Re: computer-go: Authenticating the identity of a remote go-playing computer program

[Don Dailey <drd@xxxxxxxxxxxxxxxxxxxxx>]

   From: birk@xxxxxxxxxxxxxxxxx

   Hi Don,

   > Everything about  this is  based  on   making sure the  programs   are
   > deterministic

   If the program is deterministic, then the problem is solved. I will
   play the same moves again and again. You don't need any sophisticated
   authenticating mechanisms. 

NO,  deterministic does   NOT mean that  your program   is required to
always play  the same move  in the  same position.   No  one seems  to
understand this  point.   The solution  is  that  you are allowed   to
introduce known state (such  as a random  number seed) to the program,
or publish one   at the start of  the  game.  This  makes the  program
completely  determnistic, and  yet unpredictable   from your opponents
point of view.

You   are  basically   right  about  not    needing  a   sophisticated
authentication mechanism, all  you need is  a deterministic program, a
random starting seed  that the opponents can  share, and a copy of the
program sent to  a trusted arbiter who can  resolve any  disputes that
arise.  In principle it's that simple.  I already  have a program that
complies with  these requirements, but  it's easy  to imagine programs
written in a way that makes this a lot more difficult.

   Christoph