Author: Uri Blass
Date: 17:49:09 12/12/05
Go up one level in this thread
On December 12, 2005 at 20:36:33, Dann Corbit wrote:
>On December 12, 2005 at 20:07:08, Uri Blass wrote:
>
>>On December 12, 2005 at 19:30:50, stuart taylor wrote:
>>
>>>It's very interesting, and I don't think it ever happened before, that
>>>non-commercial programs came out, and became rulers over the commercial ones.
>>>And not only that, but SEVERAL have now done just that.
>>>
>>>How did it suddenly happen all at once?
>>>
>>>Perhaps there were a few very talented programmers who all thought they were
>>>going to outsmart the best, and then one released his, and then the other one
>>>broke his silence, and showed he could do the same (or better), and then they
>>>ALL went wild together? There are about atleast 4 programs which should make
>>>stunning headlines, BUT, they all have rivals!
>>>
>>>Or perhaps there was a good book that came out, for programming something
>>>special, and all those who saw it, went for it. But alas, they had company.
>>>
>>>How and why now?
>>
>>You can blame Fabien about it.
>>After he released fruit2.1 other learned.
>
>I doubt it (for the most part) {though it is also clear that he has advanced
>state of the art computer chess programs}.
>
>Zappa was world champion before any Fruit code became available (IIRC).
No
Fruit2.1 was avialbale before zappa became world champion.
>Rybka seems rather different than Fruit.
>Similarly for Spike.
I did not say that they are clones of fruit but only that programmers learned
from fruit.
>
>I think that the situation is similar to what was going on in Newton's time
>frame.
>
>Newton, Leibnitz [and Pascal to some degree] simultaneously developed calculus.
>Hooke claimed also some of it to be his invention.
>
>But the precursors had built up to the point where *someone* was going to invent
>it eventually.
>
>I attribute the leaps and bounds improvments to the following:
>1. Lots of good, public, new ideas
>2. Some good private new ideas
>3. Some really good programmers getting interested in computer chess.
>
>Ten years from now, Fruit-level programs will be the norm and not the exception,
>even for a beginner. That is because information is advancing and smart people
>are adding their input. It's like a snowball rolling downhill. It can't help
>but get bigger.
>
>Fabien is responsible for some innovations. I guess that 2/3 of the chess
>programmers don't even understand them yet. At some point, it will become clear
>but by then there will be new fresh ideas. Probably just as important as having
>new ideas is implementing the ideas in a robust and correct manner. I think
>that is the most important lesson we can glean from Fabien's code.
>
>P.S.
>I predict he will have a very hard time writing a SMP version because of all the
>global variables.
I do not know that fruit has global variables.
Unlike tscp that has one file of external variables fruit has many files of
external variables ane I know no variable that it external to all files.
So catching Anthony Cozzie might be hard in the near term
>[before the end of 2006]. I do not know about the organization of Rybka's code.
We are going to see but my guess is that fruit is going to run with parallel
search at the middle of 2006.
Uri
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.