Robodini Q&A

[bob]

Well, your analogy with the family is a bit out of scope, although I can understand that you and your family eventually live out of Houdini sales.

Obviously for you and for Richard all this is just a game - "for a bet".
For Don, myself and other commercial authors it's different. We not only have the right but the obligation to take the necessary steps to protect our livelihood.

The "game we all have chosen to play" is that good ideas in chess engines remain exclusive for very little time, pretend the opposite is unrealistic.

I have what appears to be an unrealistic expectation that RE of commercial engines by fellow engine authors is not encouraged on this forum.

Finally I don't defend nor endorse any action here.

By not reacting to Richard's misconduct, you endorse and encourage him.
It is utterly revealing that the only sentence of my post to which you replied was what you called my "my lack of self-control". You obviously didn't feel any need to write a word about Richard's lack of self-control...

Robert

I have the unrealistic expectation that engine authors are going to be HONEST about the origins of their engine, too. Seems we are in similar "boats" here, eh?

[bob]

So if you did not solicit the idea and someone like Richard publishes it, then you are in good conscience free to use it, right? But if I pump Richard for this information then I would be guilty. In other words, once the cat is out of the bag and it becomes common knowledge, there is not much the original creator of the idea can do except go after the one who revealed it and/or anyone else involved actively in the quest. Did I get that right?

Yes, you got it right.

But what if I have a truly revolutionary idea that will add 100 ELO to any program and for example Richard or someone reverse engineers the idea and then publishes it on a web site or forum and it becomes common knowledge. Your last paragraph implies that this is the same as a patent, than nobody can use (even if they independently come up with the idea on their own) because the author can "prove" it was not published prior. So I don't see how anything I said was wrong.

Here, we were talking of an idea never expressed before and then published in the open after reverse-engineering: the person doing that is guilty. If, yourself, you got the same idea independently and had it in your program already, there is nothing that can be reproached to you (you could produce your sources in court to prove you had this idea for a while, if you were challenged to do so.)*

For patents it is different: if two inventors get the same invention independently, only the one who will patent it first (and providing that this invention has never been disclosed anywhere before) will get a patent for it... If you try to patent the same invention, fruit of your work, the day after, you won't get granted a patent... and will have to get a license from the other inventor. Dura lex sed lex.


*Note anyway that copyright don't protect ideas (ideas aren't protected, not by the copyright, nor by patents, only their realization, embodiment, is), but the source code.

That is a technicality however with patents. Usually the idea IS the realization so for all practical purposes the idea is protected, at least in many cases. For example if there is only one feasible way to do something and you figure it out then your idea is HOW to realize the process and so you patent it. You can say that only the "realization" of the idea is protected and that nobody is actually forbidden from having the same idea and that might be technically correct, but you are not allowed to do anything with that. Imagine that you had knowledge of how to save someones life but was forbidden to actually use that knowledge. Saying that you were not forbidden to have the knowledge would not be very comforting to the ones losing their dead loved one because you couldn't use your knowledge.

So the prohibition of disclosing ideas from reverse-engineering doesn't come from intellectual-property but more from criminal law and business law notions. (I'm just telling this for the sake of being exact.) Reverse engineering of program is only allowed in two cases:
- it is allowed to the USER (not the competitor), for the sake of understanding the program (and only if this is not documented). And of course, disclosing his findings is forbidden.
- to the competitor for interoperability. A typical example is Open-Office foundation Reverse-engineering Word's file format, for inter-operability. Same thing for communication interfaces.

So the reverse-engineerer disclosing such ideas wouldn't be guilty on the ground of intellectual property but on the ground of criminal law and business law.

Thanks Julian. It's nice to have someone with real legal expertise on the forum!

Don

Just a little addition. Software patents in European Union are _very_ limited:
http://en.wikipedia.org/wiki/Software_patent#Europe

In the US as well. In that you simply can't patent software any longer. This changed at least 30 years ago. The only real protection here is copyright, which can extend to hardware since EEPROMS can be copied, which is copying a binary, which is copyrighted.

[mar]

Speaking of clones in this situation is ridiculous, and it means that you have not understood anything!

...If anything, there should be expressed with the powers: clone^2, clone^3, clone^4...

I understand very well
Edit: And I'm glad you're having fun:)

[bob]

So if you did not solicit the idea and someone like Richard publishes it, then you are in good conscience free to use it, right? But if I pump Richard for this information then I would be guilty. In other words, once the cat is out of the bag and it becomes common knowledge, there is not much the original creator of the idea can do except go after the one who revealed it and/or anyone else involved actively in the quest. Did I get that right?

Yes, you got it right.

