Re: [computer-go] KGS Tournaments: Uniqueness of players

[Evan Daniel <evanbd@xxxxxxxxx>]

On Mon, 7 Mar 2005 16:18:50 +0000, Stuart Yeates <syeates@xxxxxxxxxxxxxxxxx> wrote:
> On Mon, 7 Mar 2005 13:23:22 +0000, Nick Wedd <nick@xxxxxxxxxxxxxxxxx> wrote:
> 
> > I am therefore going to express a number of opinions (not all
> > consistent), and hope that people cleverer than me can find a sensible
> > way of defining "duplicate entries".
> 
> One effective way of handling this might be to require that any
> competitors who share "game playing logic" (as opposed to I/O bridges
> / gateways and other software used to connect the game playing logic
> to the server) play in a preliminary round to select a single winner
> to go forward into the main round(s).

What about players who share a board library and nothing else?  And
aren't GNU Go, SlugGo, and GoFigure all legitimately different
programs?  If not, then how do you account for the fact that they
produce significantly different moves?  What about programs that use
their own logic most of the time, but ask GNU Go to handle dead stone
identification at the end of the game?

I think that any exact definition is more of a pain to come up with
and enforce than it's worth.  I think a couple basic criteria should
handle most cases:

Duplicates of programs aren't allowed, even if settings are different.
 Substantial changes that are basically separate but derivative works
are allowed (only with the consent of the original author, obviously),
but they need to produce substantially different moves.

I think the big things you want to prohibit are plagiarism and random
people entering copies of publically available programs.  My guess is
that there actually won't be that many cases that are actually in
question.

Evan
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/