Robodini Q&A

[Michel]

"He has done what no other programmer could do" should be rewritten to "he has done what most other programmers REFUSE to do."

Dear Prof. Robert Hyatt, could you please clarify what it is that Robert Houdart did that most other programmers REFUSE to do? Do most programmers refuse to use public domain code in their commercial programs?

I think he meant to say using code without proper attribution.

[Michel]

I do not believe it works even for industrial inventions.

Well I said "one may argue...".

The point is that if an invention requires a substantial investment it seems not unreasonable to offer the inventor a structured way of protecting his investment, rather than forcing him to rely on trade secrets.

"Software inventions" do not fit the "substantial investment" criterion.

[Houdini]

I am afraid that any future update of Criitter will be marked with a black flag....

Critter will never be the same regards,
Dr.D

This is just a rhetoric cheap shot. You should have a better clue of from where the first 5 top engines come from. All of them, and I mean all of them, have reached top ELO using ideas RE from somewhere, or directly or indirectly, through an open source built with heavy use of RE. And the current top engine, that in this case is the offended part, even takes the crown and is the king of the "helped by RE" somewhere in his genealogy tree. Critter is one of the best engines and always will be, even better in the future.

Please take a careful look at Kai's most recent similarity dendogram. Note the position of Critter 1.2, 1.4 and 1.6 with respect to Houdini, and compare to what we know are 100% RE clones of Houdini: Strelka 5 and Robodini 1.1. Quite remarkable, wouldn't you agree?
Other than Critter, AFAIK no other top 5 engine uses direct RE material of another top 5 engine. It's rather annoying to have a vampire sitting on the back of Houdini.

While it's very nice to see your unconditional endorsement of Richard's actions, is it possible that your judgement is blurred by the fact that you received the Critter 1.X sources from your generous friend with the express intent of improving Stockfish?

Regarding the spread of the Robodini virus, have you been infected (have you downloaded Robodini) ? In this case you could "sanitize" yourself deleting it from your hard disk.

The internet being what it is, the engine will remain forever in circulation. Just like Strelka 5 it will find its way into rating lists and engine tournaments.
The damage to Houdini and to Richard is permanent, I'm afraid.

Robert

[pohl4711]

"and compare to what we know are 100% RE clones of Houdini: Strelka 5 and Robodini 1.1"

But Strelka plays a lot of different moves (compared with Houdini 1.5a) and with a different style of play: If you look in my LS ratinglist you can see, that Strelka has a very high draw-rate (like all Ippos (and Critter)) and Houdini 1.5a has a very low draw-rate. So I believe it is strongly recommended for any engine-tester, to test Strelka.
The Robodini-test is running at the moment, so I cannot say anything about this one (because I prefer own testing-data for my own opinion...)
In a small test with Robodini and Houdini 3 running at the same time in the FritzGUI (one core per engine), Robodini played sometimes a different move, but not very often (around 5-10%).

Stefan

http://ls-ratinglist.beepworld.de/

[mcostalba]

While it's very nice to see your unconditional endorsement of Richard's actions, is it possible that your judgement is blurred by the fact that you received the Critter 1.X sources from your generous friend with the express intent of improving Stockfish?

I'd prefer to discuss with you when you stick to facts, indulging in imagining hidden agendas and obscure plots does not make you honour.

I am not blurred by anything regarding the Critter sources that I have read (as have done all the top engines developers, with perhaps the exception of Vas) to get some good idea to improve the engine I develop both in strength and in clearness/simplicity: when I say that Richard is a talented developer I say it for a reason.

And regarding the facts, you can say all you want about Richard, but you can't deny that all he has done has been transparent and clear, even when he RE Houdini for the first time he immediately said it. For me this is a value....and I won't do comparisons at this point to avoid reading another sub-par reply from your side.

[Houdini]

I'd prefer to discuss with you when you stick to facts, indulging in imagining hidden agendas and obscure plots does not make you honour.

I was talking about blurred judgement, not about "hidden agendas" and "obscure plots".

So far you've only been commenting on side-issues and expressing your unconditional support for Richard, while carefully avoiding anything of substance.
I am interested in knowing your comment to:
1) The position of Critter in the current similarity dendogram;
2) Richard's RE of Houdini and subsequent publishing of a binary and sources (in the past).