But what if I have a truly revolutionary idea that will add 100 ELO to any program and for example Richard or someone reverse engineers the idea and then publishes it on a web site or forum and it becomes common knowledge. Your last paragraph implies that this is the same as a patent, than nobody can use (even if they independently come up with the idea on their own) because the author can "prove" it was not published prior. So I don't see how anything I said was wrong.

Here, we were talking of an idea never expressed before and then published in the open after reverse-engineering: the person doing that is guilty. If, yourself, you got the same idea independently and had it in your program already, there is nothing that can be reproached to you (you could produce your sources in court to prove you had this idea for a while, if you were challenged to do so.)*

For patents it is different: if two inventors get the same invention independently, only the one who will patent it first (and providing that this invention has never been disclosed anywhere before) will get a patent for it... If you try to patent the same invention, fruit of your work, the day after, you won't get granted a patent... and will have to get a license from the other inventor. Dura lex sed lex.


*Note anyway that copyright don't protect ideas (ideas aren't protected, not by the copyright, nor by patents, only their realization, embodiment, is), but the source code.

That is a technicality however with patents. Usually the idea IS the realization so for all practical purposes the idea is protected, at least in many cases. For example if there is only one feasible way to do something and you figure it out then your idea is HOW to realize the process and so you patent it. You can say that only the "realization" of the idea is protected and that nobody is actually forbidden from having the same idea and that might be technically correct, but you are not allowed to do anything with that. Imagine that you had knowledge of how to save someones life but was forbidden to actually use that knowledge. Saying that you were not forbidden to have the knowledge would not be very comforting to the ones losing their dead loved one because you couldn't use your knowledge.

So the prohibition of disclosing ideas from reverse-engineering doesn't come from intellectual-property but more from criminal law and business law notions. (I'm just telling this for the sake of being exact.) Reverse engineering of program is only allowed in two cases:
- it is allowed to the USER (not the competitor), for the sake of understanding the program (and only if this is not documented). And of course, disclosing his findings is forbidden.
- to the competitor for interoperability. A typical example is Open-Office foundation Reverse-engineering Word's file format, for inter-operability. Same thing for communication interfaces.

So the reverse-engineerer disclosing such ideas wouldn't be guilty on the ground of intellectual property but on the ground of criminal law and business law.

Thanks Julian. It's nice to have someone with real legal expertise on the forum!

Don

Just a little addition. Software patents in European Union are _very_ limited:
http://en.wikipedia.org/wiki/Software_patent#Europe

In the US as well. In that you simply can't patent software any longer. This changed at least 30 years ago. The only real protection here is copyright, which can extend to hardware since EEPROMS can be copied, which is copying a binary, which is copyrighted.

[Carotino]

The copy of the copy of the copy... Wow!

I had stopped at the "copy of the copy", has now added a new dimension and things become very complicated... I look forward to exciting new combinations (StoccoLini? RobboKomo? CraftoLito? ...Fantastic!)

I think it's fantastic to have Vitruvius, Saros and Akkad. Do you plan a new clone/derivative Roberto?

I was always clear, since the first publication. I have never denied... My father!

Speaking of clones in this situation is ridiculous, and it means that you have not understood anything!

...If anything, there should be expressed with the powers: clone^2, clone^3, clone^4...

Not a criticism Roberto, but I reckon you should stick to developing Vitruvius further and forget about your other efforts. One Ippo based engine per person is surely enough?

The gist of the whole matter is the lack of new ideas... Or at least have a few. The provocation of Richard put salt on the wound.
Personally, after almost a year of work, I have not been able to significantly improve my engine. Robert, however, he succeeded.

If it is true that Houdini comes from RobboLito (I think so, I'm sorry Robert!) Is also true that he has introduced many changes and many improvements. RobboLito is a totally free software (It is not enough to make a translation and a few small changes to change the license.), and all (ALL) were able to enjoy it. The only one who has done it, significantly, was Robert, and this is its great merit.
With Vitruvius, I started with an old version of Ivanhoe, then I continued on my way, I put my ideas. There have been many changes, some good, some bad, but now I've come to a dead end.
I believe that nearly all the "top engines" have come to a dead end. What is needed now are new ideas, not controversy.

I also noticed, among many fans, the misuse of the word "clone". Many fill their mouths, often inappropriately, with this word... Although, in the end, I care little or nothing.

Roberto.

[bob]

Exactly! Ray, I respect you highly, but in this case the evidence is really overwhelming. Complete consensus among programmers, and legally it is beyond "reasonable" doubt, not a matter of opinion. You can shut your eyes to the evidence if you wish but this attitude hurts computer chess. So please think about it and reconsider

Best wishes,
Jerry

No need to reconsider. There is zero evidence based on comparison of original source code. After the Rybka investigation debacle, that is the only evidence I accept. What he did or didn't do I don't know, but the point is that it is irrelevant to the vast majority of Houdini customers. And Chessbase clearly have no concerns about it. So learn to live with it

This is not a legal issue, since the original ippolit was made public domain. It is an ethical decision. ChessBase has not exactly been a paragon of virtue there... the same company that INTENTIONALLY crippled winboard programs when they played under the Fritz GUI, to make Fritz look good, etc.

