New Engine: Robodini 1.1 w32 & x64 !!!

Discussion about chess-playing software (engines, hosts, opening books, platforms, etc...)
ThinkingALot
Posts: 144
Joined: Sun Jun 13, 2010 7:32 am
Contact:

Re: New Engine: Robodini 1.1 w32 & x64 !!!

Post by ThinkingALot » Fri Feb 01, 2013 9:44 am

Jeremy, could you please extend the post editing time a bit? I wasn't able to correct a grammar mistake in my previous post :)
I think it should be "made it clear they didn't care" instead of "made it clear they do not care"...
Last edited by ThinkingALot on Fri Feb 01, 2013 9:49 am, edited 1 time in total.

Jeremy Bernstein
Site Admin
Posts: 1226
Joined: Wed Jun 09, 2010 7:49 am
Real Name: Jeremy Bernstein
Location: Berlin, Germany
Contact:

Re: New Engine: Robodini 1.1 w32 & x64 !!!

Post by Jeremy Bernstein » Fri Feb 01, 2013 9:47 am

ThinkingALot wrote:Jeremy, could you please extend the post editing time a bit? I wasn't able to correct a grammar mistake in my previous post :)
You now have 15 instead of 12 minutes. If you haven't caught your error in 15 minutes, you own it!

jb

Prima
Posts: 328
Joined: Tue Dec 14, 2010 6:12 am

Re: New Engine: Robodini 1.1 w32 & x64 !!!

Post by Prima » Fri Feb 01, 2013 10:04 am

ThinkingALot wrote:
Prima wrote:What seems to continually escape these folks is; Robert Houdart did not use the Public Domain Robbos to commence on development of Houdini. He used a GPL'ed RobboLito.
1) Impossible to prove.
2) Irrelevant from the moral point of view if true. GPL Robbo is virtually identical to the public domain one, Italian to English translation and time management fix being the main differences. Starting with the public domain version instead wouldn't have costed Houdini a single ELO point.
Prima wrote:eventually commercialize it AGAINST the said GPL expression
Zero evidence supporting this claim. There's no obvious connection between Houdini 1.03 and 2.0 besides the name itself :)
1. This has been proven solidly. Please see these 2 links:
http://www.open-chess.org/viewtopic.php ... 7&start=10
and
http://www.talkchess.com/forum/viewtopi ... t&start=44

2. Actually this IS relevant in the aspect of morals. Licenses, in this case GPL, are there for a reason: to deter such behaviors as previously demonstrated by Vas Rajlich and now Robert Houdart. Exactly WHY the owners of RobboLito 0.085g GPL'ed it. But let's look at it from another angle: why would Robert Houdart make a false claim that Houdini was/is original and lie about NOT using the GPL'ed RobboLito codes and/or any public domain codes? Findings turned out to be contrary to his "originality" claims. THIS is the crux of the matter. The LIES is what makes it morally relevant, else why did he try to hide it from the get go?... only to be exposed later on - as enumerated in the above links. And when asked about these findings, he dodges the questions and pretends not to see them. But the problem currently following him just isn't going to vanish away.

By the way, Kranium enumerated the fixes/polishes he and Sentinel put into THEIR RobboLito 0.085g. Lot's of them and not just translations to English and/or readable programming codes and Time-fix. See here:
http://talkchess.com/forum/viewtopic.ph ... 99&t=47045

These efforts/manual work entitles them to put a license to THEIR work. Just as Robert Houdart wants his work to be protected, albeit HE himself refuses to abide by others' rights. Hypocrisy and greed comes to mind. So the question can be asked of Robert Houdart: why didn't he use the non-English and non time-fixed and/or the non-polished versions of RobboLito which happened to be Public Domain to build Houdini? If he has the right to use this GPL RobboLito, is he not bound to abide by the GPL attached to the ONE version of RobboLito he consciously used to develop Houdini? Why should Robert Houdart be exempted from repacursions that results from NOT obeying a [GPL] license and the law? Let me guess: because he added ELO to RobboLito and is commercial? Not going to work...

He owes his co-authors (Kranium & Sentinel) a big apology and a large sum of money, either from out-of-court or in-court settlement, from the illegal sales & profits of a GPL code - expressly prohibited by the EU law in which he resides at the time/distribution/sales of Houdini. No matter how one looks at this, ethically/morally or legally, it's not looking good for Robert Houdart when everything is considered.

ThinkingALot
Posts: 144
Joined: Sun Jun 13, 2010 7:32 am
Contact:

Re: New Engine: Robodini 1.1 w32 & x64 !!!

