Author: Robert Hyatt
Date: 05:39:22 01/07/99
Go up one level in this thread
On January 07, 1999 at 01:06:59, blass uri wrote: > >On January 06, 1999 at 19:05:58, Robert Hyatt wrote: > >>On January 06, 1999 at 18:43:38, Bert Seifriz wrote: >> >>> >>>>>>>Microsoft could probably have the strongest engine in the world in 2 years and >>>>>>>if they wanted to, >>>>> >>>>> >>>>>This is exactly the point. Why should Microsoft care to build a >>>>>chess engine? They do not have the slightest reason! >>>>>And if? They would either buy an existing company or >>>>>engage some of the chess programmers you already know. >>>>>After all why do you think the persons we know are in any >>>>>way inferior to a Microsoft engineer? >>>>>There are for example only a handful of persons in the world >>>>>who are able to program chess in assembly language! And these >>>>>are not MS engineers! Bert >>>> >>>>Sorry, but I'm absolutely sure that I *can* write reasonable >>>>good chess engine. Actually, I had done that 9 years ago (you >>>>can download Siberian Chess from GambitSoft chess site; >>> >>>Hi Eugene, >>>I know that because I put it there! >>> >>>>was not modified after 1991, I think; and then I knew much less, >>>>had no access to Western publications/experience, etc). Yes, >>>>it was written mainly in C, only absolutely time-critical parts >>>>of it were in assembly (first 8080, later I rewrote in 8086). >>>>But if necessary, I can write engine in "100% pure assembly". >>>> >>>>And I'm Microsoft engineer :-) >>> >>>I know that, too! What I wanted to say is that the chess programmers >>>around are in no way inferior to any other software engineers. >>>You certainly know >>>more about Microsoft than others who do not work there. But the original >>>question was somehow what MS could develop within a year. My >>>opinion is they could not overtake the current status of chess >>>programming within that time. Remember IBM's endless efforts with >>>Deep Blue. Although that was maybe more a hardware problem. >>>But when MS has more engineers with a past like you it might be >>>different!! Bert >> >> >>I don't follow your argument now. IE WRT IBM's DB project. Deep Thought >>came along around 1987, and it was *always* superior to any commercial chess >>engine. DB took it a step forward beyond that level... > >The arguement was that they needed more than one year to develop deeper blue. > >I do not think that it is relevant because they did not throw enough money at >depper blue(they did not throw 1,000,000,000$) > >Deeper blue is also not relevant because we are interested in the program and >not in the hardware. > >The question if they can do a better program that I can use. > >Uri this is not quite accurate. The first "real" chiptest played in 1987, one year after the project was started. It blew _everyone_ away at the 1987 ACM computer chess event. _easily_. And it did nothing but get better. Once you understand silicon compilers and automated fab processes (like project MOSIS) you discover that "casting a chess program into silicon is not nearly as difficult as you'd originally think, if you feel comfortable with hardware design in the first place." However, chiptest was a quantum leap forward and took less than one year to pass everybody. And they kept improving it, adding more processors, until they eventually reached the machine that beat Kasparov... It didn't take 10 years to produce something that worked well...
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.