Author: Enrique Irazoqui
Date: 08:25:48 10/31/99
Go up one level in this thread
On October 31, 1999 at 09:47:36, Robert Hyatt wrote: >On October 31, 1999 at 05:49:51, Enrique Irazoqui wrote: > >>On October 31, 1999 at 02:23:47, Ed Schröder wrote: >> >>>>Posted by Robert Hyatt on October 30, 1999 at 10:16:22: >>>> >>>>>Joining public conversations like in CCC has a lot of side effects which >>>>>may force them to stay away or limit themselves to the read-only status. >>>> >>>>Then don't "join in" here. If you have something that you think is clever, >>>>release your program. After the 'new batch of releases' are out by everyone >>>>else, write it up and submit it to the JICCA. You get no questions, not >>>>bugging about release dates, etc. >>>> >>>>Remember that _I_ get the _same_ questions, criticisms, etc here also, and I >>>>am _not_ a commercial vendor. It just comes with the territory. >>> >>>Unreasonable to ask. >>> >>>Not everybody likes to write for a magazine. Writing in CCC is easy, some >>>quick notes and that's it. Writing for a magazine is quite different. I did >>>it once (1984 I believe) for the dutch computer chess magazine "Computer >>>Schaak" when I was asked to explain about Rebel's selective search and >>>my own invented A/B algorithm because everybody was surprised to see Rebel >>>hitting 6 plies on a TRS-80 1.77 Mhz. >>> >>>I hated it. I hated it because after 3-4 days of hard work the editors told me >>>they didn't understand the contents and they wanted me to do it all over. Later >>>much later after version 4.0 (and with the help of the editor) the article was >>>ready for publication and frankly I must say that it looked a lot better than my >>>first try. I said never again. The ICCA has asked my several times to >>>contribute which I kindly refused for the above reason. >>> >>>You know, writing good articles requires some talent. I know I don't have it. >>> >>> >>>>>I don't feel obliged to contribute my sources or idea's in public >>>>>because you choose to do so. You have chosen to make your sources >>>>>public and a lot of people love you for that. But you can't force >>>>>others to do as you do and then if they don't do that, or don't >>>>>do that enough in your opinion, make them look bad in public for >>>>>reasons which are not so clear after some consideration. >>>>> >>>>>Ed >>>> >>>> >>>>I am simply trying to show that progress for _all_ would go faster. Or do >>>>you think that your ideas are good, but that there are no other good ideas >>>>to compensate you for exposing yours? That is the issue >>> >>>Oh yes, if all would open their sources we would have a super breakthrough >>>in computer chess. >>> >>>In Holland we call this "a shot for open goal" :) >>> >>>Ed >> >>After reading this long "bickering" thread, I don't think it makes any sense at >>all to deny that commercials owe a great deal to the research and publications >>of non-commercials, from alpha-beta to null-move. A case has been made about who >>gave null-move as used today, but even if the publication of Chrilly Donninger's >>paper was the direct source of information for commercial programmers, Chrilly >>himself would have developed his null-move implementation after the findings and >>publications of others. So directly or indirectly the level of achievement of >>today's commercial chess programs would not have been possible without this body >>of knowledge provided by "amateurs", which by the way is the wrong word that has >>been used in this discussion to refer to non-commercial chess programmers. >>Programmers, freaks like me and end-users should be grateful to them. >> >>On the other hand, even if I understand and even sympathize with Bob's >>frustration about commercial programmers not publishing their findings, I think >>that his idea of asking commercials to release all information after a year is >>not realistic. In the first place, I am not sure that they make significant and >>steady progress every year. Secondly, even if they make this progress one year, >>they have no guarantee that during the next year they would find noticeable >>improvements, so they must save for a rainy day. Finally, the "let's put >>together what we all know" has been traditionally the goal of University people, >>but never of the industry, and commercials are industry. It is like criticizing >>commercials for being commercial. >> >>So I find not reasonable that commercials deny what they owe to researchers and >>as unreasonable that researchers demand the industry to act as if they were not >>industry. >> >>Enrique > > >I think even industry is not that secret. As I mentioned every microprocessor >manufacturer has released internal technical details very quickly, so that they >get 'on record' as the inventor of a specific idea. As I said I have video >cassettes of lectures given by each microprocessor manufacturer (by members of >the design team in fact) as specific microprocessor architectures were released. >Alpha, HP-PA, IBM SP, Intel, etc... And the journals are full of their reports. I guess that this depends on to what extent the industry can protect their products after publishing their findings. In chess programming there is no protection: everybody can use what you publish. Probably the weapon industry is an extreme example, but it may work. What I read about it doesn't go beyond John Le Carré novels, but I doubt that Los Alamos published in the 40s their findings about nuclear weapons, or more recently about the neutron bomb. That's why competing powers were and are spying each other like nuts, and that's why industrial espionage is flourishing: they don't reveal, they don't publish and you don't expect them to. And commercial chess programmers must compete quite intensely (certainly not war, even if at times it looks like it) for a very small niche in the market. Enrique
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.