Post by ThinkingALot » Fri Feb 01, 2013 11:06 am

Prima wrote:This has been proven solidly. Please see these 2 links
It has been proven solidly that Houdini is an Ippolit derivative. In order to prove that it's based on 0.85g3 (and not on some public domain Robbo like 0.85d3) one need to find a piece of code which is
1) not some obvious public domain stuff like PV output format or bitscan/popcnt functions;
2) present in 0.85g3;
3) absent in 0.85d3 and every other public domain Ippolit/Robbo or Ivanhoe. Every one.
Do you see now why it's impossible to prove?
Lot's of them and not just translations to English and/or readable programming codes and Time-fix.
Some of these improvements are already present in 0.85d3. Some are trivial. Some are obviously absent in Houdini 2.0+. And a skilled programmer can implement all of them in a day.

Prima
Posts: 328
Joined: Tue Dec 14, 2010 6:12 am

Re: New Engine: Robodini 1.1 w32 & x64 !!!

Post by Prima » Fri Feb 01, 2013 8:55 pm

First, let me thank you (and simultaneously self-correct myself) for the very Robbolito version in discussion: Robbolito 0.085g3. I tend to just end at the "g", which I shouldn't do.
ThinkingALot wrote: It has been proven solidly that Houdini is an Ippolit derivative. In order to prove that it's based on 0.85g3 (and not on some public domain Robbo like 0.85d3) one need to find a piece of code which is
1) not some obvious public domain stuff like PV output format or bitscan/popcnt functions;
2) present in 0.85g3;
3) absent in 0.85d3 and every other public domain Ippolit/Robbo or Ivanhoe. Every one.
Do you see now why it's impossible to prove?
Exactly why Robert Houdart needs to release the source code of Houdini from 1.0 up to 1.5a. If I'm not mistaken, Robert Houdart had promised to do just that but later decided against it(I can't pull up the specific thread backing this up, for now). My point?.....when evidence linked Houdini to RobboLito 0.085g3, RH knew that the charade was over because it's virtually impossible for one's engine to produce the EXACT same PV outputs, analysis, and even choice-moves - as Houdini did/does when compared to its parent Ippo/RobboLito - unless codes were copied verbatim. The codes copied happened to be copied FROM RobboLito 0.085g3. Way beyond just getting "ideas" as he selectively professes in his Houdini web site.
That's why, IMHO, talented individuals like Richard Vida et.al who caught what RH was doing and exposing it should be thanked. Keep in mind, RV said he will NOT release Robbodini source code. Not to mention Kai, Kingliveson, and others who compared PV/Analysis/move outputs of Houdini to that of R0.085g3. In fact as we now know, the Robbodini bet was to prove that H3 still contains codes FROM Robbolito 0.085g3 and how easy it was to RE H3 using it's parent code; Robbolito. "Perfect" or "close resemblance" RE-ing would have been nil to impossible IF Houdini was original. That was the whole point of Robbodini, unless I'm mistken in my interpretations of recent event. Quite contrary to the notion that H2 or H3 is "Scott free" from Robbolito 0.085g3.
ThinkingALot wrote:Some of these improvements are already present in 0.85d3. Some are trivial. Some are obviously absent in Houdini 2.0+. And a skilled programmer can implement all of them in a day.
Then WHY did RH go after a [GPL] refined Robbolito 0.085g3 and not stick with 0.085d3 that was Public Domain and free of any legal obligations/compliance? What Kranium & Sentinel did was not trivial. And they also added bug fixes and features which is WHAT made Robbolito 0.085g3 a lot better & smoother than its Public Domain predecessors - and even still better than its newer version, R0.09. In the event he (RH) chose to use "the" GPL code, then proper

Code: Select all

 attribution AND compliance of the said GPL-code should be adhered to. Neither of which Robert Houdart has done to this day...

Somehow I get the sense that most people, including RH, feels that producing a stronger, multi-threaded Robbolito version entitles him to break GPL of Robbolito 0.085g3 and absolve him from any legal/ethical consequences. Might work in RH's world. Not going to work in ANY real-life court.

User avatar
kingliveson
Posts: 1388
Joined: Thu Jun 10, 2010 1:22 am
Real Name: Franklin Titus
Location: 28°32'1"N 81°22'33"W

Re: New Engine: Robodini 1.1 w32 & x64 !!!

Post by kingliveson » Fri Feb 01, 2013 10:15 pm

ThinkingALot wrote:
Prima wrote:What seems to continually escape these folks is; Robert Houdart did not use the Public Domain Robbos to commence on development of Houdini. He used a GPL'ed RobboLito.
1) Impossible to prove.
2) Irrelevant from the moral point of view if true. GPL Robbo is virtually identical to the public domain one, Italian to English translation and time management fix being the main differences. Starting with the public domain version instead wouldn't have costed Houdini a single ELO point.
Prima wrote:eventually commercialize it AGAINST the said GPL expression
Zero evidence supporting this claim. There's no obvious connection between Houdini 1.03 and 2.0 besides the name itself :)
Robert started Houdini from GPL RobboLito version put out by Norm and Milos. The validity of GPL on top Public Domain is another discussion entirely, but with great confidence, I assure you, Houdini's code base began from works of aforementioned individuals.
PAWN : Knight >> Bishop >> Rook >>Queen

User avatar
Carlos Ylich
Posts: 56
Joined: Thu Jun 10, 2010 3:04 am
Location: São Paulo

Re: New Engine: Robodini 1.1 w32 & x64 !!!

