Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: About Fafis...

Author: Vasik Rajlich

Date: 12:48:48 05/30/05

Go up one level in this thread


On May 30, 2005 at 05:04:31, Uri Blass wrote:

>On May 30, 2005 at 04:00:22, Mridul Muralidharan wrote:
>
>>On May 29, 2005 at 19:09:14, Vasik Rajlich wrote:
>>
>>>On May 29, 2005 at 15:08:17, Mridul Muralidharan wrote:
>>>
>>>>On May 29, 2005 at 11:35:42, Vasik Rajlich wrote:
>>>>
>>>>>On May 29, 2005 at 08:25:02, Mridul Muralidharan wrote:
>>>>>
>>>>>>On May 28, 2005 at 12:29:33, GŁnther Simon wrote:
>>>>>>
>>>>>>>On May 28, 2005 at 12:20:00, Vladimir Elin wrote:
>>>>>>>
>>>>>>>>Hi Alex,
>>>>>>>>I think that reason for you to use only engines with open sources and delete
>>>>>>>>all engines were you can see words : alpha, beta, prunning and many many etc.
>>>>>>>
>>>>>>>Vladimir that is really a dumb post after all...
>>>>>>>
>>>>>>>Guenther
>>>>>>>
>>>>>>>>You idea that Patriot 2.0 is clone - full absurd.
>>>>>>>>Best.
>>>>>>>>Vladimir.
>>>>>>
>>>>>>Is it really such a dumb post ? I am not sure - whenever someone mentions
>>>>>>"clone" , I am skeptical - inspite of the number of clones that are discovered.
>>>>>>I prefer to give the author the benifit of doubt - a genourously large benifit
>>>>>>at that.
>>>>>>
>>>>>>Nothing is opensource programs is a "secret" , I mean - even if they were not
>>>>>>opensource , the amateur (and definitely professional) authors will have arrived
>>>>>>(or already have) at them independently : by expiriments or through available
>>>>>>docs. I dont really see any ground breaking code or idea in any of the
>>>>>>opensource engines today - all are straight forward implementations of the
>>>>>>standard theory.
>>>>>>
>>>>>>Most , if not all , clone accusations show as "proof" something really dumb -
>>>>>>like string search , result in a single position , behaviour of a parser (which
>>>>>>is _not_ part of the engine as such people !) , etc - maybe these are the only
>>>>>>possible ways to identify clones (I am not sure - and as I have posted before ,
>>>>>>I dont really care much) , but based on such flimsy grounds people should not
>>>>>>accuse others.
>>>>>>When you accuse a program as a clone - you are also maligning the reputation of
>>>>>>the author : which is the more serious thing IMHO.
>>>>>>
>>>>>>What Vladimir Elin is hinting at is that (I think) , people (usually non-chess
>>>>>>engine programmers who know quiet little about the programming aspects) see
>>>>>>something/anything strange (in their eyes) and cry wolf.
>>>>>>Like a string search which returns strings - which might be what is defined in
>>>>>>the pgn spec , or a binary search which returns data match (whcih might be de to
>>>>>>a generated parser for pgn handling for book) , etc !
>>>>>>
>>>>>>Ofcourse wachful people are always needs to see the hints which will lead to the
>>>>>>identification of many clones , but IMO we need a better way to decide how to
>>>>>>identify clones.
>>>>>>
>>>>>>The current process seems to be : 1) Accuse 2) Flame 3) Author defends 4)
>>>>>>mudslinging 5) Nasty posts - brining the author's whole family history to the
>>>>>>ground 6) Challenge (to show source) 7) If 6 accepted , cleared , else branded
>>>>>>as clone.
>>>>>>  I dont know about others, but no I am never going to send my source code to
>>>>>>someone I dont personally trust - even if the rest of the computer chess
>>>>>>community might seem to.
>>>>>>Not everyone knows what the non-opensource guys are doing in their code : and
>>>>>>personally I do many a stupid things , but I might have something interesting
>>>>>>too :)
>>>>>>
>>>>>>So why have opensource engines ? - different question anyway , we wont discuss
>>>>>>that since it is largely an authors decision.
>>>>>>
>>>>>>We should try to promote the number of amateur engines so that as many people as
>>>>>>possible should enter this field - not discourage people. (both within
>>>>>>reasonable limits ofcourse)
>>>>>>
>>>>>>Mridul
>>>>>
>>>>>Mridul --
>>>>>
>>>>>First of all, your post makes me wonder if you are familiar with the Patriot 2.0