Robert

[Rebel]

I'd prefer to discuss with you when you stick to facts, indulging in imagining hidden agendas and obscure plots does not make you honour.

I was talking about blurred judgement, not about "hidden agendas" and "obscure plots".

So far you've only been commenting on side-issues and expressing your unconditional support for Richard, while carefully avoiding anything of substance.
I am interested in knowing your comment to:
1) The position of Critter in the current similarity dendogram;
2) Richard's RE of Houdini and subsequent publishing of a binary and sources (in the past).

Robert

Did the thought ever occurred to you that the key of your problems with Richard is your denial of the Houdini origin?

[chessmann]

I found an interesting post (#19):

Robodini is here under IPPOLIT family engines , where it belongs

Robodini is a SP RE of Houdini 3

why is Houdini 3 not here where it belongs ?

I just wonder what will happen now, it seems that there are two solutions:
1) Robodini will be moved to Normal Engines section
2) Houdini will be moved to Ippolit Engines section

It seems there are no other ways.

[kranium]

Please take a careful look at Kai's most recent similarity dendogram. Note the position of Critter 1.2, 1.4 and 1.6 with respect to Houdini, and compare to what we know are 100% RE clones of Houdini: Strelka 5 and Robodini 1.1. Quite remarkable, wouldn't you agree?
Other than Critter, AFAIK no other top 5 engine uses direct RE material of another top 5 engine. It's rather annoying to have a vampire sitting on the back of Houdini.

Robert

the relationship between Houdini, Strelka 5, and Robodini 1.1 is no different than the relationship between Houdini 1 and Robbolito,
except, of course, for the 'method' used to obtain the source code...
both equally unacceptable

[Lavir]

Sorry, your choice of words is poor. "He has done what no other programmer could do" should be rewritten to "he has done what most other programmers REFUSE to do." There is a HUGE difference between the two statements...

That's the perfect alibi, nothing to say about it. It works perfectly.

However I suppose that also if as you say it's against your principles to do something like that, it is not so for what it concerns proving a point, does it?

So why instead of cheap talk you don't prove the thing practically? Since you say that Robert has done nothing special and it's just a matter of ethics that Houdini is not just "one of a many", I suppose it would not be very difficult for you - a genius programmer - to obtain the same result better utilizing the same methodology, isn't it?

As such why don't you prove the matter practically with doing what Robert did yourself and obtaining the same result? It will not be morally unacceptable for you because you naturally do this just to prove a point (just like Richard has RE Houdini just for a bet and not naturally for himself; I mean, it's obvious!) to everybody and to let understand even those that don't believe it that everybody can do the same (and naturally you are advantaged herein because you are not a simple "everybody", isn't it?). It will indeed remove all the diatribe from the root, don't you think? So what best way?

So, to prove your point, you could do the same as Robert did (or is believed he did). You can take Ippo/Robbo (for example) plus whatever other source/idea you want and come with an engine with the same strength (or even more given that you are what you are). Naturally the only thing you cannot do is to RE Houdini 3 itself or use the source of Richard of the same (nor Robodini). What do you say about it?

I will even higher the stake up. I will bet with you on the line with what George proposed some time ago, and bet with you 5.000 Euros vs 1.000 on this point. If in a plausible time (let's say six months) you can come up starting from whatever base you want (apart the RE or Houdini) with a program with the same strength or higher than Houdini 3 (or even 2, I could accept even that) I would give you 5.000 euros, but if you cannot you will give me 1.000.

Do you accept? If you do, let's sign the contract in front of everyone.

Enough of cheap talk, it is time words are tied with some proof.

P.S: To be clear. I don't do this for Robert personally (in fact the thing could also work against him, as it could naturally be the opposite). I do this to prove what I was talking about in the other thread that people as you know absolutely nothing of what constitutes original work and a copy and what are the factors that determine one or another.

If you fail to win the bet or don't accept it, it is enough proof for me and everyone that it's so. In fact if it was just a matter of the act (copy/paste) if something is a reproduction or not, you (or anybody else) could do the same easily.