Frank, I'm a physicist. The fact that I work now as an engineer does not mean that I'm a bad physicist. Let me explain it to you.
First, all mathematically logical theories are based on some axioms. These axioms are hypothesises that are most likely correct and have not been able to be proved. Thus, based on axioms alone, the validity of a theory cannot be determined. Thus, there comes the second criteria for the validity of a theory, the methematical consistency of the theory. For physical science above two criteria are still not enough, a third one is used. This third criteria is that the theory must be able to describe the physical process of the nature to a certain extent. If you attack a theory by attacking the axiom, you better either disaprove the axiom or you establish a new theory on an opposite axiom which satisfy the second criteria (in case of mathematics) or satisfy both the second and third criteria (in case of physics). I would like to mention that you didn't accomplish it. Of course you may attack the above three criteria themselves. However, these three criteria are the foundation of the human intellect. So you better come up something that's real good. Otherwise your opinion is of no difference than the birds' singing.
Now go back to physics. What's physics? Physics is to describe the nature consistently with mathematics. You noticed that it didn't say that physics is to describe the nature as exactly what it is. Special theory of relativity not only described the nature correctly in experiments performed so far, it also predicted important new phenomena not know at the time. Such as E=mc2. Ok, you say the time didn't change in the case of the atomic oscillator and the ocsillator slowed instead. As I mentioned early, you need a mathematic theory to establish your arguement. Otherwise you didn't make any statement in the physics sense. Since you didn't make any statement, no body can argue with you.
Back to Go. If the correct formulism is used, is it possible to extract the ultimate (or best available) Go knowledge from the Go game records? Yes, maybe or maybe not, because nobody has proved or disprove it yet. So should we implement go knowledge in the program to win the game or should we do nothing until someone prove or find a way to automatically extract the ultimate Go knowledge from game records? The answer is obvious.
Daniel Liu
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/