>>>>>situation, but anyway those things are boring for me so let's talk philosophy
>>>>>for a second :)
>>>>>
>>>>>I've had the good or maybe bad luck to spend at least 3 years living in five
>>>>>different countries, and I can make a certain observation. In two of these
>>>>>countries - USA and Germany - society essentially works. Wages are decent, crime
>>>>>is kept down, things just work. In three of these countries - Hungary, Czech
>>>>>Republic and Poland - no offense intended to anyone, but they just don't work as
>>>>>well. People steal from the government without getting punished, people cheat on
>>>>>their jobs, nobody is willing to deal with various problems, etc.
>>>>>
>>>>>What's the difference? A huge difference is that in Germany and USA, people
>>>>>essentially care. If they see something wrong, they report it and attempt to
>>>>>rectify it. This goes from cleaning up a small mess on the road, to calling the
>>>>>police if the neighbor is beating his family, etc. Throughout Eastern Europe,
>>>>>people are apathetic - and everybody suffers as a result.
>>>>>
>>>>>Sometimes, it can seem a bit too much. I remember I had this impression when I
>>>>>first came to the US - why is everybody so concerned with things that aren't
>>>>>their business. In the overall picture, though, society is better for it.
>>>>>
>>>>>So I certainly appreciate that there are people who are going to look into these
>>>>>things and do something about it, rather than just endlessly holding their
>>>>>tongue for fear of being out of line. Without it, computer chess will just be a
>>>>>mess.
>>>>>
>>>>>Vas
>>>>
>>>>Hi Vas,
>>>>
>>>>  Like I mentioned myself , we need people who will point out the
>>>>errors/suspicions.
>>>>But these are just that suspicions - a 35% binary match of the executables (egtb
>>>>will account for that ;-) ) , a small set of common strings , a bug in the fen
>>>>parser (I have seen multiple people misread/misinterpret the same spec - there
>>>>will be grey areas even in the most well written specs) , etc are not enough by
>>>>a long shot to accuse something as a clone - they can be indicators of a
>>>>potential clone at best.
>>>>
>>>>I visit CCC less and less nowadays - and each time I do so , a new program seems
>>>>to be accused of being a clone : personally I dont care , it is a hobby for me ,
>>>>something I use to fill my remaining freetime with when I am not busy with other
>>>>opensource projects I am involved with - but true , there are people who take it
>>>>seriously and for them and for the future (if not for other reasons) we should
>>>>try to keep this field as clean as possible.
>>>>
>>>>But that should not be at the expense of any tom dick and harry coming out and
>>>>accusing programs of being a clone.
>>>>That is why I said - we should have a better process for clone issue : how clone
>>>>suspicions are raised , how they are probed into , how they are proved/disproved
>>>>, etc : a bunch of amateurish tests should not be the basis of flame wars here.
>>>>Makes the whole forum (and field for that matter) more and more unreadable and
>>>>uninteresting.
>>>>
>>>>  The analogy you raised is not really valid in this context (IMHO :) ).
>>>>I wrote a long response in this space on that - and then removed it.
>>>>That is not the matter we are discussing here :) - it will most probably only
>>>>expose my ignorance of the issues concerned since I have never physically been
>>>>to the places mentioned like you though I am made aware of the ground realities
>>>>through my friends.
>>>>
>>>>Anyway , you are correct about the first point - my understanding (from what
>>>>little I read among the accusations and counteraccusations that kept flying
>>>>around) was that Patriot 2 was accused of being a clone , author did not expose
>>>>the source code , branded as a clone by the community here based on the
>>>>circumstantial evidence found (I read a few - maybe I missed a lot more) and the
>>>>ones I read looked not very solid to me (I have not done any research on Patriot