Post by Carlos Ylich » Sat Feb 02, 2013 1:40 am

kingliveson wrote:
ThinkingALot wrote:
Prima wrote:What seems to continually escape these folks is; Robert Houdart did not use the Public Domain Robbos to commence on development of Houdini. He used a GPL'ed RobboLito.
1) Impossible to prove.
2) Irrelevant from the moral point of view if true. GPL Robbo is virtually identical to the public domain one, Italian to English translation and time management fix being the main differences. Starting with the public domain version instead wouldn't have costed Houdini a single ELO point.
Prima wrote:eventually commercialize it AGAINST the said GPL expression
Zero evidence supporting this claim. There's no obvious connection between Houdini 1.03 and 2.0 besides the name itself :)
Robert started Houdini from GPL RobboLito version put out by Norm and Milos. The validity of GPL on top Public Domain is another discussion entirely, but with great confidence, I assure you, Houdini's code base began from works of aforementioned individuals.

In short, Robert Houdart is a fucking thief :|

ThinkingALot
Posts: 144
Joined: Sun Jun 13, 2010 7:32 am
Contact:

Re: New Engine: Robodini 1.1 w32 & x64 !!!

Post by ThinkingALot » Sat Feb 02, 2013 8:18 am

kingliveson wrote:Robert started Houdini from GPL RobboLito version put out by Norm and Milos.
We don't know for sure :)
Prima wrote:Then WHY did RH go after a [GPL] refined Robbolito 0.085g3 and not stick with 0.085d3 that was Public Domain and free of any legal obligations/compliance?
Probably simple recklessness if this is indeed the case.
Prima wrote:What Kranium & Sentinel did was not trivial
Trivial is too strong a word. Simple fits it better.
I noticed some guys on CCC believing the translation to be a great accomplishment (for example: "Here Norman and Milos did a huge translation job and made the free Robbolito source code understandable in English", Ed Shroder, http://talkchess.com/forum/viewtopic.php?t=47073). Well, this is how it is done. You just pick the source files one by one (starting with headers), look for Italian words, pick one, open the quick replace dialog in visual studio, look up the word translation using Google Translate, pick the best fitting translation, click the "Replace All" button. As you progress through the files it becomes easier to understand the code. I was half through the process myself when Kranium's version appeared. My estimate is 2h-4h for the entire translation. And this translation is already present in 0.85d3.
Btw, bugfixing and porting to Windows was probably done by Yuri Osipov. I managed to google this: "RobboLito RobboLito 0.084 x86 sources and exe for Windows by Y.Osipov" (from a post on immortal forum).

User avatar
kingliveson
Posts: 1388
Joined: Thu Jun 10, 2010 1:22 am
Real Name: Franklin Titus
Location: 28°32'1"N 81°22'33"W

Re: New Engine: Robodini 1.1 w32 & x64 !!!

Post by kingliveson » Sat Feb 02, 2013 5:54 pm

ThinkingALot wrote:
kingliveson wrote:Robert started Houdini from GPL RobboLito version put out by Norm and Milos.
We don't know for sure :)
<snip>
Who is we?
PAWN : Knight >> Bishop >> Rook >>Queen

ThinkingALot
Posts: 144
Joined: Sun Jun 13, 2010 7:32 am
Contact:

Re: New Engine: Robodini 1.1 w32 & x64 !!!

Post by ThinkingALot » Sun Feb 03, 2013 8:22 am

A reply to Kranium's post on CCC (http://talkchess.com/forum/viewtopic.ph ... &start=120):
0.85d3 was released by me...it's one of our 1st releases
I know. Not GPL'ed yet.
the translation is really rough, as it had really just started...
But pretty much sufficient for a chess programmer to understand the code easily. 0.09 indeed features a more polished translation, but everyone claims H1.03 was derived from 0.085g3, not 0.09. I still don't understand how are you going to prove that it was 0.085g3 which served as a base for Houdini and not 0.085d3 (Or, say, 0.085d8).
Perhaps you should post a list of changes in 0.085g3 compared to the last public domain Robbo, so we know which features exactly were protected by GPL? Something like
1) a bit better translation
2) somewhat improved time management
3) ...
...
he suggests using Google translate, but i can assure you that would not suffice 100% in this effort...
Sufficient to obtain 0.085d3.
ABYSS became 'depth'
According to Google Translate "depth" is one of the meanings of "abisso". Not that difficult, isn't it?
i can assure you that Osipov had nothing to do with our windows porting and debugging efforts

You didn't start with his Robbo? It would be logical. Why duplicate the porting effort?

Post Reply