? He absolutely, clearly, and unmistakeably complied with the GPL license.
There were several innovations by Albert, the most important of which are: 1. Doubling the size of the net. 2. Using guided analysis, including other engines.
BOTH of these have been copied by the Stockfish team. It is not wrong to make money on GPL software. Albert's beaming pride is hardly a fault. Everyone does that.
I literally see nothing whatsoever wrong with what Albert did.
As for chessbase boasting, did you ever see (for instance) a software vendor say that their program was "WORLD CHAMPION" because it beat some weak sisters of the poor at the ICGA contest, with all of the strong programs staying home? Because I have seen that, each and every year, by whoever won the contest.
Every software vendor pushed the envelope of the truth to show how good their software is. I do not like that, but it is what it is.
To me, Albert is as pure as the driven snow, and the people attacking him are utterly, completely, hopelessly wrong.
You surprise me there
Dann Corbit
Posts : 189 Join date : 2020-11-26
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 12:13 pm
One more exposition: The saw cuts both ways. In other words, the SF team could simply have done a copy/paste of Albert's net double idea into SF and been well within their rights. That is the nature of GPL software.
As for copying of the net itself, I am not so sure. It might be simply data, in which case it is free game. But it might be a creative work, in which case it is not.
But his idea of guiding the analysis and using other engines besides SF to perform the calculations are both up for grabs.
Admin Admin
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 12:37 pm
My view - I am not even sure about the FF2 net, it's the net with the highest similarity to SF12.
Not surprising. The fork of SF used for the project is not the latest version of SF, so it is closer to SF 12. Albert may have used SF 12 for much of the analysis. I believe he also used LC0 (and I do not know which other engines).
Admin Admin
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Alternative Chess Test 4 Fri Jul 09, 2021 8:53 pm
There is a country saying, "You can't push a rope."
For those not familiar with this expression, it has two different meanings.
The first is that those in positions of leadership must lead. The correct example must be set.
The second is that some kinds of tasks cannot be performed without cooperation. You can throw a man a rope, but if he refuses to take hold of it, it is for naught.
mwyoung
Posts : 880 Join date : 2020-11-25 Location : USA
Subject: Re: Alternative Chess Test 4 Sat Jul 10, 2021 4:43 am
Admin wrote:
Dann Corbit wrote:
Admin wrote:
Yes, you understood the intent of my post well, your department from CCC. But why should a few people (in my case one) take away the fun and passion of our hobby? That was what I was trying to communicate.
I have not abandoned computer chess, only CCC.
I have been as busy as ever in my computer chess fun. The project with dorszcz probably represents a thousand hours of effort. We may have analyzed 1000 positions at long time control with many engines to obtain the final set. For some of the positions, we analyzed carefully dozens of variant possibilities to ensure correctness. I also have my own projects ongoing.
My joy in computer chess is not diminished, but I see an atmosphere of persecution based on whim in CCC and that I cannot tolerate.
Good to hear Dann.
Welcome aboard Dann!
Not this it is important what I think. But I will say my only issue in what ChessBase has done. Not just A.S. Is not that ChessBase made a net for Stockfish. But they used the Stockfish frontend, and renamed the frontend, and then sold it to a unknowing public under a different name. To make them think it was a new engine.
If my understanding is flawed. Let me know. I could change my mind.
Somethings are just unethical, but not illegal.
Dann Corbit
Posts : 189 Join date : 2020-11-26
Subject: Re: Alternative Chess Test 4 Sat Jul 10, 2021 5:22 am
Subject: Re: Alternative Chess Test 4 Sat Jul 10, 2021 5:27 am
I initiated an email conversation with Albert and he said this about the generation of the net (I did not have it quite right): " - The FF2 NNUE net was trained on two sources of data exclusively:
1) 90% FF1 (my lc0 net) data converted to NNUE training data with an exclusively developed converter unavailable to the wider public. This converted the Lc0 training data, positions, evaluations and results, into NNUE training data. I will add that I could easily have simply downloaded tens of millions of Leela data and used it instead, and saved myself a bundle in both time and money. Of course this would be incredibly unethical, even if legal theoretically. In fact, this is what SF14 has done now, though in their case it is fine as this is a joint effort with Leela to show me the error of my ways (tongue-in-cheek).
2) 10% FF2 Beta data (this is data using the resulting NNUE net I had and trying to build on it). I spent $2700, renting 112 thread Xeon Platinums for 3 months, but only managed to squeeze out some 12 Elo this way.
Some uninformed people have suggested I simply used the data I had generated for FF1 to do this, implying my cost claims were bogus, but this is obviously false if you stop to consider this for a minute. The FF1 net was an evolving project, and as such the data at the midway stage was still that of a much much weaker net, and thus of no interest for the NNUE project. Only the very final stages, and data, were of use, which really left me with some 300-400 thousand games. I generated upwards of 15 million for the NNUE net, all using top-of-the-line GPU-machines, which was beastly expensive. My contention was that the evaluations and variety of the FF1 data was far superior to the SF12 data the SF project used for their NNUE net, and it was my goal to use this superior data to produce a richer net. I will add that the FF1 data also used its own divergences compared to the development methods of the Leela group."