[bob]

It seems to me some double moral standards are at work here.

I could not agree more. People like Robert H are portrayed as villains. Let's assume Robert started with Ippo source, and has improved it by nearly 150 Elo, and even more on big hardware. He is a villain. Other authors pillage Ippo source code, pillage stockfish source code, and use ideas and modified code in their own engines. They are original engine heroes !! And in the case of Ippo, assuming that it started off life as some reverse engineered Rybka, is proprietary code that those original engine heroes were never meant to see, but they seem to have no moral objection to pillaging it. And those same original engine heroes, on their moral high ground, would have no problem in wanting to get details of Richard's reverse engineered Houdini closed proprietray source code.

Computer chess is a real mess now, of that there is no doubt.

In terms of Houdini, what are the facts ? Saying something is a fact doesn't make it so. Show me original Houdini source code side by side with Ippo source code. Then we will see the facts. And while we are at it, let's also get some other commercial engine original sources and compare them too. I wonder what we would find. Maye something, maybe nothing.

The houdini issue has been shown with 100% clarity on open-chess. Direct comparison of houdini ASM vs Robolito ASM. Zero doubt.

Robert do you have any idea how much Crafty code is in Robolito.
Best,
Gerold.

Not a clue. Only comparison I have seen, that was 100% convincing, was the comparison of robolito/ivanhoe to Houdini (1.5 I think). Absolutely no doubt there. Whether that borrows any code from Crafty or not I do not know and have not looked due to a complete lack of interest in trying to read the robolito gibberish.

[bob]

Exactly! Ray, I respect you highly, but in this case the evidence is really overwhelming. Complete consensus among programmers, and legally it is beyond "reasonable" doubt, not a matter of opinion. You can shut your eyes to the evidence if you wish but this attitude hurts computer chess. So please think about it and reconsider

Best wishes,
Jerry

No need to reconsider. There is zero evidence based on comparison of original source code. After the Rybka investigation debacle, that is the only evidence I accept. What he did or didn't do I don't know, but the point is that it is irrelevant to the vast majority of Houdini customers. And Chessbase clearly have no concerns about it. So learn to live with it

This is not a legal issue, since the original ippolit was made public domain. It is an ethical decision. ChessBase has not exactly been a paragon of virtue there... the same company that INTENTIONALLY crippled winboard programs when they played under the Fritz GUI, to make Fritz look good, etc.

[Darkmoon]

Exactly! Ray, I respect you highly, but in this case the evidence is really overwhelming. Complete consensus among programmers, and legally it is beyond "reasonable" doubt, not a matter of opinion. You can shut your eyes to the evidence if you wish but this attitude hurts computer chess. So please think about it and reconsider

Best wishes,
Jerry

No need to reconsider. There is zero evidence based on comparison of original source code. After the Rybka investigation debacle, that is the only evidence I accept.

Which means, you will NEVER accept any evidence - of course RH is not going to show his code. Fortunately most here are reasonable enough to accept this evidence, which as I stated is legally beyond "reasonable" doubt. However, the one who has to learn to live with his dishonesty (and Robodini!) is Robert Houdart. Nevertheless, we won't make much progress here I fear, so let's end the debate here. Who knows, we may agree in a year's time

Best wishes,
Jerry

Unless you are a programmer this side of the argument means absolutely zilch to a consumer.

A consumer can state his/her position in favor of a product for its advancement; and, all the guilt tripping from programmers over its presumed illegality isn't going to amount to hill of beans as long as the product performs excellently and exceeded all previous attempts at development- and, the programmer continues to make great strides in its advancement.

You are shoveling against the tide.

[M ANSARI]

The fact that is so easy to clone a chess engine is nothing new, and is the reason that Vas figured that the best solution is to do a Cloud based engine. I think that it is one thing to clone or use code gleaned from a chess engine for your own use and improvement in your own engine, but it is a totally different thing when you actually post and distribute what you have gleaned on the internet. That is what happened with Rybka 3 with Ippolit and I hope this doesn't happen with H3. So far Richard has made it clear that he was just trying to make a point to how easy it is to RE a chess engine, and he has proved a point that many already know. He has not distributed the source code or posted it "anonymously" on the internet which Ippolit did. Maybe the only way to protect your code these days is some sort of hardware or USB based protection. Actually a hardware based PCI Express solution seems like a great idea to me as it would allow some extra stuff that would provide an even stronger engine. A hardware solution is a lot harder to crack than a software solution ... probably orders of magnitude harder.