Page 1 of 20
I'm in doubt if RobboLito is a clone
Posted: Sat Oct 24, 2009 1:18 pm
by Volker Pittlik
On the Ippolito website is the source code of another program called RobboLito. Although this Ippolito seems to be of a questionable origin and also crashes a lot I took a look at this RobboLito program.
- It seems to be a "normal" organized program in several c- and header files. I can't find any similarities to an open-source programs I know. The program seems to be written by an Italian, but my knowledge of that language is not good enough that I can decide if the source code is written by a human or generated somehow. Also there aren't any comments, but in total is much better readable and understandable than the Ippolito sources.
- If RobboLitto is a successor of Ippolito then some progress has been made in the regard of stability. There wasn't a single crash in some hundreds games here.
- It comes with a program to generate its own kind of endgame tablebases. At least this part seems to be an original work.
vp
P.S. It seems to be a real monster. In a test at blitz time controls (2+2) it got this results here:
Code: Select all
1 RobboLito verzione 0.085 443 108.0 ( 94.0 : 14.0)
12.0 ( 9.0 : 3.0) Stockfish 1.5.1 278
24.0 ( 20.0 : 4.0) Toga II 1.2.1a 228
12.0 ( 9.5 : 2.5) Rybka 2.2n2 32-bit 204
12.0 ( 10.0 : 2.0) Shredder Classic 4 189
12.0 ( 11.5 : 0.5) Fruit 05/11/03 128
12.0 ( 11.0 : 1.0) Crafty-23.0 84
12.0 ( 11.0 : 1.0) Scorpio_2.2 ACCA -5
12.0 ( 12.0 : 0.0) Shredder 6 -31
I neither claim it's an original work nor I'd like to defend any cloner. But it seems to be a good idea to take a second look at the new sources too.
vp
Re: I'm in doubt if RobboLito is a clone
Posted: Sat Oct 24, 2009 1:43 pm
by swami
Robbolito's playing style is uber aggressive, positionally accurate, every single move that was made had deeper meaning - and un-computer like... simply stunning!
Re: I'm in doubt if RobboLito is a clone
Posted: Sat Oct 24, 2009 1:45 pm
by Alexander Schmidt
Volker Pittlik wrote:The program seems to be written by an Italian,
It would be nice to know if this is done by a native italian. Could one of the italians here take a look at this?
This would be the answer to the question if this is the work of a group off one.
Re: I'm in doubt if RobboLito is a clone
Posted: Sat Oct 24, 2009 1:49 pm
by zamar
I'm also starting to be in doubt. In the beginning I was mislead by the fact that Ippolito source code was looking so clearly decompiled. But instead Robbolito is clear C.
Vas is angry because someone decompiled Rybka3 and took all his brilliant ideas. But you can't have copyright on ideas! If the hacker simply took ideas and wrote everything from scratch, then Robbolito is fully legal!!
After saying this, I must say that I still have great doubts on this issue, mainly because hacker is not revealing his own name, code flow is very similar to Strelka. But I'm not anymore at all 99% certain of the issue.
Re: I'm in doubt if RobboLito is a clone
Posted: Sat Oct 24, 2009 2:07 pm
by George Tsavdaris
Volker Pittlik wrote:On the Ippolito website is the source code of another program called RobboLito. Although this Ippolito seems to be of a questionable origin and also crashes a lot I took a look at this RobboLito program.
- It seems to be a "normal" organized program in several c- and header files. I can't find any similarities to an open-source programs I know. The program seems to be written by an Italian, but my knowledge of that language is not good enough that I can decide if the source code is written by a human or generated somehow. Also there aren't any comments, but in total is much better readable and understandable than the Ippolito sources.
- If RobboLitto is a successor of Ippolito then some progress has been made in the regard of stability. There wasn't a single crash in some hundreds games here.
- It comes with a program to generate its own kind of endgame tablebases. At least this part seems to be an original work.
vp
P.S. It seems to be a real monster. In a test at blitz time controls (2+2) it got this results here:
Can you provide some instructions of how to compile Robollito to a Windows EXE or a Linux one (if Windows is not possible)?
Re: Wait for Igorrit-please !
Posted: Sat Oct 24, 2009 2:13 pm
by Sylwy
Wait for IGORRIT ( a MP version of RobboLito ) !
I think Rybka 4 is in great danger !
Re: I'm in doubt if RobboLito is a clone
Posted: Sat Oct 24, 2009 2:16 pm
by Volker Pittlik
George Tsavdaris wrote:...
Can you provide some instructions of how to compile Robollito to a Windows EXE or a Linux one (if Windows is not possible)?
It comes with a Makefile for Linux so typing "make" should be enough under that operating system. (If the c developement environment is installed what is usually true.) No idea for Windows.
vp
Re: I'm in doubt if RobboLito is a clone
Posted: Sat Oct 24, 2009 2:18 pm
by Alexander Schmidt
zamar wrote:Vas is angry because someone decompiled Rybka3 and took all his brilliant ideas.
It was OK for Vas to take other ideas when he was the one who "took" them.
Re: I'm in doubt if RobboLito is a clone
Posted: Sat Oct 24, 2009 3:49 pm
by mcostalba
Alexander Schmidt wrote:Volker Pittlik wrote:The program seems to be written by an Italian,
It would be nice to know if this is done by a native italian. Could one of the italians here take a look at this?
I gave a really quick glance at that puppy. I have to say I prefer the english version. The transaltion is very "funny" and IMHO technical terms should not be translated because they are a kind of a jargon.
Anyhow there are no comments in the code so without reading full sentences it is difficult to understand if it was machine transalted or not.
There are some spell errors here and there, for instance:
Code: Select all
A = attaco_re[POSIZIONE.bianco_re_quadretto];
That in english would be:
Code: Select all
A = king_attack[POSITION.white_king_square];
It has a spell error because attack should be written with double 'c' : 'attacco' not 'attaco'
And also transaltion of square as 'quadretto' is very funny becasue sounds absolutly silly : 'quadretto' litterally means 'little painting'
I understand the author intention was to write 'little square' because the geometrical form 'square' in italian is 'quadrato' but the correct diminutive to get 'little square' is 'quadratino' not 'quadretto' that is instead the diminutive form of 'quadro' that in italian means 'painting' (as a noun).
On another side translation of king_attack as attaco_re correctly switches the position of the verb and the noun because in italian they are reversed compared to english.
Re: I'm in doubt if RobboLito is a clone
Posted: Sat Oct 24, 2009 4:01 pm
by Alexander Schmidt
Thank you.
So in italian there are the same wrong word creations as in the english text. So he is probably no native italian speaker, and i would guess he is the same person with some english and italian knowledge.