>>>>and never used it for that matter , so likely that I am missing the finer points
>>>>of the Patriot2 clone issue).
>>>>
>>>>My main problem with these accusations is that :
>>>>
>>>>1) People here follow the maxim guilty until proven innocent.
>>>>
>>>>2) Sensationalism in the accusations - I see more and more of this in the media
>>>>where it is better to say something bad to get the max amount of publicity and
>>>>attention : same thing is being "imported" into CCC.
>>>>
>>>>3) In general , it is the author's reputation which is more at stake than the
>>>>program as such , and mudslinging is affecting the author's reputation (the
>>>>accusations might or mightnot be correct).
>>>>Hence , even if something is disproved - the result is not going to remove the
>>>>damage already done to it !
>>>>
>>>>"An arrow which leaves the bow and a word which leaves the mouth cannot be taken
>>>>back" - an old saying here :)
>>>>
>>>>I believe in "Innocent until proven guilty" too much (it might be a
>>>>cultural/upbringing thing, not sure) , which is why these discussions disturb me
>>>>more.
>>>>Very few people seem to stick up for the author in general - like Peter Skinner
>>>>seems to be doing right now (just skimmed through the posts now) , the more
>>>>vocal group are the ones who are accusing. The others seem to be maintaining a
>>>>studied silence - true , you should try not to react until you get all the facts
>>>>- mark of a wise man , but sometimes it galls me when the more vocal group makes
>>>>the community believe in an issue just because they keep repeating it and the
>>>>others dont challenge or respond until everyone believes it !
>>>>I have seen way too many "discussions" of this nature in other forums online and
>>>>now recently in CCC also.
>>>>Justice happens when both sides are looked at impartially : assuming people are
>>>>really interested in getting to the roots of the problem. (which most of us in
>>>>CCC here are I assume).
>>>>
>>>>Note : even now I am not really saying whether the programs are clones or not ,
>>>>I dont have the data , unfortunately neither the time to investigate , or the
>>>>patience for it right now and really appreciate the work people are puttig in
>>>>this work.
>>>>
>>>>
>>>>Thanks,
>>>>Mridul
>>>
>>>Hi Mridul --
>>>
>>>Ok there is no way I can write something intelligent at this hour :)
>>>
>>>There is a balance of course between persecuting too many innocents, and
>>>defending too many who are guilty.
>>>
>>>I guess I don't see this particular cause as very attractive.
>>>
>>>Here is some more stuff about it:
>>>
>>>http://www.uciengines.de/UCI-Engines/Patriot/Patriot2/hauptteil_patriot2.html
>>>
>>>But frankly - I don't really care. :)
>>>
>>>Vas
>>
>>Hi Vas,
>>
>>  Thanks for the link - will look into it later today.
>>My understanding of the legal system (atleast over here) is that even if a
>>thousand guilty escape , one innocent should not be punished.
>
>Maybe 1000 is not a number that is big enough but I believe that there is X that
>is big enough that  it is better that one innocent person will be punished and
>not X quilty people escape.
>
>The only way to be 100% sure of not putting innocent people in prison is to put
>nobody in prison.
>
>It is always possible that all the people are lying or that the person is
>unlucky to be similiar to the real criminal so people say that they saw him but
>they really saw someone else.
>
>even finger print is no proof because it may be possible that 2 humans have
>similiar finger.
>
>I see no 100% proof.
>Even if the source code is public and 2 programmers have exactly the same code
>then in theory it is possible that both thought the same things without copying.
>
>probability is practically 0 but I cannot say that it is 0 because even
>1/(10^10000) is not 0.
>
>Uri

BTW - what you say here is equivalent to saying that punishing a guilty person
has some positive value.

(This is a clearer form I think.)

Vas



This page took 0.11 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.